1.SlowFast Networks for Video Recognition pdf
We present SlowFast networks for video recognition. Our model involves (i) a Slow pathway, operating at low frame rate, to capture spatial semantics, and (ii) a Fast pathway, operating at high frame rate, to capture motion at fine temporal resolution. The Fast pathway can be made very lightweight by reducing its channel capacity, yet can learn useful temporal information for video recognition. Our models achieve strong performance for both action classification and detection in video, and large improvements are pin-pointed as contributions by our SlowFast concept. We report 79.0% accuracy on the Kinetics dataset without using any pre-training, largely surpassing the previous best results of this kind. On AVA action detection we achieve a new state-of-the-art of 28.3 mAP. Code will be made publicly available.
2.Accuracy, Uncertainty, and Adaptability of Automatic Myocardial ASL Segmentation using Deep CNN pdf
PURPOSE: To apply deep convolutional neural networks (CNN) to the left ventricular segmentation task in myocardial arterial spin labeled (ASL) perfusion imaging. To develop methods that measure uncertainty and that adapt segmentation based on a specific false positive vs. false negative tradeoff.
METHODS: We utilized a modified UNET architecture with Monte Carlo (MC) dropout. The model was trained on data from 22 subjects and tested on data from 6 heart transplant recipients. Manual segmentation and quantitative myocardial blood flow (MBF) maps were available for comparison. We consider two global scores of segmentation uncertainty, "Dice Uncertainty" and "MC Uncertainty", which were calculated with and without the use of manual segmentation, respectively. Tversky loss function with a hyperparameter$β$ was used to adapt the model to a specific false positive vs. false negative tradeoff.
RESULTS: The modified UNET model achieved Dice coefficient of mean(std) =$0.91(0.04)$ on the test set. MBF measured using automatic segmentation was highly correlated to that measured using the manual segmentation ($R^2 = 0.96$ ). Dice Uncertainty and MC Uncertainty were in good agreement ($R^2 = 0.64$ ). As$β$ increased, the false positive rate systematically decreased and false negative rate systematically increased.
CONCLUSION: We demonstrate the feasibility of using CNN for automatic segmentation of the left ventricle in myocardial ASL data. This is a particularly challenging application because of low and inconsistent blood-myocardium contrast-to-noise ratio. We also demonstrate novel methods that measure uncertainty and adapt to a desired tradeoff between false positive and false negative rates.
3.A New Ensemble Learning Framework for 3D Biomedical Image Segmentation pdf
3D image segmentation plays an important role in biomedical image analysis. Many 2D and 3D deep learning models have achieved state-of-the-art segmentation performance on 3D biomedical image datasets. Yet, 2D and 3D models have their own strengths and weaknesses, and by unifying them together, one may be able to achieve more accurate results. In this paper, we propose a new ensemble learning framework for 3D biomedical image segmentation that combines the merits of 2D and 3D models. First, we develop a fully convolutional network based meta-learner to learn how to improve the results from 2D and 3D models (base-learners). Then, to minimize over-fitting for our sophisticated meta-learner, we devise a new training method that uses the results of the base-learners as multiple versions of "ground truths". Furthermore, since our new meta-learner training scheme does not depend on manual annotation, it can utilize abundant unlabeled 3D image data to further improve the model. Extensive experiments on two public datasets (the HVSMR 2016 Challenge dataset and the mouse piriform cortex dataset) show that our approach is effective under fully-supervised, semi-supervised, and transductive settings, and attains superior performance over state-of-the-art image segmentation methods.
4.Data Fine-tuning pdf
In real-world applications, commercial off-the-shelf systems are utilized for performing automated facial analysis including face recognition, emotion recognition, and attribute prediction. However, a majority of these commercial systems act as black boxes due to the inaccessibility of the model parameters which makes it challenging to fine-tune the models for specific applications. Stimulated by the advances in adversarial perturbations, this research proposes the concept of Data Fine-tuning to improve the classification accuracy of a given model without changing the parameters of the model. This is accomplished by modeling it as data (image) perturbation problem. A small amount of "noise" is added to the input with the objective of minimizing the classification loss without affecting the (visual) appearance. Experiments performed on three publicly available datasets LFW, CelebA, and MUCT, demonstrate the effectiveness of the proposed concept.
5.Self-Contained Stylization via Steganography for Reverse and Serial Style Transfer pdf
Style transfer has been widely applied to give real-world images a new artistic look. However, given a stylized image, the attempts to use typical style transfer methods for de-stylization or transferring it again into another style usually lead to artifacts or undesired results. We realize that these issues are originated from the content inconsistency between the original image and its stylized output. Therefore in this paper we advance to keep the content information of the input image during process of style transfer by the power of steganography, with two approaches proposed: a two-stage model and an end-to-end model. We conduct extensive experiments to successfully verify the capacity of our models, in which both of them are able to not only generate stylized images of quality comparable with the ones produced by state-of-the-art style transfer methods, but also effectively eliminate the artifacts introduced in reconstructing original input from a stylized image as well as performing multiple times of style transfer in series.
6.Attention-guided Unified Network for Panoptic Segmentation pdf
This paper studies panoptic segmentation, a recently proposed task which segments foreground (FG) objects at the instance level as well as background (BG) contents at the semantic level. Existing methods mostly dealt with these two problems separately, but in this paper, we reveal the underlying relationship between them, in particular, FG objects provide complementary cues to assist BG understanding. Our approach, named the Attention-guided Unified Network (AUNet), is an unified framework with two branches for FG and BG segmentation simultaneously. Two sources of attentions are added to the BG branch, namely, RPN and FG segmentation mask to provide object-level and pixel-level attentions, respectively. Our approach is generalized to different backbones with consistent accuracy gain in both FG and BG segmentation, and also sets new state-of-the-arts in the MS-COCO (46.5% PQ) benchmarks.
7.Facial Landmark Machines: A Backbone-Branches Architecture with Progressive Representation Learning pdf
Facial landmark localization plays a critical role in face recognition and analysis. In this paper, we propose a novel cascaded backbone-branches fully convolutional neural network~(BB-FCN) for rapidly and accurately localizing facial landmarks in unconstrained and cluttered settings. Our proposed BB-FCN generates facial landmark response maps directly from raw images without any preprocessing. BB-FCN follows a coarse-to-fine cascaded pipeline, which consists of a backbone network for roughly detecting the locations of all facial landmarks and one branch network for each type of detected landmark for further refining their locations. Furthermore, to facilitate the facial landmark localization under unconstrained settings, we propose a large-scale benchmark named SYSU16K, which contains 16000 faces with large variations in pose, expression, illumination and resolution. Extensive experimental evaluations demonstrate that our proposed BB-FCN can significantly outperform the state-of-the-art under both constrained (i.e., within detected facial regions only) and unconstrained settings. We further confirm that high-quality facial landmarks localized with our proposed network can also improve the precision and recall of face detection.
8.Video Colorization using CNNs and Keyframes extraction: An application in saving bandwidth pdf
In this paper we tackle the problem of Colorization of Grayscale Videos given a few colored keyframes of the video. We propose a model that extracts keyframes from the video and trains a Convolutional Neural Network from scratch on these colored frames. Once the model has been trained we colorize the rest of the frames of the video given their grayscale versions. An application of this technique is in saving bandwidth while sending raw colored videos that haven't gone through any compression. A raw colored video takes upto three times more memory than it's grayscale version. We can exploit this fact and send a grayscale video along with our trained model instead of a colored one. This will reduce bandwidth usage to upto three times.
9.Weakly Supervised Dense Event Captioning in Videos pdf
Dense event captioning aims to detect and describe all events of interest contained in a video. Despite the advanced development in this area, existing methods tackle this task by making use of dense temporal annotations, which is dramatically source-consuming. This paper formulates a new problem: weakly supervised dense event captioning, which does not require temporal segment annotations for model training. Our solution is based on the one-to-one correspondence assumption, each caption describes one temporal segment, and each temporal segment has one caption, which holds in current benchmark datasets and most real-world cases. We decompose the problem into a pair of dual problems: event captioning and sentence localization and present a cycle system to train our model. Extensive experimental results are provided to demonstrate the ability of our model on both dense event captioning and sentence localization in videos.
10.Occupancy Networks: Learning 3D Reconstruction in Function Space pdf
With the advent of deep neural networks, learning-based approaches for 3D reconstruction have gained popularity. However, unlike for images, in 3D there is no canonical representation which is both computationally and memory efficient yet allows for representing high-resolution geometry of arbitrary topology. Many of the state-of-the-art learning-based 3D reconstruction approaches can hence only represent very coarse 3D geometry or are limited to a restricted domain. In this paper, we propose occupancy networks, a new representation for learning-based 3D reconstruction methods. Occupancy networks implicitly represent the 3D surface as the continuous decision boundary of a deep neural network classifier. In contrast to existing approaches, our representation encodes a description of the 3D output at infinite resolution without excessive memory footprint. We validate that our representation can efficiently encode 3D structure and can be inferred from various kinds of input. Our experiments demonstrate competitive results, both qualitatively and quantitatively, for the challenging tasks of 3D reconstruction from single images, noisy point clouds and coarse discrete voxel grids. We believe that occupancy networks will become a useful tool in a wide variety of learning-based 3D tasks.
11.Learning to Drive from Simulation without Real World Labels pdf
Simulation can be a powerful tool for understanding machine learning systems and designing methods to solve real-world problems. Training and evaluating methods purely in simulation is often "doomed to succeed" at the desired task in a simulated environment, but the resulting models are incapable of operation in the real world. Here we present and evaluate a method for transferring a vision-based lane following driving policy from simulation to operation on a rural road without any real-world labels. Our approach leverages recent advances in image-to-image translation to achieve domain transfer while jointly learning a single-camera control policy from simulation control labels. We assess the driving performance of this method using both open-loop regression metrics, and closed-loop performance operating an autonomous vehicle on rural and urban roads.
12.Image-based Navigation using Visual Features and Map pdf
Building on progress in feature representations for image retrieval, image-based localization has seen a surge of research interest. Image-based localization has the advantage of being inexpensive and efficient, often avoiding the use of 3D metric maps altogether. This said, the need to maintain a large number of reference images as an effective support of localization in a scene, nonetheless calls for them to be organized in a map structure of some kind.
The problem of localization often arises as part of a navigation process. We are, therefore, interested in summarizing the reference images as a set of landmarks, which meet the requirements for image-based navigation. A contribution of the paper is to formulate such a set of requirements for the two sub-tasks involved: map construction and self localization. These requirements are then exploited for compact map representation and accurate self-localization, using the framework of a network flow problem. During this process, we formulate the map construction and self-localization problems as convex quadratic and second-order cone programs, respectively. We evaluate our methods on publicly available indoor and outdoor datasets, where they outperform existing methods significantly.
13.Can we learn where people go? pdf
In most agent-based simulators, pedestrians navigate from origins to destinations. Consequently, destinations are essential input parameters to the simulation. While many other relevant parameters as positions, speeds and densities can be obtained from sensors, like cameras, destinations cannot be observed directly. Our research question is: Can we obtain this information from video data using machine learning methods? We use density heatmaps, which indicate the pedestrian density within a given camera cutout, as input to predict the destination distributions. For our proof of concept, we train a Random Forest predictor on an exemplary data set generated with the VADERE microscopic simulator. The scenario is a crossroad where pedestrians can head left, straight or right. In addition, we gain first insights on suitable placement of the camera. The results motivate an in-depth analysis of the methodology.
14.Efficient Condition-based Representations for Long-Term Visual Localization pdf
We propose an approach to localization from images that is designed to explicitly handle the strong variations in appearance happening when capturing conditions change throughout the day or across seasons. As revealed by recent long-term localization benchmarks, both traditional feature-based and retrieval-based approaches still struggle to handle such changes. Our novel retrieval-based method introduces condition-specific sub-networks allowing the computation of global image descriptors that are explicitly dependent of the capturing conditions. We compare our approach to previous localization methods on very recent challenging benchmarks, and observe that our method outperforms them by a large margin in case of day-night variation, where repeatable feature points cannot be identified or matched.
15.Defending against Universal Perturbations with Shared Adversarial Training pdf
Classifiers such as deep neural networks have been shown to be vulnerable against adversarial perturbations on problems with high-dimensional input space. While adversarial training improves the robustness of image classifiers against such adversarial perturbations, it leaves them sensitive to perturbations on a non-negligible fraction of the inputs. In this work, we show that adversarial training is more effective in preventing universal perturbations, where the same perturbation needs to fool a classifier on many inputs. Moreover, we investigate the trade-off between robustness against universal perturbations and performance on unperturbed data and propose an extension of adversarial training that handles this trade-off more gracefully. We present results for image classification and semantic segmentation to showcase that universal perturbations that fool a model hardened with adversarial training become clearly perceptible and show patterns of the target scene.
16.SMIT: Stochastic Multi-Label Image-to-Image Translation pdf
Cross-domain mapping has been a very active topic in recent years. Given one image, its main purpose is to translate it to the desired target domain, or multiple domains in the case of multiple labels. This problem is highly challenging due to three main reasons: (i) unpaired datasets, (ii) multiple attributes, and (iii) the multimodality associated with the translation. Most of the existing state-of-the-art has focused only on two reasons, i.e. producing disentangled representations from unpaired datasets in a one-to-one domain translation or producing multiple unimodal attributes from unpaired datasets. In this work, we propose a joint framework of diversity and multi-mapping image-to-image translations, using a single generator to conditionally produce countless and unique fake images that hold the underlying characteristics of the source image. Extensive experiments over different datasets demonstrate the effectiveness of our proposed approach with comparisons to the state-of-the-art in both multi-label and multimodal problems. Additionally, our method is able to generalize under different scenarios: continuous style interpolation, continuous label interpolation, and multi-label mapping.
17.Auto-Encoder-BoF/HMM System for Arabic Text Recognition pdf
The recognition of Arabic text, in both handwritten and printed forms, represents a fertile provenance of technical difficulties for Optical Character Recognition (OCR). Indeed, the printed is commonly governed by well-established calligraphy rules and the characters are well aligned. However, there is not always a system capable of reading Arabic printed text in an unconstrained environments such as unlimited vocabulary, multi styles, mixed-font and their great morphological variability. This diversity complicates the choice of features to extract and algorithm of segmentation. In this context, we adopt a new solution for unlimited-vocabulary and mixed-font Arabic printed text recognition. The proposed system is based on the adoption of Bag of Features (BoF) model using Sparse Auto-Encoder (SAE) for features representation and Hidden Markov Models (HMM) for recognition. As results, the obtained average accuracies of recognition vary between 99.65% and 99.96% for the mono-font and exceed 99% for mixed-font.
18.Spatial Knowledge Distillation to aid Visual Reasoning pdf
For tasks involving language and vision, the current state-of-the-art methods tend not to leverage any additional information that might be present to gather relevant (commonsense) knowledge. A representative task is Visual Question Answering where large diagnostic datasets have been proposed to test a system's capability of answering questions about images. The training data is often accompanied by annotations of individual object properties and spatial locations. In this work, we take a step towards integrating this additional privileged information in the form of spatial knowledge to aid in visual reasoning. We propose a framework that combines recent advances in knowledge distillation (teacher-student framework), relational reasoning and probabilistic logical languages to incorporate such knowledge in existing neural networks for the task of Visual Question Answering. Specifically, for a question posed against an image, we use a probabilistic logical language to encode the spatial knowledge and the spatial understanding about the question in the form of a mask that is directly provided to the teacher network. The student network learns from the ground-truth information as well as the teachers prediction via distillation. We also demonstrate the impact of predicting such a mask inside the teachers network using attention. Empirically, we show that both the methods improve the test accuracy over a state-of-the-art approach on a publicly available dataset.
19.EDF: Ensemble, Distill, and Fuse for Easy Video Labeling pdf
We present a way to rapidly bootstrap object detection on unseen videos using minimal human annotations. We accomplish this by combining two complementary sources of knowledge (one generic and the other specific) using bounding box merging and model distillation. The first (generic) knowledge source is obtained from ensembling pre-trained object detectors using a novel bounding box merging and confidence reweighting scheme. We make the observation that model distillation with data augmentation can train a specialized detector that outperforms the noisy labels it was trained on, and train a Student Network on the ensemble detections that obtains higher mAP than the ensemble itself. The second (specialized) knowledge source comes from training a detector (which we call the Supervised Labeler) on a labeled subset of the video to generate detections on the unlabeled portion. We demonstrate on two popular vehicular datasets that these techniques work to emit bounding boxes for all vehicles in the frame with higher mean average precision (mAP) than any of the reference networks used, and that the combination of ensembled and human-labeled data produces object detections that outperform either alone.
20.3D Scene Parsing via Class-Wise Adaptation pdf
We propose the method that uses only computer graphics datasets to parse the real world 3D scenes. 3D scene parsing based on semantic segmentation is required to implement the categorical interaction in the virtual world. Convolutional Neural Networks (CNNs) have recently shown state-of-theart performance on computer vision tasks including semantic segmentation. However, collecting and annotating a huge amount of data are needed to train CNNs. Especially in the case of semantic segmentation, annotating pixel by pixel takes a significant amount of time and often makes mistakes. In contrast, computer graphics can generate a lot of accurate annotated data and easily scale up by changing camera positions, textures and lights. Despite these advantages, models trained on computer graphics datasets cannot perform well on real data, which is known as the domain shift. To address this issue, we first present that depth modal and synthetic noise are effective to reduce the domain shift. Then, we develop the class-wise adaptation which obtains domain invariant features of CNNs. To reduce the domain shift, we create computer graphics rooms with a lot of props, and provide photo-realistic rendered images.We also demonstrate the application which is combined semantic segmentation with Simultaneous Localization and Mapping (SLAM). Our application performs accurate 3D scene parsing in real-time on an actual room.
21.Learning Non-Uniform Hypergraph for Multi-Object Tracking pdf
The majority of Multi-Object Tracking (MOT) algorithms based on the tracking-by-detection scheme do not use higher order dependencies among objects or tracklets, which makes them less effective in handling complex scenarios. In this work, we present a new near-online MOT algorithm based on non-uniform hypergraph, which can model different degrees of dependencies among tracklets in a unified objective. The nodes in the hypergraph correspond to the tracklets and the hyperedges with different degrees encode various kinds of dependencies among them. Specifically, instead of setting the weights of hyperedges with different degrees empirically, they are learned automatically using the structural support vector machine algorithm (SSVM). Several experiments are carried out on various challenging datasets (i.e., PETS09, ParkingLot sequence, SubwayFace, and MOT16 benchmark), to demonstrate that our method achieves favorable performance against the state-of-the-art MOT methods.
22.Task-Free Continual Learning pdf
Methods proposed in the literature towards continual deep learning typically operate in a task-based sequential learning setup. A sequence of tasks is learned, one at a time, with all data of current task available but not of previous or future tasks. Task boundaries and identities are known at all times. This setup, however, is rarely encountered in practical applications. Therefore we investigate how to transform continual learning to an online setup. We develop a system that keeps on learning over time in a streaming fashion, with data distributions gradually changing and without the notion of separate tasks. To this end, we build on the work on Memory Aware Synapses, and show how this method can be made online by providing a protocol to decide i) when to update the importance weights, ii) which data to use to update them, and iii) how to accumulate the importance weights at each update step. Experimental results show the validity of the approach in the context of two applications: (self-)supervised learning of a face recognition model by watching soap series and learning a robot to avoid collisions.
23.PoseFix: Model-agnostic General Human Pose Refinement Network pdf
Multi-person pose estimation from a 2D image is an essential technique for human behavior understanding. In this paper, we propose a human pose refinement network that estimates a refined pose from a tuple of an input image and input pose. The pose refinement was performed mainly through an end-to-end trainable multi-stage architecture in previous methods. However, they are highly dependent on pose estimation models and require careful model design. By contrast, we propose a model-agnostic pose refinement method. According to a recent study, state-of-the-art 2D human pose estimation methods have similar error distributions. We use this error statistics as prior information to generate synthetic poses and use the synthesized poses to train our model. In the testing stage, pose estimation results of any other methods can be input to the proposed method. Moreover, the proposed model does not require code or knowledge about other methods, which allows it to be easily used in the post-processing step. We show that the proposed approach achieves better performance than the conventional multi-stage refinement models and consistently improves the performance of various state-of-the-art pose estimation methods on the commonly used benchmark. We will release the code and pre-trained model for easy access.
24.Learning Style Compatibility for Furniture pdf
When judging style, a key question that often arises is whether or not a pair of objects are compatible with each other. In this paper we investigate how Siamese networks can be used efficiently for assessing the style compatibility between images of furniture items. We show that the middle layers of pretrained CNNs can capture essential information about furniture style, which allows for efficient applications of such networks for this task. We also use a joint image-text embedding method that allows for the querying of stylistically compatible furniture items, along with additional attribute constraints based on text. To evaluate our methods, we collect and present a large scale dataset of images of furniture of different style categories accompanied by text attributes.
25.Deep Spectral Reflectance and Illuminant Estimation from Self-Interreflections pdf
In this work, we propose a CNN-based approach to estimate the spectral reflectance of a surface and the spectral power distribution of the light from a single RGB image of a V-shaped surface. Interreflections happening in a concave surface lead to gradients of RGB values over its area. These gradients carry a lot of information concerning the physical properties of the surface and the illuminant. Our network is trained with only simulated data constructed using a physics-based interreflection model. Coupling interreflection effects with deep learning helps to retrieve the spectral reflectance under an unknown light and to estimate the spectral power distribution of this light as well. In addition, it is more robust to the presence of image noise than the classical approaches. Our results show that the proposed approach outperforms the state of the art learning-based approaches on simulated data. In addition, it gives better results on real data compared to other interreflection-based approaches.
26.A Structured Model For Action Detection pdf
A dominant paradigm for learning-based approaches in computer vision is training generic models, such as ResNet for image recognition, or I3D for video understanding, on large datasets and allowing them to discover the optimal representation for the problem at hand. While this is an obviously attractive approach, it is not applicable in all scenarios. We claim that action detection is one such challenging problem - the models that need to be trained are large, and labeled data is expensive to obtain. To address this limitation, we propose to incorporate domain knowledge into the structure of the model, simplifying optimization. In particular, we augment a standard I3D network with a tracking module to aggregate long term motion patterns, and use a graph convolutional network to reason about interactions between actors and objects. Evaluated on the challenging AVA dataset, the proposed approach improves over the I3D baseline by 5.5% mAP and over the state-of-the-art by 4.8% mAP.
27.A Deep Multi-task Learning Approach to Skin Lesion Classification pdf
Skin lesion identification is a key step toward dermatological diagnosis. When describing a skin lesion, it is very important to note its body site distribution as many skin diseases commonly affect particular parts of the body. To exploit the correlation between skin lesions and their body site distributions, in this study, we investigate the possibility of improving skin lesion classification using the additional context information provided by body location. Specifically, we build a deep multi-task learning (MTL) framework to jointly optimize skin lesion classification and body location classification (the latter is used as an inductive bias). Our MTL framework uses the state-of-the-art ImageNet pretrained model with specialized loss functions for the two related tasks. Our experiments show that the proposed MTL based method performs more robustly than its standalone (single-task) counterpart.
28.Skin Disease Classification versus Skin Lesion Characterization: Achieving Robust Diagnosis using Multi-label Deep Neural Networks pdf
In this study, we investigate what a practically useful approach is in order to achieve robust skin disease diagnosis. A direct approach is to target the ground truth diagnosis labels, while an alternative approach instead focuses on determining skin lesion characteristics that are more visually consistent and discernible. We argue that, for computer-aided skin disease diagnosis, it is both more realistic and more useful that lesion type tags should be considered as the target of an automated diagnosis system such that the system can first achieve a high accuracy in describing skin lesions, and in turn facilitate disease diagnosis using lesion characteristics in conjunction with other evidence. To further meet such an objective, we employ convolutional neural networks (CNNs) for both the disease-targeted and lesion-targeted classifications. We have collected a large-scale and diverse dataset of 75,665 skin disease images from six publicly available dermatology atlantes. Then we train and compare both disease-targeted and lesion-targeted classifiers, respectively. For disease-targeted classification, only 27.6% top-1 accuracy and 57.9% top-5 accuracy are achieved with a mean average precision (mAP) of 0.42. In contrast, for lesion-targeted classification, we can achieve a much higher mAP of 0.70.
29.More Knowledge is Better: Cross-Modality Volume Completion and 3D+2D Segmentation for Intracardiac Echocardiography Contouring pdf
Using catheter ablation to treat atrial fibrillation increasingly relies on intracardiac echocardiography (ICE) for an anatomical delineation of the left atrium and the pulmonary veins that enter the atrium. However, it is a challenge to build an automatic contouring algorithm because ICE is noisy and provides only a limited 2D view of the 3D anatomy. This work provides the first automatic solution to segment the left atrium and the pulmonary veins from ICE. In this solution, we demonstrate the benefit of building a cross-modality framework that can leverage a database of diagnostic images to supplement the less available interventional images. To this end, we develop a novel deep neural network approach that uses the (i) 3D geometrical information provided by a position sensor embedded in the ICE catheter and the (ii) 3D image appearance information from a set of computed tomography cardiac volumes. We evaluate the proposed approach over 11,000 ICE images collected from 150 clinical patients. Experimental results show that our model is significantly better than a direct 2D image-to-image deep neural network segmentation, especially for less-observed structures.
30.From Coarse to Fine: Robust Hierarchical Localization at Large Scale pdf
Robust and accurate visual localization is a fundamental capability for numerous applications, such as autonomous driving, mobile robotics, or augmented reality. It remains, however, a challenging task, particularly for large-scale environments and in presence of significant appearance changes. State-of-the-art methods not only struggle with such scenarios, but are often too resource intensive for certain real-time applications. In this paper we propose HF-Net, a hierarchical localization approach based on a monolithic CNN that simultaneously predicts local features and global descriptors for accurate 6-DoF localization. We exploit the coarse-to-fine localization paradigm: we first perform a global retrieval to obtain location hypotheses and only later match local features within those candidate places. This hierarchical approach incurs significant runtime savings and makes our system suitable for real-time operation. By leveraging learned descriptors, our method achieves remarkable localization robustness across large variations of appearance. Consequently, we demonstrate new state-of-the-art performance on two challenging benchmarks for large-scale 6-DoF localization. The code of our method will be made publicly available.
31.Adversarial Sparse-View CBCT Artifact Reduction pdf
We present an effective post-processing method to reduce the artifacts from sparsely reconstructed cone-beam CT (CBCT) images. The proposed method is based on the state-of-the-art, image-to-image generative models with a perceptual loss as regulation. Unlike the traditional CT artifact-reduction approaches, our method is trained in an adversarial fashion that yields more perceptually realistic outputs while preserving the anatomical structures. To address the streak artifacts that are inherently local and appear across various scales, we further propose a novel discriminator architecture based on feature pyramid networks and a differentially modulated focus map to induce the adversarial training. Our experimental results show that the proposed method can greatly correct the cone-beam artifacts from clinical CBCT images reconstructed using 1/3 projections, and outperforms strong baseline methods both quantitatively and qualitatively.
32.Joint Vertebrae Identification and Localization in Spinal CT Images by Combining Short- and Long-Range Contextual Information pdf
Automatic vertebrae identification and localization from arbitrary CT images is challenging. Vertebrae usually share similar morphological appearance. Because of pathology and the arbitrary field-of-view of CT scans, one can hardly rely on the existence of some anchor vertebrae or parametric methods to model the appearance and shape. To solve the problem, we argue that one should make use of the short-range contextual information, such as the presence of some nearby organs (if any), to roughly estimate the target vertebrae; due to the unique anatomic structure of the spine column, vertebrae have fixed sequential order which provides the important long-range contextual information to further calibrate the results.
We propose a robust and efficient vertebrae identification and localization system that can inherently learn to incorporate both the short-range and long-range contextual information in a supervised manner. To this end, we develop a multi-task 3D fully convolutional neural network (3D FCN) to effectively extract the short-range contextual information around the target vertebrae. For the long-range contextual information, we propose a multi-task bidirectional recurrent neural network (Bi-RNN) to encode the spatial and contextual information among the vertebrae of the visible spine column. We demonstrate the effectiveness of the proposed approach on a challenging dataset and the experimental results show that our approach outperforms the state-of-the-art methods by a significant margin.
33.Comixify: Transform video into a comics pdf
In this paper, we propose a solution to transform a video into a comics. We approach this task using a neural style algorithm based on Generative Adversarial Networks (GANs). Several recent works in the field of Neural Style Transfer showed that producing an image in the style of another image is feasible. In this paper, we build up on these works and extend the existing set of style transfer use cases with a working application of video comixification. To that end, we train an end-to-end solution that transforms input video into a comics in two stages. In the first stage, we propose a state-of-the-art keyframes extraction algorithm that selects a subset of frames from the video to provide the most comprehensive video context and we filter those frames using image aesthetic estimation engine. In the second stage, the style of selected keyframes is transferred into a comics. To provide the most aesthetically compelling results, we selected the most state-of-the art style transfer solution and based on that implement our own ComixGAN framework. The final contribution of our work is a Web-based working application of video comixification available at this http URL.
34.A Comparison of Embedded Deep Learning Methods for Person Detection pdf
Recent advancements in parallel computing, GPU technology and deep learning provide a new platform for complex image processing tasks such as person detection to flourish. Person detection is fundamental preliminary operation for several high level computer vision tasks. One industry that can significantly benefit from person detection is retail. In recent years, various studies attempt to find an optimal solution for person detection using neural networks and deep learning. This study conducts a comparison among the state of the art deep learning base object detector with the focus on person detection performance in indoor environments. Performance of various implementations of YOLO, SSD, RCNN, R-FCN and SqueezeDet have been assessed using our in-house proprietary dataset which consists of over 10 thousands indoor images captured form shopping malls, retails and stores. Experimental results indicate that, Tiny YOLO-416 and SSD (VGG-300) are the fastest and Faster-RCNN (Inception ResNet-v2) and R-FCN (ResNet-101) are the most accurate detectors investigated in this study. Further analysis shows that YOLO v3-416 delivers relatively accurate result in a reasonable amount of time, which makes it an ideal model for person detection in embedded platforms.
35.FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search pdf
Designing accurate and efficient ConvNets for mobile devices is challenging because the design space is combinatorially large. Due to this, previous neural architecture search (NAS) methods are computationally expensive. ConvNet architecture optimality depends on factors such as input resolution and target devices. However, existing approaches are too expensive for case-by-case redesigns. Also, previous work focuses primarily on reducing FLOPs, but FLOP count does not always reflect actual latency. To address these, we propose a differentiable neural architecture search (DNAS) framework that uses gradient-based methods to optimize ConvNet architectures, avoiding enumerating and training individual architectures separately as in previous methods. FBNets, a family of models discovered by DNAS surpass state-of-the-art models both designed manually and generated automatically. FBNet-B achieves 74.1% top-1 accuracy on ImageNet with 295M FLOPs and 23.1 ms latency on a Samsung S8 phone, 2.4x smaller and 1.5x faster than MobileNetV2-1.3 with similar accuracy. Despite higher accuracy and lower latency than MnasNet, we estimate FBNet-B's search cost is 420x smaller than MnasNet's, at only 216 GPU-hours. Searched for different resolutions and channel sizes, FBNets achieve 1.5% to 6.4% higher accuracy than MobileNetV2. The smallest FBNet achieves 50.2% accuracy and 2.9 ms latency (345 frames per second) on a Samsung S8. Over a Samsung-optimized FBNet, the iPhone-X-optimized model achieves a 1.4x speedup on an iPhone X.
36.Real-Time Referring Expression Comprehension by Single-Stage Grounding Network pdf
In this paper, we propose a novel end-to-end model, namely Single-Stage Grounding network (SSG), to localize the referent given a referring expression within an image. Different from previous multi-stage models which rely on object proposals or detected regions, our proposed model aims to comprehend a referring expression through one single stage without resorting to region proposals as well as the subsequent region-wise feature extraction. Specifically, a multimodal interactor is proposed to summarize the local region features regarding the referring expression attentively. Subsequently, a grounder is proposed to localize the referring expression within the given image directly. For further improving the localization accuracy, a guided attention mechanism is proposed to enforce the grounder to focus on the central region of the referent. Moreover, by exploiting and predicting visual attribute information, the grounder can further distinguish the referent objects within an image and thereby improve the model performance. Experiments on RefCOCO, RefCOCO+, and RefCOCOg datasets demonstrate that our proposed SSG without relying on any region proposals can achieve comparable performance with other advanced models. Furthermore, our SSG outperforms the previous models and achieves the state-of-art performance on the ReferItGame dataset. More importantly, our SSG is time efficient and can ground a referring expression in a 416416 image from the RefCOCO dataset in 25ms (40 referents per second) on average with a Nvidia Tesla P40, accomplishing more than 9 speedups over the existing multi-stage models.
37.Learning Transferable Adversarial Examples via Ghost Networks pdf
The recent development of adversarial attack has proven that ensemble-based methods can perform black-box attack better than the traditional, non-ensemble ones. However, those methods generally suffer from high complexity. They require a family of diverse models, and ensembling them afterward, both of which are computationally expensive.
In this paper, we propose Ghost Networks to efficiently learn transferable adversarial examples. The key principle of ghost networks is to perturb an existing model, which potentially generates a huge set of diverse models. Those models are subsequently fused by longitudinal ensemble. Both steps almost require no extra time and space consumption.
Extensive experimental results suggest that the number of networks is essential for improving the transferability of adversarial examples, but it is less necessary to independently train different networks and then ensemble them in an intensive aggregation way. Instead, our work can be a computationally cheap plug-in, which can be easily applied to improve adversarial approaches both in single-model attack and multi-model attack, compatible with both residual and non-residual networks. In particular, by re-producing the NIPS 2017 adversarial competition, our work outperforms the No.1 attack submission by a large margin, which demonstrates its effectiveness and efficiency.
38.Feature Denoising for Improving Adversarial Robustness pdf
Adversarial attacks to image classification systems present challenges to convolutional networks and opportunities for understanding them. This study suggests that adversarial perturbations on images lead to noise in the features constructed by these networks. Motivated by this observation, we develop new network architectures that increase adversarial robustness by performing feature denoising. Specifically, our networks contain blocks that denoise the features using non-local means or other filters; the entire networks are trained end-to-end. When combined with adversarial training, our feature denoising networks substantially improve the state-of-the-art in adversarial robustness in both white-box and black-box attack settings. On ImageNet, under 10-iteration PGD white-box attacks where prior art has 27.9% accuracy, our method achieves 55.7%; even under extreme 2000-iteration PGD white-box attacks, our method secures 42.6% accuracy. A network based on our method was ranked first in Competition on Adversarial Attacks and Defenses (CAAD) 2018 --- it achieved 50.6% classification accuracy on a secret, ImageNet-like test dataset against 48 unknown attackers, surpassing the runner-up approach by ~10%. Code and models will be made publicly available.
39.Beyond Domain Adaptation: Unseen Domain Encapsulation via Universal Non-volume Preserving Models pdf
Recognition across domains has recently become an active topic in the research community. However, it has been largely overlooked in the problem of recognition in new unseen domains. Under this condition, the delivered deep network models are unable to be updated, adapted or fine-tuned. Therefore, recent deep learning techniques, such as: domain adaptation, feature transferring, and fine-tuning, cannot be applied. This paper presents a novel Universal Non-volume Preserving approach to the problem of domain generalization in the context of deep learning. The proposed method can be easily incorporated with any other ConvNet framework within an end-to-end deep network design to improve the performance. On digit recognition, we benchmark on four popular digit recognition databases, i.e. MNIST, USPS, SVHN and MNIST-M. The proposed method is also experimented on face recognition on Extended Yale-B, CMU-PIE and CMU-MPIE databases and compared against other the state-of-the-art methods. In the problem of pedestrian detection, we empirically observe that the proposed method learns models that improve performance across a priori unknown data distributions.
40.Semantically-Aware Attentive Neural Embeddings for Image-based Visual Localization pdf
We present a novel method for fusing appearance and semantic information using visual attention for 2D image-based localization (2D-VL) across extreme changes in viewing conditions. Our deep learning based method is motivated by the intuition that specific scene regions remain stable in the semantic modality even in the presence of vast differences in the appearance modality. The proposed attention-based module learns to focus not only on discriminative visual regions for place recognition but also on consistently stable semantic regions to perform 2D-VL. We show the effectiveness of this model by comparing against state-of-the-art (SOTA) methods on several challenging localization datasets. We report an average absolute improvement of 19% over current SOTA 2D-VL methods. Furthermore, we present an extensive study demonstrating the effectiveness and contribution of each component of our model, showing 8%-15% absolute improvement from adding semantic information, and an additional 4% from our proposed attention module, over both prior methods as well as a competitive baseline.
41.Biometric Recognition System (Algorithm) pdf
Fingerprints are the most widely deployed form of biometric identification. No two individuals share the same fingerprint because they have unique biometric identifiers. This paper presents an efficient fingerprint verification algorithm which improves matching accuracy. Fingerprint images get degraded and corrupted due to variations in skin and impression conditions. Thus, image enhancement techniques are employed prior to singular point detection and minutiae extraction. Singular point is the point of maximum curvature. It is determined by the normal of each fingerprint ridge, and then following them inward towards the centre. The local ridge features known as minutiae is extracted using cross-number method to find ridge endings and ridge bifurcations. The proposed algorithm chooses a radius and draws a circle with core point as centre, making fingerprint images rotationally invariant and uniform. The radius can be varied according to the accuracy depending on the particular application. Morphological techniques such as clean, spur and H-break is employed to remove noise, followed by removing spurious minutiae. Templates are created based on feature vector extraction and databases are made for verification and identification for the fingerprint images taken from Fingerprint Verification Competition (FVC2002). Minimum Euclidean distance is calculated between saved template and the test fingerprint image template and compared with the set threshold for matching decision. For the performance evaluation of the proposed algorithm various measures, equal error rate (EER), Dmin at EER, accuracy and threshold are evaluated and plotted. The measures demonstrate that the proposed algorithm is more effective and robust.
42.Unsupervised Learning of Monocular Depth Estimation with Bundle Adjustment, Super-Resolution and Clip Loss pdf
We present a novel unsupervised learning framework for single view depth estimation using monocular videos. It is well known in 3D vision that enlarging the baseline can increase the depth estimation accuracy, and jointly optimizing a set of camera poses and landmarks is essential. In previous monocular unsupervised learning frameworks, only part of the photometric and geometric constraints within a sequence are used as supervisory signals. This may result in a short baseline and overfitting. Besides, previous works generally estimate a low resolution depth from a low resolution impute image. The low resolution depth is then interpolated to recover the original resolution. This strategy may generate large errors on object boundaries, as the depth of background and foreground are mixed to yield the high resolution depth. In this paper, we introduce a bundle adjustment framework and a super-resolution network to solve the above two problems. In bundle adjustment, depths and poses of an image sequence are jointly optimized, which increases the baseline by establishing the relationship between farther frames. The super resolution network learns to estimate a high resolution depth from a low resolution image. Additionally, we introduce the clip loss to deal with moving objects and occlusion. Experimental results on the KITTI dataset show that the proposed algorithm outperforms the state-of-the-art unsupervised methods using monocular sequences, and achieves comparable or even better result compared to unsupervised methods using stereo sequences.
43.GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloud pdf
We introduce a novel 3D object proposal approach named Generative Shape Proposal Network (GSPN) for instance segmentation in point cloud data. Instead of treating object proposal as a direct bounding box regression problem, we take an analysis-by-synthesis strategy and generate proposals by reconstructing shapes from noisy observations in a scene. We incorporate GSPN into a novel 3D instance segmentation framework named Region-based PointNet (R-PointNet) which allows flexible proposal refinement and instance segmentation generation. We achieve state-of-the-art performance on several 3D instance segmentation tasks. The success of GSPN largely comes from its emphasis on geometric understandings during object proposal, which greatly reducing proposals with low objectness.
44.Detecting Adversarial Examples in Convolutional Neural Networks pdf
The great success of convolutional neural networks has caused a massive spread of the use of such models in a large variety of Computer Vision applications. However, these models are vulnerable to certain inputs, the adversarial examples, which although are not easily perceived by humans, they can lead a neural network to produce faulty results. This paper focuses on the detection of adversarial examples, which are created for convolutional neural networks that perform image classification. We propose three methods for detecting possible adversarial examples and after we analyze and compare their performance, we combine their best aspects to develop an even more robust approach. The first proposed method is based on the regularization of the feature vector that the neural network produces as output. The second method detects adversarial examples by using histograms, which are created from the outputs of the hidden layers of the neural network. These histograms create a feature vector which is used as the input of an SVM classifier, which classifies the original input either as an adversarial or as a real input. Finally, for the third method we introduce the concept of the residual image, which contains information about the parts of the input pattern that are ignored by the neural network. This method aims at the detection of possible adversarial examples, by using the residual image and reinforcing the parts of the input pattern that are ignored by the neural network. Each one of these methods has some novelties and by combining them we can further improve the detection results. For the proposed methods and their combination, we present the results of detecting adversarial examples on the MNIST dataset. The combination of the proposed methods offers some improvements over similar state of the art approaches.
45.Explainability by Parsing: Neural Module Tree Networks for Natural Language Visual Grounding pdf
Grounding natural language in images essentially requires composite visual reasoning. However, existing methods over-simplify the composite nature of language into a monolithic sentence embedding or a coarse composition of subject-predicate-object triplet. They might perform well on short phrases, but generally fail in longer sentences, mainly due to the over-fitting to certain vision-language bias. In this paper, we propose to ground natural language in an intuitive, explainable, and composite fashion as it should be. In particular, we develop a novel modular network called Neural Module Tree network (NMTree) that regularizes the visual grounding along the dependency parsing tree of the sentence, where each node is a module network that calculates or accumulates the grounding score in a bottom-up direction where as needed. NMTree disentangles the visual grounding from the composite reasoning, allowing the former to only focus on primitive and easy-to-generalize patterns. To reduce the impact of parsing errors, we train the modules and their assembly end-to-end by using the Gumbel-Softmax approximation and its straight-through gradient estimator, accounting for the discrete process of module selection. Overall, the proposed NMTree not only consistently outperforms the state-of-the-arts on several benchmarks and tasks, but also shows explainable reasoning in grounding score calculation. Therefore, NMTree shows a good direction in closing the gap between explainability and performance.
46.Attend More Times for Image Captioning pdf
Most attention-based image captioning models attend to the image once per word. However, attending once per word is rigid and is easy to miss some information. Attending more times can adjust the attention position, find the missing information back and avoid generating the wrong word. In this paper, we show that attending more times per word can gain improvements in the image captioning task. We propose a flexible two-LSTM merge model to make it convenient to encode more attentions than words. Our captioning model uses two LSTMs to encode the word sequence and the attention sequence respectively. The information of the two LSTMs and the image feature are combined to predict the next word. Experiments on the MSCOCO caption dataset show that our method outperforms the state-of-the-art. Using bottom up features and self-critical training method, our method gets BLEU-4, METEOR, ROUGE-L and CIDEr scores of 0.381, 0.283, 0.580 and 1.261 on the Karpathy test split.
47.Spatial-Temporal Person Re-identification pdf
Most of current person re-identification (ReID) methods neglect a spatial-temporal constraint. Given a query image, conventional methods compute the feature distances between the query image and all the gallery images and return a similarity ranked table. When the gallery database is very large in practice, these approaches fail to obtain a good performance due to appearance ambiguity across different camera views. In this paper, we propose a novel two-stream spatial-temporal person ReID (st-ReID) framework that mines both visual semantic information and spatial-temporal information. To this end, a joint similarity metric with Logistic Smoothing (LS) is introduced to integrate two kinds of heterogeneous information into a unified framework. To approximate a complex spatial-temporal probability distribution, we develop a fast Histogram-Parzen (HP) method. With the help of the spatial-temporal constraint, the st-ReID model eliminates lots of irrelevant images and thus narrows the gallery database. Without bells and whistles, our st-ReID method achieves rank-1 accuracy of 98.1% on Market-1501 and 94.4% on DukeMTMC-reID, improving from the baselines 91.2% and 83.8%, respectively, outperforming all previous state-of-the-art methods by a large margin.
48.SANTIS: Sampling-Augmented Neural neTwork with Incoherent Structure for MR image reconstruction pdf
Deep learning holds great promise in the reconstruction of undersampled Magnetic Resonance Imaging (MRI) data, providing new opportunities to escalate the performance of rapid MRI. In existing deep learning-based reconstruction methods, supervised training is performed using artifact-free reference images and their corresponding undersampled pairs. The undersampled images are generated by a fixed undersampling pattern in the training, and the trained network is then applied to reconstruct new images acquired with the same pattern in the inference. While such a training strategy can maintain a favorable reconstruction for a pre-selected undersampling pattern, the robustness of the trained network against any discrepancy of undersampling schemes is typically poor. We developed a novel deep learning-based reconstruction framework called SANTIS for efficient MR image reconstruction with improved robustness against sampling pattern discrepancy. SANTIS uses a data cycle-consistent adversarial network combining efficient end-to-end convolutional neural network mapping, data fidelity enforcement and adversarial training for reconstructing accelerated MR images more faithfully. A training strategy employing sampling augmentation with extensive variation of undersampling patterns was further introduced to promote the robustness of the trained network. Compared to conventional reconstruction and standard deep learning methods, SANTIS achieved consistent better reconstruction performance, with lower errors, greater image sharpness and higher similarity with respect to the reference regardless of the undersampling patterns during inference. This novel concept behind SANTIS can particularly be useful towards improving the robustness of deep learning-based image reconstruction against discrepancy between training and evaluation, which is currently an important but less studied open question.
49.Neural Abstract Style Transfer for Chinese Traditional Painting pdf
Chinese traditional painting is one of the most historical artworks in the world. It is very popular in Eastern and Southeast Asia due to being aesthetically appealing. Compared with western artistic painting, it is usually more visually abstract and textureless. Recently, neural network based style transfer methods have shown promising and appealing results which are mainly focused on western painting. It remains a challenging problem to preserve abstraction in neural style transfer. In this paper, we present a Neural Abstract Style Transfer method for Chinese traditional painting. It learns to preserve abstraction and other style jointly end-to-end via a novel MXDoG-guided filter (Modified version of the eXtended Difference-of-Gaussians) and three fully differentiable loss terms. To the best of our knowledge, there is little work study on neural style transfer of Chinese traditional painting. To promote research on this direction, we collect a new dataset with diverse photo-realistic images and Chinese traditional paintings. In experiments, the proposed method shows more appealing stylized results in transferring the style of Chinese traditional painting than state-of-the-art neural style transfer methods.
50.Face Completion with Semantic Knowledge and Collaborative Adversarial Learning pdf
Unlike a conventional background inpainting approach that infers a missing area from image patches similar to the background, face completion requires semantic knowledge about the target object for realistic outputs. Current image inpainting approaches utilize generative adversarial networks (GANs) to achieve such semantic understanding. However, in adversarial learning, the semantic knowledge is learned implicitly and hence good semantic understanding is not always guaranteed. In this work, we propose a collaborative adversarial learning approach to face completion to explicitly induce the training process. Our method is formulated under a novel generative framework called collaborative GAN (collaGAN), which allows better semantic understanding of a target object through collaborative learning of multiple tasks including face completion, landmark detection, and semantic segmentation. Together with the collaGAN, we also introduce an inpainting concentrated scheme such that the model emphasizes more on inpainting instead of autoencoding. Extensive experiments show that the proposed designs are indeed effective and collaborative adversarial learning provides better feature representations of the faces. In comparison with other generative image inpainting models and single task learning methods, our solution produces superior performances on all tasks.
51.Deep ChArUco: Dark ChArUco Marker Pose Estimation pdf
ChArUco boards are used for camera calibration, monocular pose estimation, and pose verification in both robotics and augmented reality. Such fiducials are detectable via traditional computer vision methods (as found in OpenCV) in well-lit environments, but classical methods fail when the lighting is poor or when the image undergoes extreme motion blur. We present Deep ChArUco, a real-time pose estimation system which combines two custom deep networks, ChArUcoNet and RefineNet, with the Perspective-n-Point (PnP) algorithm to estimate the marker's 6DoF pose. ChArUcoNet is a two-headed marker-specific convolutional neural network (CNN) which jointly outputs ID-specific classifiers and 2D point locations. The 2D point locations are further refined into subpixel coordinates using RefineNet. Our networks are trained using a combination of auto-labeled videos of the target marker, synthetic subpixel corner data, and extreme data augmentation. We evaluate Deep ChArUco in challenging low-light, high-motion, high-blur scenarios and demonstrate that our approach is superior to a traditional OpenCV-based method for ChArUco marker detection and pose estimation.
52.Self-Improving Visual Odometry pdf
We propose a self-supervised learning framework that uses unlabeled monocular video sequences to generate large-scale supervision for training a Visual Odometry (VO) frontend, a network which computes pointwise data associations across images. Our self-improving method enables a VO frontend to learn over time, unlike other VO and SLAM systems which require time-consuming hand-tuning or expensive data collection to adapt to new environments. Our proposed frontend operates on monocular images and consists of a single multi-task convolutional neural network which outputs 2D keypoints locations, keypoint descriptors, and a novel point stability score. We use the output of VO to create a self-supervised dataset of point correspondences to retrain the frontend. When trained using VO at scale on 2.5 million monocular images from ScanNet, the stability classifier automatically discovers a ranking for keypoints that are not likely to help in VO, such as t-junctions across depth discontinuities, features on shadows and highlights, and dynamic objects like people. The resulting frontend outperforms both traditional methods (SIFT, ORB, AKAZE) and deep learning methods (SuperPoint and LF-Net) in a 3D-to-2D pose estimation task on ScanNet.
53.PIRC Net : Using Proposal Indexing, Relationships and Context for Phrase Grounding pdf
Phrase Grounding aims to detect and localize objects in images that are referred to and are queried by natural language phrases. Phrase grounding finds applications in tasks such as Visual Dialog, Visual Search and Image-text co-reference resolution. In this paper, we present a framework that leverages information such as phrase category, relationships among neighboring phrases in a sentence and context to improve the performance of phrase grounding systems. We propose three modules: Proposal Indexing Network(PIN); Inter-phrase Regression Network(IRN) and Proposal Ranking Network(PRN) each of which analyze the region proposals of an image at increasing levels of detail by incorporating the above information. Also, in the absence of ground-truth spatial locations of the phrases(weakly-supervised), we propose knowledge transfer mechanisms that leverages the framework of PIN module. We demonstrate the effectiveness of our approach on the Flickr 30k Entities and ReferItGame datasets, for which we achieve improvements over state-of-the-art approaches in both supervised and weakly-supervised variants.
54.Harmonic Networks: Integrating Spectral Information into CNNs pdf
Convolutional neural networks (CNNs) learn filters in order to capture local correlation patterns in feature space. In contrast, in this paper we propose harmonic blocks that produce features by learning optimal combinations of spectral filters defined by the Discrete Cosine Transform. The harmonic blocks are used to replace conventional convolutional layers to construct partial or fully harmonic CNNs. We extensively validate our approach and show that the introduction of harmonic blocks into state-of-the-art CNN baseline architectures results in comparable or better performance in classification tasks on small NORB, CIFAR10 and CIFAR100 datasets.
55.Removal of Parameter Adjustment of Frangi Filters in Case of Coronary Angiograms pdf
Frangi Filters are one of the widely used filters for enhancing vessels in medical images. Since they were first proposed, the threshold of the vesselness function of Frangi Filters is to be arranged for each individual application. These thresholds are changed manually for individual fluoroscope, for enhancing coronary angiogram images. Hence it is felt, there is a need of mitigating the tuning procedure of threshold values for every fluoroscope. The current papers approach has been devised in order to treat the coronary angiogram images uniformly, irrespective of the fluoroscopes through which they were obtained and the patient demographics for further stenosis detection. This problem to the best of our knowledge has not been addressed yet. In the approach, before feeding the image to Frangi Filters, non uniform illumination of the input image is removed using homomorphic filters and the image is enhanced using Non Subsampled Contourlet Transform (NSCT). The experiment was conducted on the data that has been accumulated from various hospitals in India and the results obtained verifies dependency removal of parameters without compromising the results obtained by Frangi filters.
56.An Intelligent Safety System for Human-Centered Semi-Autonomous Vehicles pdf
Nowadays, automobile manufacturers make efforts to develop ways to make cars fully safe. Monitoring driver's actions by computer vision techniques to detect driving mistakes in real-time and then planning for autonomous driving to avoid vehicle collisions is one of the most important issues that has been investigated in the machine vision and Intelligent Transportation Systems (ITS). The main goal of this study is to prevent accidents caused by fatigue, drowsiness, and driver distraction. To avoid these incidents, this paper proposes an integrated safety system that continuously monitors the driver's attention and vehicle surroundings, and finally decides whether the actual steering control status is safe or not. For this purpose, we equipped an ordinary car called FARAZ with a vision system consisting of four mounted cameras along with a universal car tool for communicating with surrounding factory-installed sensors and other car systems, and sending commands to actuators. The proposed system leverages a scene understanding pipeline using deep convolutional encode-decoder network and a driver state detection pipeline. We have been identifying and assessing domestic capabilities for the development of technologies specifically of the ordinary vehicles in order to manufacture smart cars and eke providing an intelligent system to increase safety and to assist the driver in various conditions/situations.
57.Learning Representations of Sets through Optimized Permutations pdf
Representations of sets are challenging to learn because operations on sets should be permutation-invariant. To this end, we propose a Permutation-Optimisation module that learns how to permute a set end-to-end. The permuted set can be further processed to learn a permutation-invariant representation of that set, avoiding a bottleneck in traditional set models. We demonstrate our model's ability to learn permutations and set representations with either explicit or implicit supervision on four datasets, on which we achieve state-of-the-art results: number sorting, image mosaics, classification from image mosaics, and visual question answering.
58.Hierarchical Bipartite Graph Convolution Networks pdf
Recently, graph neural networks have been adopted in a wide variety of applications ranging from relational representations to modeling irregular data domains such as point clouds and social graphs. However, the space of graph neural network architectures remains highly fragmented impeding the development of optimized implementations similar to what is available for convolutional neural networks. In this work, we present BiGraphNet, a graph neural network architecture that generalizes many popular graph neural network models and enables new efficient operations similar to those supported by ConvNets. By explicitly separating the input and output nodes, BiGraphNet: (i) generalizes the graph convolution to support new efficient operations such as coarsened graph convolutions (similar to strided convolution in convnets), multiple input graphs convolution and graph expansions (unpooling) which can be used to implement various graph architectures such as graph autoencoders, and graph residual nets; and (ii) accelerates and scales the computations and memory requirements in hierarchical networks by performing computations only at specified output nodes.
59.Unsupervised Deep Learning for Structured Shape Matching pdf
We present a novel method for computing correspondences across shapes using unsupervised learning. Our method allows to compute a non-linear transformation of given descriptor functions, while optimizing for global structural properties of the resulting maps, such as their bijectivity or approximate isometry. To this end, we use the functional maps framework, and build upon the recently proposed FMNet architecture for descriptor learning. Unlike the method proposed in that work, however, we show that learning can be done in a purely unsupervised setting, without having access to any ground truth correspondences. This results in a very general shape matching method, which can be used to establish correspondences within shape collections or even just a single shape pair, without any prior information. We demonstrate on a wide range of challenging benchmarks, that our method leads to significant improvement compared to the existing axiomatic methods and achieves comparable, and in some cases superior results to even the supervised learning techniques.
60.Learning Embedding Adaptation for Few-Shot Learning pdf
Learning with limited data is a key challenge for visual recognition. Few-shot learning methods address this challenge by learning an instance embedding function from seen classes and apply the function to instances from unseen classes with limited labels. This style of transfer learning is task-agnostic: the embedding function is not learned optimally discriminative with respect to the unseen classes, where discerning among them is the target task. In this paper, we propose a novel approach to adapt the embedding model to the target classification task, yielding embeddings that are task-specific and are discriminative. To this end, we employ a type of self-attention mechanism called Transformer to transform the embeddings from task-agnostic to task-specific by focusing on relating instances from the test instances to the training instances in both seen and unseen classes. Our approach also extends to both transductive and generalized few-shot classification, two important settings that have essential use cases. We verify the effectiveness of our model on two standard benchmark few-shot classification datasets --- MiniImageNet and CUB, where our approach demonstrates state-of-the-art empirical performance.
61.A New Variational Model for Joint Image Reconstruction and Motion Estimation in Spatiotemporal Imaging pdf
We propose a new variational model for joint image reconstruction and motion estimation in spatiotemporal imaging, which is investigated along a general framework that we present with shape theory. This model consists of two components, one for conducting modified static image reconstruction, and the other performs sequentially indirect image registration. For the latter, we generalize the large deformation diffeomorphic metric mapping framework into the sequentially indirect registration setting. The proposed model is compared theoretically against alternative approaches (optical flow based model and diffeomorphic motion models), and we demonstrate that the proposed model has desirable properties in terms of the optimal solution. The theoretical derivations and efficient algorithms are also presented for a time-discretized scenario of the proposed model, which show that the optimal solution of the time-discretized version is consistent with that of the time-continuous one, and most of the computational components is the easy-implemented linearized deformation. The complexity of the algorithm is analyzed as well. This work is concluded by some numerical examples in 2D space + time tomography with very sparse and/or highly noisy data.
62.Area-preserving mapping of 3D ultrasound carotid artery images using density-equalizing reference map pdf
Carotid atherosclerosis is a focal disease at the bifurcations of the carotid artery. To quantitatively monitor the local changes in the vessel-wall-plus-plaque thickness (VWT) and compare the VWT distributions for different patients or for the same patients at different ultrasound scanning sessions, a mapping technique is required to adjust for the geometric variability of different carotid artery models. In this work, we propose a novel method called density-equalizing reference map (DERM) for mapping 3D carotid surfaces to a standardized 2D carotid template, with an emphasis on preserving the local geometry of the carotid surface by minimizing the local area distortion. The initial map was generated by a previously described arc-length scaling (ALS) mapping method, which projects a 3D carotid surface onto a 2D non-convex L-shaped domain. A smooth and area-preserving flattened map was subsequently constructed by deforming the ALS map using the proposed algorithm that combines the density-equalizing map and the reference map techniques. This combination allows, for the first time, one-to-one mapping from a 3D surface to a standardized non-convex planar domain in an area-preserving manner. Evaluations using 20 carotid surface models show that the proposed method reduced the area distortion of the flattening maps by over 80% as compared to the ALS mapping method.
63.AutoGAN: Robust Classifier Against Adversarial Attacks pdf
Classifiers fail to classify correctly input images that have been purposefully and imperceptibly perturbed to cause misclassification. This susceptability has been shown to be consistent across classifiers, regardless of their type, architecture or parameters. Common defenses against adversarial attacks modify the classifer boundary by training on additional adversarial examples created in various ways. In this paper, we introduce AutoGAN, which counters adversarial attacks by enhancing the lower-dimensional manifold defined by the training data and by projecting perturbed data points onto it. AutoGAN mitigates the need for knowing the attack type and magnitude as well as the need for having adversarial samples of the attack. Our approach uses a Generative Adversarial Network (GAN) with an autoencoder generator and a discriminator that also serves as a classifier. We test AutoGAN against adversarial samples generated with state-of-the-art Fast Gradient Sign Method (FGSM) as well as samples generated with random Gaussian noise, both using the MNIST dataset. For different magnitudes of perturbation in training and testing, AutoGAN can surpass the accuracy of FGSM method by up to 25% points on samples perturbed using FGSM. Without an augmented training dataset, AutoGAN achieves an accuracy of 89% compared to 1% achieved by FGSM method on FGSM testing adversarial samples.