1.Precision Gating: Improving Neural Network Efficiency with Dynamic Dual-Precision Activations ⬇️
We propose precision gating (PG), an end-to-end trainable dynamic dual-precision quantization technique for deep neural networks. PG computes most features in a low precision and only a small proportion of important features in a higher precision to preserve accuracy. The proposed approach is applicable to a variety of DNN architectures and significantly reduces the computational cost of DNN execution with almost no accuracy loss. Our experiments indicate that PG achieves excellent results on CNNs, including statically compressed mobile-friendly networks such as ShuffleNet. Compared to the state-of-the-art prediction-based quantization schemes, PG achieves the same or higher accuracy with 2.4$\times$ less compute on ImageNet. PG furthermore applies to RNNs. Compared to 8-bit uniform quantization, PG obtains a 1.2% improvement in perplexity per word with 2.7$\times$ computational cost reduction on LSTM on the Penn Tree Bank dataset.
2.Lake Ice Detection from Sentinel-1 SAR with Deep Learning ⬇️
Lake ice, as part of the Essential Climate Variable (ECV) lakes, is an important indicator to monitor climate change and global warming. The spatio-temporal extent of lake ice cover, along with the timings of key phenological events such as freeze-up and break-up, provides important cues about the local and global climate. We present a lake ice monitoring system based on the automatic analysis of Sentinel-1 Synthetic Aperture Radar (SAR) data with a deep neural network. In previous studies that used optical satellite imagery for lake ice monitoring, frequent cloud cover was a main limiting factor, which we overcome thanks to the ability of microwave sensors to penetrate clouds and observe the lakes regardless of the weather and illumination conditions. We cast ice detection as a two class (frozen, non-frozen) semantic segmentation problem and solve it using a state-of-the-art deep convolutional network (CNN). We report results on two winters (
$2016-17$ and$2017-18$ ) and three alpine lakes in Switzerland, including cross-validation tests to assess the generalisation to unseen lakes and winters. The proposed model reaches mean Intersection-over-Union (mIoU) scores >90% on average, and >84% even for the most difficult lake.
3.Learning Architectures for Binary Networks ⬇️
Backbone architectures of most binary networks are well-known floating point architectures, such as the ResNet family. Questioning that the architectures designed for floating-point networks would not be the best for binary networks, we propose to search architectures for binary networks (BNAS). Specifically, based on the cell based search method, we define a new set of layer types, design a new cell template, and rediscover the utility of and propose to use the Zeroise layer to learn well-performing binary networks. In addition, we propose to diversify early search to learn better performing binary architectures. We show that our searched binary networks outperform state-of-the-art binary networks on CIFAR10 and ImageNet datasets.
4.Patient-Specific Finetuning of Deep Learning Models for Adaptive Radiotherapy in Prostate CT ⬇️
Contouring of the target volume and Organs-At-Risk (OARs) is a crucial step in radiotherapy treatment planning. In an adaptive radiotherapy setting, updated contours need to be generated based on daily imaging. In this work, we leverage personalized anatomical knowledge accumulated over the treatment sessions, to improve the segmentation accuracy of a pre-trained Convolution Neural Network (CNN), for a specific patient. We investigate a transfer learning approach, fine-tuning the baseline CNN model to a specific patient, based on imaging acquired in earlier treatment fractions. The baseline CNN model is trained on a prostate CT dataset from one hospital of 379 patients. This model is then fine-tuned and tested on an independent dataset of another hospital of 18 patients, each having 7 to 10 daily CT scans. For the prostate, seminal vesicles, bladder and rectum, the model fine-tuned on each specific patient achieved a Mean Surface Distance (MSD) of
$1.64 \pm 0.43$ mm,$2.38 \pm 2.76$ mm,$2.30 \pm 0.96$ mm, and$1.24 \pm 0.89$ mm, respectively, which was significantly better than the baseline model. The proposed personalized model adaptation is therefore very promising for clinical implementation in the context of adaptive radiotherapy of prostate cancer.
5.Amplifying The Uncanny ⬇️
Deep neural networks have become remarkably good at producing realistic deepfakes, images of people that are (to the untrained eye) indistinguishable from real images. These are produced by algorithms that learn to distinguish between real and fake images and are optimised to generate samples that the system deems realistic. This paper, and the resulting series of artworks Being Foiled explore the aesthetic outcome of inverting this process and instead optimising the system to generate images that it sees as being fake. Maximising the unlikelihood of the data and in turn, amplifying the uncanny nature of these machine hallucinations.
6.Hierarchical Rule Induction Network for Abstract Visual Reasoning ⬇️
Abstract reasoning refers to the ability to analyze information, discover rules at an intangible level, and solve problems in innovative ways. Raven's Progressive Matrices (RPM) test is typically used to examine the capability of abstract reasoning. In the test, the subject is asked to identify the correct choice from the answer set to fill the missing panel at the bottom right of RPM (e.g., a 3$\times$3 matrix), following the underlying rules inside the matrix. Recent studies, taking advantage of Convolutional Neural Networks (CNNs), have achieved encouraging progress to accomplish the RPM test problems. Unfortunately, simply relying on the relation extraction at the matrix level, they fail to recognize the complex attribute patterns inside or across rows/columns of RPM. To address this problem, in this paper we propose a Hierarchical Rule Induction Network (HriNet), by intimating human induction strategies. HriNet extracts multiple granularity rule embeddings at different levels and integrates them through a gated embedding fusion module. We further introduce a rule similarity metric based on the embeddings, so that HriNet can not only be trained using a tuplet loss but also infer the best answer according to the similarity score. To comprehensively evaluate HriNet, we first fix the defects contained in the very recent RAVEN dataset and generate a new one named Balanced-RAVEN. Then extensive experiments are conducted on the large-scale dataset PGM and our Balanced-RAVEN, the results of which show that HriNet outperforms the state-of-the-art models by a large margin.
7.DeepDualMapper: A Gated Fusion Network for Automatic Map Extraction using Aerial Images and Trajectories ⬇️
Automatic map extraction is of great importance to urban computing and location-based services. Aerial image and GPS trajectory data refer to two different data sources that could be leveraged to generate the map, although they carry different types of information. Most previous works on data fusion between aerial images and data from auxiliary sensors do not fully utilize the information of both modalities and hence suffer from the issue of information loss. We propose a deep convolutional neural network called DeepDualMapper which fuses the aerial image and trajectory data in a more seamless manner to extract the digital map. We design a gated fusion module to explicitly control the information flows from both modalities in a complementary-aware manner. Moreover, we propose a novel densely supervised refinement decoder to generate the prediction in a coarse-to-fine way. Our comprehensive experiments demonstrate that DeepDualMapper can fuse the information of images and trajectories much more effectively than existing approaches, and is able to generate maps with higher accuracy.
8.Text Perceptron: Towards End-to-End Arbitrary-Shaped Text Spotting ⬇️
Many approaches have recently been proposed to detect irregular scene text and achieved promising results. However, their localization results may not well satisfy the following text recognition part mainly because of two reasons: 1) recognizing arbitrary shaped text is still a challenging task, and 2) prevalent non-trainable pipeline strategies between text detection and text recognition will lead to suboptimal performances. To handle this incompatibility problem, in this paper we propose an end-to-end trainable text spotting approach named Text Perceptron. Concretely, Text Perceptron first employs an efficient segmentation-based text detector that learns the latent text reading order and boundary information. Then a novel Shape Transform Module (abbr. STM) is designed to transform the detected feature regions into regular morphologies without extra parameters. It unites text detection and the following recognition part into a whole framework, and helps the whole network achieve global optimization. Experiments show that our method achieves competitive performance on two standard text benchmarks, i.e., ICDAR 2013 and ICDAR 2015, and also obviously outperforms existing methods on irregular text benchmarks SCUT-CTW1500 and Total-Text.
9.On the Similarity of Deep Learning Representations Across Didactic and Adversarial Examples ⬇️
The increasing use of deep neural networks (DNNs) has motivated a parallel endeavor: the design of adversaries that profit from successful misclassifications. However, not all adversarial examples are crafted for malicious purposes. For example, real world systems often contain physical, temporal, and sampling variability across instrumentation. Adversarial examples in the wild may inadvertently prove deleterious for accurate predictive modeling. Conversely, naturally occurring covariance of image features may serve didactic purposes. Here, we studied the stability of deep learning representations for neuroimaging classification across didactic and adversarial conditions characteristic of MRI acquisition variability. We show that representational similarity and performance vary according to the frequency of adversarial examples in the input space.
10.Discernible Compressed Images via Deep Perception Consistency ⬇️
Image compression, as one of the fundamental low-level image processing tasks, is very essential for computer vision. Conventional image compression methods tend to obtain compressed images by minimizing their appearance discrepancy with the corresponding original images, but pay little attention to their efficacy in downstream perception tasks, e.g., image recognition and object detection. In contrast, this paper aims to produce compressed images by pursuing both appearance and perception consistency. Based on the encoder-decoder framework, we propose using a pre-trained CNN to extract features of original and compressed images. In addition, the maximum mean discrepancy (MMD) is employed to minimize the difference between feature distributions. The resulting compression network can generate images with high image quality and preserve the consistent perception in the feature domain, so that these images can be well recognized by pre-trained machine learning models. Experiments on benchmarks demonstrate the superiority of the proposed algorithm over comparison methods.
11.CQ-VQA: Visual Question Answering on Categorized Questions ⬇️
This paper proposes CQ-VQA, a novel 2-level hierarchical but end-to-end model to solve the task of visual question answering (VQA). The first level of CQ-VQA, referred to as question categorizer (QC), classifies questions to reduce the potential answer search space. The QC uses attended and fused features of the input question and image. The second level, referred to as answer predictor (AP), comprises of a set of distinct classifiers corresponding to each question category. Depending on the question category predicted by QC, only one of the classifiers of AP remains active. The loss functions of QC and AP are aggregated together to make it an end-to-end model. The proposed model (CQ-VQA) is evaluated on the TDIUC dataset and is benchmarked against state-of-the-art approaches. Results indicate competitive or better performance of CQ-VQA.
12.Deep Domain Adaptive Object Detection: a Survey ⬇️
Deep learning (DL) based object detection has achieved great progress. These methods typically assume that large amount of labeled training data is available, and training and test data are drawn from an identical distribution. However, the two assumptions are not always hold in practice. Deep domain adaptive object detection (DDAOD) has emerged as a new learning paradigm to address the above mentioned challenges. This paper aims to review the state-of-the-art progress on deep domain adaptive object detection approaches. Firstly, we introduce briefly the basic concepts of deep domain adaptation. Secondly, the deep domain adaptive detectors are classified into four categories and detailed descriptions of representative methods in each category are provided. Finally, insights for future research trend are presented.
13.Unsupervised Image-generation Enhanced Adaptation for Object Detection in Thermal images ⬇️
Object detection in thermal images is an important computer vision task and has many applications such as unmanned vehicles, robotics, surveillance and night vision. Deep learning based detectors have achieved major progress, which usually need large amount of labelled training data. However, labelled data for object detection in thermal images is scarce and expensive to collect. How to take advantage of the large number labelled visible images and adapt them into thermal image domain, is expected to solve. This paper proposes an unsupervised image-generation enhanced adaptation method for object detection in thermal images. To reduce the gap between visible domain and thermal domain, the proposed method manages to generate simulated fake thermal images that are similar to the target images, and preserves the annotation information of the visible source domain. The image generation includes a CycleGAN based image-to-image translation and an intensity inversion transformation. Generated fake thermal images are used as renewed source domain. And then the off-the-shelf Domain Adaptive Faster RCNN is utilized to reduce the gap between generated intermediate domain and the thermal target domain. Experiments demonstrate the effectiveness and superiority of the proposed method.
14.Superpixel Segmentation via Convolutional Neural Networks with Regularized Information Maximization ⬇️
We propose an unsupervised superpixel segmentation method by optimizing a randomly-initialized convolutional neural network (CNN) in inference time. Our method generates superpixels via CNN from a single image without any labels by minimizing a proposed objective function for superpixel segmentation in inference time. There are three advantages to our method compared with many of existing methods: (i) leverages an image prior of CNN for superpixel segmentation, (ii) adaptively changes the number of superpixels according to the given images, and (iii) controls the property of superpixels by adding an auxiliary cost to the objective function. We verify the advantages of our method quantitatively and qualitatively on BSDS500 and SBD datasets.
15.Directional Deep Embedding and Appearance Learning for Fast Video Object Segmentation ⬇️
Most recent semi-supervised video object segmentation (VOS) methods rely on fine-tuning deep convolutional neural networks online using the given mask of the first frame or predicted masks of subsequent frames. However, the online fine-tuning process is usually time-consuming, limiting the practical use of such methods. We propose a directional deep embedding and appearance learning (DDEAL) method, which is free of the online fine-tuning process, for fast VOS. First, a global directional matching module, which can be efficiently implemented by parallel convolutional operations, is proposed to learn a semantic pixel-wise embedding as an internal guidance. Second, an effective directional appearance model based statistics is proposed to represent the target and background on a spherical embedding space for VOS. Equipped with the global directional matching module and the directional appearance model learning module, DDEAL learns static cues from the labeled first frame and dynamically updates cues of the subsequent frames for object segmentation. Our method exhibits state-of-the-art VOS performance without using online fine-tuning. Specifically, it achieves a J & F mean score of 74.8% on DAVIS 2017 dataset and an overall score G of 71.3% on the large-scale YouTube-VOS dataset, while retaining a speed of 25 fps with a single NVIDIA TITAN Xp GPU. Furthermore, our faster version runs 31 fps with only a little accuracy loss. Our code and trained networks are available at this https URL.
16.Generator From Edges: Reconstruction of Facial Images ⬇️
Applications that involve supervised training require paired images. Researchers of single image super-resolution (SISR) create such images by artificially generating blurry input images from the corresponding ground truth. Similarly we can create paired images with the canny edge. We propose Generator From Edges (GFE) [Figure 2]. Our aim is to determine the best architecture for GFE, along with reviews of perceptual loss [1, 2]. To this end, we conducted three experiments. First, we explored the effects of the adversarial loss often used in SISR. In particular, we uncovered that it is not an essential component to form a perceptual loss. Eliminating adversarial loss will lead to a more effective architecture from the perspective of hardware resource. It also means that considerations for the problems pertaining to generative adversarial network (GAN) [3], such as mode collapse, are not necessary. Second, we reexamined VGG loss and found that the mid-layers yield the best results. By extracting the full potential of VGG loss, the overall performance of perceptual loss improves significantly. Third, based on the findings of the first two experiments, we reevaluated the dense network to construct GFE. Using GFE as an intermediate process, reconstructing a facial image from a pencil sketch can become an easy task.
17.AOL: Adaptive Online Learning for Human Trajectory Prediction in Dynamic Video Scenes ⬇️
We present a novel adaptive online learning (AOL) framework to predict human movement trajectories in dynamic video scenes. Our framework learns and adapts to changes in the scene environment and generates best network weights for different scenarios. The framework can be applied to prediction models and improve their performance as it dynamically adjusts when it encounters changes in the scene and can apply the best training weights for predicting the next locations. We demonstrate this by integrating our framework with two existing prediction models: LSTM [3] and Future Person Location (FPL) [1]. Furthermore, we analyze the number of network weights for optimal performance and show that we can achieve real-time with a fixed number of networks using the least recently used (LRU) strategy for maintaining the most recently trained network weights. With extensive experiments, we show that our framework increases prediction accuracies of LSTM and FPL by ~17% and 28% on average, and up to ~50% for FPL on the worst case while achieving real-time (20fps).
18.PeelNet: Textured 3D reconstruction of human body using single view RGB image ⬇️
Reconstructing human shape and pose from a single image is a challenging problem due to issues like severe self-occlusions, clothing variations, and changes in lighting to name a few. Many applications in the entertainment industry, e-commerce, health-care (physiotherapy), and mobile-based AR/VR platforms can benefit from recovering the 3D human shape, pose, and texture. In this paper, we present PeelNet, an end-to-end generative adversarial framework to tackle the problem of textured 3D reconstruction of the human body from a single RGB image. Motivated by ray tracing for generating realistic images of a 3D scene, we tackle this problem by representing the human body as a set of peeled depth and RGB maps which are obtained by extending rays beyond the first intersection with the 3D object. This formulation allows us to handle self-occlusions efficiently. Current parametric model-based approaches fail to model loose clothing and surface-level details and are proposed for the underlying naked human body. Majority of non-parametric approaches are either computationally expensive or provide unsatisfactory results. We present a simple non-parametric solution where the peeled maps are generated from a single RGB image as input. Our proposed peeled depth maps are back-projected to 3D volume to obtain a complete 3D shape. The corresponding RGB maps provide vertex-level texture details. We compare our method against current state-of-the-art methods in 3D reconstruction and demonstrate the effectiveness of our method on BUFF and MonoPerfCap datasets.
19.Latent Normalizing Flows for Many-to-Many Cross-Domain Mappings ⬇️
Learned joint representations of images and text form the backbone of several important cross-domain tasks such as image captioning. Prior work mostly maps both domains into a common latent representation in a purely supervised fashion. This is rather restrictive, however, as the two domains follow distinct generative processes. Therefore, we propose a novel semi-supervised framework, which models shared information between domains and domain-specific information separately. The information shared between the domains is aligned with an invertible neural network. Our model integrates normalizing flow-based priors for the domain-specific information, which allows us to learn diverse many-to-many mappings between the two domains. We demonstrate the effectiveness of our model on diverse tasks, including image captioning and text-to-image synthesis.
20.Block Annotation: Better Image Annotation for Semantic Segmentation with Sub-Image Decomposition ⬇️
Image datasets with high-quality pixel-level annotations are valuable for semantic segmentation: labelling every pixel in an image ensures that rare classes and small objects are annotated. However, full-image annotations are expensive, with experts spending up to 90 minutes per image. We propose block sub-image annotation as a replacement for full-image annotation. Despite the attention cost of frequent task switching, we find that block annotations can be crowdsourced at higher quality compared to full-image annotation with equal monetary cost using existing annotation tools developed for full-image annotation. Surprisingly, we find that 50% pixels annotated with blocks allows semantic segmentation to achieve equivalent performance to 100% pixels annotated. Furthermore, as little as 12% of pixels annotated allows performance as high as 98% of the performance with dense annotation. In weakly-supervised settings, block annotation outperforms existing methods by 3-4% (absolute) given equivalent annotation time. To recover the necessary global structure for applications such as characterizing spatial context and affordance relationships, we propose an effective method to inpaint block-annotated images with high-quality labels without additional human effort. As such, fewer annotations can also be used for these applications compared to full-image annotation.
21.CRL: Class Representative Learning for Image Classification ⬇️
Building robust and real-time classifiers with diverse datasets are one of the most significant challenges to deep learning researchers. It is because there is a considerable gap between a model built with training (seen) data and real (unseen) data in applications. Recent works including Zero-Shot Learning (ZSL), have attempted to deal with this problem of overcoming the apparent gap through transfer learning. In this paper, we propose a novel model, called Class Representative Learning Model (CRL), that can be especially effective in image classification influenced by ZSL. In the CRL model, first, the learning step is to build class representatives to represent classes in datasets by aggregating prominent features extracted from a Convolutional Neural Network (CNN). Second, the inferencing step in CRL is to match between the class representatives and new data. The proposed CRL model demonstrated superior performance compared to the current state-of-the-art research in ZSL and mobile deep learning. The proposed CRL model has been implemented and evaluated in a parallel environment, using Apache Spark, for both distributed learning and recognition. An extensive experimental study on the benchmark datasets, ImageNet-1K, CalTech-101, CalTech-256, CIFAR-100, shows that CRL can build a class distribution model with drastic improvement in learning and recognition performance without sacrificing accuracy compared to the state-of-the-art performances in image classification.
22.Key Points Estimation and Point Instance Segmentation Approach for Lane Detection ⬇️
State-of-the-art lane detection methods achieve successful performance. Despite their advantages, these methods have critical deficiencies such as the limited number of detectable lanes and high false positive. In especial, high false positive can cause wrong and dangerous control. In this paper, we propose a novel lane detection method for the arbitrary number of lanes using the deep learning method, which has the lower number of false positives than other recent lane detection methods. The architecture of the proposed method has the shared feature extraction layers and several branches for detection and embedding to cluster lanes. The proposed method can generate exact points on the lanes, and we cast a clustering problem for the generated points as a point cloud instance segmentation problem. The proposed method is more compact because it generates fewer points than the original image pixel size. Our proposed post processing method eliminates outliers successfully and increases the performance notably. Whole proposed framework achieves competitive results on the tuSimple dataset.
23.Analytic Marching: An Analytic Meshing Solution from Deep Implicit Surface Networks ⬇️
This paper studies a problem of learning surface mesh via implicit functions in an emerging field of deep learning surface reconstruction, where implicit functions are popularly implemented as multi-layer perceptrons (MLPs) with rectified linear units (ReLU). To achieve meshing from learned implicit functions, existing methods adopt the de-facto standard algorithm of marching cubes; while promising, they suffer from loss of precision learned in the MLPs, due to the discretization nature of marching cubes. Motivated by the knowledge that a ReLU based MLP partitions its input space into a number of linear regions, we identify from these regions analytic cells and analytic faces that are associated with zero-level isosurface of the implicit function, and characterize the theoretical conditions under which the identified analytic faces are guaranteed to connect and form a closed, piecewise planar surface. Based on our theorem, we propose a naturally parallelizable algorithm of analytic marching, which marches among analytic cells to exactly recover the mesh captured by a learned MLP. Experiments on deep learning mesh reconstruction verify the advantages of our algorithm over existing ones.
24.Automated Labelling using an Attention model for Radiology reports of MRI scans (ALARM) ⬇️
Labelling large datasets for training high-capacity neural networks is a major obstacle to the development of deep learning-based medical imaging applications. Here we present a transformer-based network for magnetic resonance imaging (MRI) radiology report classification which automates this task by assigning image labels on the basis of free-text expert radiology reports. Our model's performance is comparable to that of an expert radiologist, and better than that of an expert physician, demonstrating the feasibility of this approach. We make code available online for researchers to label their own MRI datasets for medical imaging applications.
25.Reinforced active learning for image segmentation ⬇️
Learning-based approaches for semantic segmentation have two inherent challenges. First, acquiring pixel-wise labels is expensive and time-consuming. Second, realistic segmentation datasets are highly unbalanced: some categories are much more abundant than others, biasing the performance to the most represented ones. In this paper, we are interested in focusing human labelling effort on a small subset of a larger pool of data, minimizing this effort while maximizing performance of a segmentation model on a hold-out set. We present a new active learning strategy for semantic segmentation based on deep reinforcement learning (RL). An agent learns a policy to select a subset of small informative image regions -- opposed to entire images -- to be labeled, from a pool of unlabeled data. The region selection decision is made based on predictions and uncertainties of the segmentation model being trained. Our method proposes a new modification of the deep Q-network (DQN) formulation for active learning, adapting it to the large-scale nature of semantic segmentation problems. We test the proof of concept in CamVid and provide results in the large-scale dataset Cityscapes. On Cityscapes, our deep RL region-based DQN approach requires roughly 30% less additional labeled data than our most competitive baseline to reach the same performance. Moreover, we find that our method asks for more labels of under-represented categories compared to the baselines, improving their performance and helping to mitigate class imbalance.
26.Topological Mapping for Manhattan-like Repetitive Environments ⬇️
We showcase a topological mapping framework for a challenging indoor warehouse setting. At the most abstract level, the warehouse is represented as a Topological Graph where the nodes of the graph represent a particular warehouse topological construct (e.g. rackspace, corridor) and the edges denote the existence of a path between two neighbouring nodes or topologies. At the intermediate level, the map is represented as a Manhattan Graph where the nodes and edges are characterized by Manhattan properties and as a Pose Graph at the lower-most level of detail. The topological constructs are learned via a Deep Convolutional Network while the relational properties between topological instances are learnt via a Siamese-style Neural Network. In the paper, we show that maintaining abstractions such as Topological Graph and Manhattan Graph help in recovering an accurate Pose Graph starting from a highly erroneous and unoptimized Pose Graph. We show how this is achieved by embedding topological and Manhattan relations as well as Manhattan Graph aided loop closure relations as constraints in the backend Pose Graph optimization framework. The recovery of near ground-truth Pose Graph on real-world indoor warehouse scenes vindicate the efficacy of the proposed framework.
27.Facial Attribute Capsules for Noise Face Super Resolution ⬇️
Existing face super-resolution (SR) methods mainly assume the input image to be noise-free. Their performance degrades drastically when applied to real-world scenarios where the input image is always contaminated by noise. In this paper, we propose a Facial Attribute Capsules Network (FACN) to deal with the problem of high-scale super-resolution of noisy face image. Capsule is a group of neurons whose activity vector models different properties of the same entity. Inspired by the concept of capsule, we propose an integrated representation model of facial information, which named Facial Attribute Capsule (FAC). In the SR processing, we first generated a group of FACs from the input LR face, and then reconstructed the HR face from this group of FACs. Aiming to effectively improve the robustness of FAC to noise, we generate FAC in semantic, probabilistic and facial attributes manners by means of integrated learning strategy. Each FAC can be divided into two sub-capsules: Semantic Capsule (SC) and Probabilistic Capsule (PC). Them describe an explicit facial attribute in detail from two aspects of semantic representation and probability distribution. The group of FACs model an image as a combination of facial attribute information in the semantic space and probabilistic space by an attribute-disentangling way. The diverse FACs could better combine the face prior information to generate the face images with fine-grained semantic attributes. Extensive benchmark experiments show that our method achieves superior hallucination results and outperforms state-of-the-art for very low resolution (LR) noise face image super resolution.
28.A Real-Time Deep Network for Crowd Counting ⬇️
Automatic analysis of highly crowded people has attracted extensive attention from computer vision research. Previous approaches for crowd counting have already achieved promising performance across various benchmarks. However, to deal with the real situation, we hope the model run as fast as possible while keeping accuracy. In this paper, we propose a compact convolutional neural network for crowd counting which learns a more efficient model with a small number of parameters. With three parallel filters executing the convolutional operation on the input image simultaneously at the front of the network, our model could achieve nearly real-time speed and save more computing resources. Experiments on two benchmarks show that our proposed method not only takes a balance between performance and efficiency which is more suitable for actual scenes but also is superior to existing light-weight models in speed.
29.Face Recognition: Too Bias, or Not Too Bias? ⬇️
We reveal critical insights into problems of bias in state-of-the-art facial recognition (FR) systems using a novel Balanced Faces In the Wild (BFW) dataset: data balanced for gender and ethnic groups. We show variations in the optimal scoring threshold for face-pairs across different subgroups. Thus, the conventional approach of learning a global threshold for all pairs resulting in performance gaps among subgroups. By learning subgroup-specific thresholds, we not only mitigate problems in performance gaps but also show a notable boost in the overall performance. Furthermore, we do a human evaluation to measure the bias in humans, which supports the hypothesis that such a bias exists in human perception. For the BFW database, source code, and more, visit this http URL.
30.Learning to Group: A Bottom-Up Framework for 3D Part Discovery in Unseen Categories ⬇️
We address the problem of discovering 3D parts for objects in unseen categories. Being able to learn the geometry prior of parts and transfer this prior to unseen categories pose fundamental challenges on data-driven shape segmentation approaches. Formulated as a contextual bandit problem, we propose a learning-based agglomerative clustering framework which learns a grouping policy to progressively group small part proposals into bigger ones in a bottom-up fashion. At the core of our approach is to restrict the local context for extracting part-level features, which encourages the generalizability to unseen categories. On the large-scale fine-grained 3D part dataset, PartNet, we demonstrate that our method can transfer knowledge of parts learned from 3 training categories to 21 unseen testing categories without seeing any annotated samples. Quantitative comparisons against four shape segmentation baselines shows that our approach achieve the state-of-the-art performance.
31.A Multiple Decoder CNN for Inverse Consistent 3D Image Registration ⬇️
The recent application of deep learning technologies in medical image registration has exponentially decreased the registration time and gradually increased registration accuracy when compared to their traditional counterparts. Most of the learning-based registration approaches considers this task as a one directional problem. As a result, only correspondence from the moving image to the target image is considered. However, in some medical procedures bidirectional registration is required to be performed. Unlike other learning-based registration, we propose a registration framework with inverse consistency. The proposed method simultaneously learns forward transformation and backward transformation in an unsupervised manner. We perform training and testing of the method on the publicly available LPBA40 MRI dataset and demonstrate strong performance than baseline registration methods.
32.HighRes-net: Recursive Fusion for Multi-Frame Super-Resolution of Satellite Imagery ⬇️
Generative deep learning has sparked a new wave of Super-Resolution (SR) algorithms that enhance single images with impressive aesthetic results, albeit with imaginary details. Multi-frame Super-Resolution (MFSR) offers a more grounded approach to the ill-posed problem, by conditioning on multiple low-resolution views. This is important for satellite monitoring of human impact on the planet -- from deforestation, to human rights violations -- that depend on reliable imagery. To this end, we present HighRes-net, the first deep learning approach to MFSR that learns its sub-tasks in an end-to-end fashion: (i) co-registration, (ii) fusion, (iii) up-sampling, and (iv) registration-at-the-loss. Co-registration of low-resolution views is learned implicitly through a reference-frame channel, with no explicit registration mechanism. We learn a global fusion operator that is applied recursively on an arbitrary number of low-resolution pairs. We introduce a registered loss, by learning to align the SR output to a ground-truth through ShiftNet. We show that by learning deep representations of multiple views, we can super-resolve low-resolution signals and enhance Earth Observation data at scale. Our approach recently topped the European Space Agency's MFSR competition on real-world satellite imagery.
33.An End-to-End Framework for Unsupervised Pose Estimation of Occluded Pedestrians ⬇️
Pose estimation in the wild is a challenging problem, particularly in situations of (i) occlusions of varying degrees and (ii) crowded outdoor scenes. Most of the existing studies of pose estimation did not report the performance in similar situations. Moreover, pose annotations for occluded parts of human figures have not been provided in any of the relevant standard datasets which in turn creates further difficulties to the required studies for pose estimation of the entire figure of occluded humans. Well known pedestrian detection datasets such as CityPersons contains samples of outdoor scenes but it does not include pose annotations. Here, we propose a novel multi-task framework for end-to-end training towards the entire pose estimation of pedestrians including in situations of any kind of occlusion. To tackle this problem for training the network, we make use of a pose estimation dataset, MS-COCO, and employ unsupervised adversarial instance-level domain adaptation for estimating the entire pose of occluded pedestrians. The experimental studies show that the proposed framework outperforms the SOTA results for pose estimation, instance segmentation and pedestrian detection in cases of heavy occlusions (HO) and reasonable + heavy occlusions (R + HO) on the two benchmark datasets.
34.Scale-Invariant Multi-Oriented Text Detection in Wild Scene Images ⬇️
Automatic detection of scene texts in the wild is a challenging problem, particularly due to the difficulties in handling (i) occlusions of varying percentages, (ii) widely different scales and orientations, (iii) severe degradations in the image quality etc. In this article, we propose a fully convolutional neural network architecture consisting of a novel Feature Representation Block (FRB) capable of efficient abstraction of information. The proposed network has been trained using curriculum learning with respect to difficulties in image samples and gradual pixel-wise blurring. It is capable of detecting texts of different scales and orientations suffered by blurring from multiple possible sources, non-uniform illumination as well as partial occlusions of varying percentages. Text detection performance of the proposed framework on various benchmark sample databases including ICDAR 2015, ICDAR 2017 MLT, COCO-Text and MSRA-TD500 improves respective state-of-the-art results significantly. Source code of the proposed architecture will be made available at github.
35.Video Face Super-Resolution with Motion-Adaptive Feedback Cell ⬇️
Video super-resolution (VSR) methods have recently achieved a remarkable success due to the development of deep convolutional neural networks (CNN). Current state-of-the-art CNN methods usually treat the VSR problem as a large number of separate multi-frame super-resolution tasks, at which a batch of low resolution (LR) frames is utilized to generate a single high resolution (HR) frame, and running a slide window to select LR frames over the entire video would obtain a series of HR frames. However, duo to the complex temporal dependency between frames, with the number of LR input frames increase, the performance of the reconstructed HR frames become worse. The reason is in that these methods lack the ability to model complex temporal dependencies and hard to give an accurate motion estimation and compensation for VSR process. Which makes the performance degrade drastically when the motion in frames is complex. In this paper, we propose a Motion-Adaptive Feedback Cell (MAFC), a simple but effective block, which can efficiently capture the motion compensation and feed it back to the network in an adaptive way. Our approach efficiently utilizes the information of the inter-frame motion, the dependence of the network on motion estimation and compensation method can be avoid. In addition, benefiting from the excellent nature of MAFC, the network can achieve better performance in the case of extremely complex motion scenarios. Extensive evaluations and comparisons validate the strengths of our approach, and the experimental results demonstrated that the proposed framework is outperform the state-of-the-art methods.
36.UniViLM: A Unified Video and Language Pre-Training Model for Multimodal Understanding and Generation ⬇️
We propose UniViLM: a Unified Video and Language pre-training Model for multimodal understanding and generation. Motivated by the recent success of BERT based pre-training technique for NLP and image-language tasks, VideoBERT and CBT are proposed to exploit BERT model for video and language pre-training using narrated instructional videos. Different from their works which only pre-train understanding task, we propose a unified video-language pre-training model for both understanding and generation tasks. Our model comprises of 4 components including two single-modal encoders, a cross encoder and a decoder with the Transformer backbone. We first pre-train our model to learn the universal representation for both video and language on a large instructional video dataset. Then we fine-tune the model on two multimodal tasks including understanding task (text-based video retrieval) and generation task (multimodal video captioning). Our extensive experiments show that our method can improve the performance of both understanding and generation tasks and achieves the state-of-the art results.
37.Cell R-CNN V3: A Novel Panoptic Paradigm for Instance Segmentation in Biomedical Images ⬇️
Instance segmentation is an important task for biomedical image analysis. Due to the complicated background components, the high variability of object appearances, numerous overlapping objects, and ambiguous object boundaries, this task still remains challenging. Recently, deep learning based methods have been widely employed to solve these problems and can be categorized into proposal-free and proposal-based methods. However, both proposal-free and proposal-based methods suffer from information loss, as they focus on either global-level semantic or local-level instance features. To tackle this issue, we present a panoptic architecture that unifies the semantic and instance features in this work. Specifically, our proposed method contains a residual attention feature fusion mechanism to incorporate the instance prediction with the semantic features, in order to facilitate the semantic contextual information learning in the instance branch. Then, a mask quality branch is designed to align the confidence score of each object with the quality of the mask prediction. Furthermore, a consistency regularization mechanism is designed between the semantic segmentation tasks in the semantic and instance branches, for the robust learning of both tasks. Extensive experiments demonstrate the effectiveness of our proposed method, which outperforms several state-of-the-art methods on various biomedical datasets.
38.Recognizing Families In the Wild (RFIW): The 4th Edition ⬇️
Recognizing Families In the Wild (RFIW): an annual large-scale, multi-track automatic kinship recognition evaluation that supports various visual kin-based problems on scales much higher than ever before. Organized in conjunction with the 15th IEEE International Conference on Automatic Face and Gesture Recognition (FG) as a Challenge, RFIW provides a platform for publishing original work and the gathering of experts for a discussion of the next steps. This paper summarizes the supported tasks (i.e., kinship verification, tri-subject verification, and search & retrieval of missing children) in the evaluation protocols, which include the practical motivation, technical background, data splits, metrics, and benchmark results. Furthermore, top submissions (i.e., leader-board stats) are listed and reviewed as a high-level analysis on the state of the problem. In the end, the purpose of this paper is to describe the 2020 RFIW challenge, end-to-end, along with forecasts in promising future directions.
39.Historical Document Processing: Historical Document Processing: A Survey of Techniques, Tools, and Trends ⬇️
Historical Document Processing is the process of digitizing written material from the past for future use by historians and other scholars. It incorporates algorithms and software tools from various subfields of computer science, including computer vision, document analysis and recognition, natural language processing, and machine learning, to convert images of ancient manuscripts, letters, diaries, and early printed texts automatically into a digital format usable in data mining and information retrieval systems. Within the past twenty years, as libraries, museums, and other cultural heritage institutions have scanned an increasing volume of their historical document archives, the need to transcribe the full text from these collections has become acute. Since Historical Document Processing encompasses multiple sub-domains of computer science, knowledge relevant to its purpose is scattered across numerous journals and conference proceedings. This paper surveys the major phases of, standard algorithms, tools, and datasets in the field of Historical Document Processing, discusses the results of a literature review, and finally suggests directions for further research.
40.Single Unit Status in Deep Convolutional Neural Network Codes for Face Identification: Sparseness Redefined ⬇️
Deep convolutional neural networks (DCNNs) trained for face identification develop representations that generalize over variable images, while retaining subject (e.g., gender) and image (e.g., viewpoint) information. Identity, gender, and viewpoint codes were studied at the "neural unit" and ensemble levels of a face-identification network. At the unit level, identification, gender classification, and viewpoint estimation were measured by deleting units to create variably-sized, randomly-sampled subspaces at the top network layer. Identification of 3,531 identities remained high (area under the ROC approximately 1.0) as dimensionality decreased from 512 units to 16 (0.95), 4 (0.80), and 2 (0.72) units. Individual identities separated statistically on every top-layer unit. Cross-unit responses were minimally correlated, indicating that units code non-redundant identity cues. This "distributed" code requires only a sparse, random sample of units to identify faces accurately. Gender classification declined gradually and viewpoint estimation fell steeply as dimensionality decreased. Individual units were weakly predictive of gender and viewpoint, but ensembles proved effective predictors. Therefore, distributed and sparse codes co-exist in the network units to represent different face attributes. At the ensemble level, principal component analysis of face representations showed that identity, gender, and viewpoint information separated into high-dimensional subspaces, ordered by explained variance. Identity, gender, and viewpoint information contributed to all individual unit responses, undercutting a neural tuning analogy for face attributes. Interpretation of neural-like codes from DCNNs, and by analogy, high-level visual codes, cannot be inferred from single unit responses. Instead, "meaning" is encoded by directions in the high-dimensional space.
41.Layered Embeddings for Amodal Instance Segmentation ⬇️
The proposed method extends upon the representational output of semantic instance segmentation by explicitly including both visible and occluded parts. A fully convolutional network is trained to produce consistent pixel-level embedding across two layers such that, when clustered, the results convey the full spatial extent and depth ordering of each instance. Results demonstrate that the network can accurately estimate complete masks in the presence of occlusion and outperform leading top-down bounding-box approaches. Source code available at this https URL
42.Why Do Line Drawings Work? A Realism Hypothesis ⬇️
Why is it that we can recognize object identity and 3D shape from line drawings, even though they do not exist in the natural world? This paper hypothesizes that the human visual system perceives line drawings as if they were approximately realistic images. Moreover, the techniques of line drawing are chosen to accurately convey shape to a human observer. Several implications and variants of this hypothesis are explored.
43.Social-WaGDAT: Interaction-aware Trajectory Prediction via Wasserstein Graph Double-Attention Network ⬇️
Effective understanding of the environment and accurate trajectory prediction of surrounding dynamic obstacles are indispensable for intelligent mobile systems (like autonomous vehicles and social robots) to achieve safe and high-quality planning when they navigate in highly interactive and crowded scenarios. Due to the existence of frequent interactions and uncertainty in the scene evolution, it is desired for the prediction system to enable relational reasoning on different entities and provide a distribution of future trajectories for each agent. In this paper, we propose a generic generative neural system (called Social-WaGDAT) for multi-agent trajectory prediction, which makes a step forward to explicit interaction modeling by incorporating relational inductive biases with a dynamic graph representation and leverages both trajectory and scene context information. We also employ an efficient kinematic constraint layer applied to vehicle trajectory prediction which not only ensures physical feasibility but also enhances model performance. The proposed system is evaluated on three public benchmark datasets for trajectory prediction, where the agents cover pedestrians, cyclists and on-road vehicles. The experimental results demonstrate that our model achieves better performance than various baseline approaches in terms of prediction accuracy.
44.Spectrum Translation for Cross-Spectral Ocular Matching ⬇️
Cross-spectral verification remains a big issue in biometrics, especially for the ocular area due to differences in the reflected features in the images depending on the region and spectrum used.
In this paper, we investigate the use of Conditional Adversarial Networks for spectrum translation between near infra-red and visual light images for ocular biometrics. We analyze the transformation based on the overall visual quality of the transformed images and the accuracy drop of the identification system when trained with opposing data.
We use the PolyU database and propose two different systems for biometric verification, the first one based on Siamese Networks trained with Softmax and Cross-Entropy loss, and the second one a Triplet Loss network. We achieved an EER of 1% when using a Triplet Loss network trained for NIR and finding the Euclidean distance between the real NIR images and the fake ones translated from the visible spectrum. We also outperform previous results using baseline algorithms.
45.Seeing Around Corners with Edge-Resolved Transient Imaging ⬇️
Non-line-of-sight (NLOS) imaging is a rapidly growing field seeking to form images of objects outside the field of view, with potential applications in search and rescue, reconnaissance, and even medical imaging. The critical challenge of NLOS imaging is that diffuse reflections scatter light in all directions, resulting in weak signals and a loss of directional information. To address this problem, we propose a method for seeing around corners that derives angular resolution from vertical edges and longitudinal resolution from the temporal response to a pulsed light source. We introduce an acquisition strategy, scene response model, and reconstruction algorithm that enable the formation of 2.5-dimensional representations -- a plan view plus heights -- and a 180$^{\circ}$ field of view (FOV) for large-scale scenes. Our experiments demonstrate accurate reconstructions of hidden rooms up to 3 meters in each dimension.
46.Query-Efficient Physical Hard-Label Attacks on Deep Learning Visual Classification ⬇️
We present Survival-OPT, a physical adversarial example algorithm in the black-box hard-label setting where the attacker only has access to the model prediction class label. Assuming such limited access to the model is more relevant for settings such as proprietary cyber-physical and cloud systems than the whitebox setting assumed by prior work. By leveraging the properties of physical attacks, we create a novel approach based on the survivability of perturbations corresponding to physical transformations. Through simply querying the model for hard-label predictions, we optimize perturbations to survive in many different physical conditions and show that adversarial examples remain a security risk to cyber-physical systems (CPSs) even in the hard-label threat model. We show that Survival-OPT is query-efficient and robust: using fewer than 200K queries, we successfully attack a stop sign to be misclassified as a speed limit 30 km/hr sign in 98.5% of video frames in a drive-by setting. Survival-OPT also outperforms our baseline combination of existing hard-label and physical approaches, which required over 10x more queries for less robust results.
47.PCSGAN: Perceptual Cyclic-Synthesized Generative Adversarial Networks for Thermal and NIR to Visible Image Transformation ⬇️
In many real world scenarios, it is difficult to capture the images in the visible light spectrum (VIS) due to bad lighting conditions. However, the images can be captured in such scenarios using Near-Infrared (NIR) and Thermal (THM) cameras. The NIR and THM images contain the limited details. Thus, there is a need to transform the images from THM/NIR to VIS for better understanding. However, it is non-trivial task due to the large domain discrepancies and lack of abundant datasets. Nowadays, Generative Adversarial Network (GAN) is able to transform the images from one domain to another domain. Most of the available GAN based methods use the combination of the adversarial and the pixel-wise losses (like L1 or L2) as the objective function for training. The quality of transformed images in case of THM/NIR to VIS transformation is still not up to the mark using such objective function. Thus, better objective functions are needed to improve the quality, fine details and realism of the transformed images. A new model for THM/NIR to VIS image transformation called Perceptual Cyclic-Synthesized Generative Adversarial Network (PCSGAN) is introduced to address these issues. The PCSGAN uses the combination of the perceptual (i.e., feature based) losses along with the pixel-wise and the adversarial losses. Both the quantitative and qualitative measures are used to judge the performance of the PCSGAN model over the WHU-IIP face and the RGB-NIR scene datasets. The proposed PCSGAN outperforms the state-of-the-art image transformation models, including Pix2pix, DualGAN, CycleGAN, PS2GAN, and PAN in terms of the SSIM, MSE, PSNR and LPIPS evaluation measures. The code is available at: \url{this https URL}.
48.Large-scale biometry with interpretable neural network regression on UK Biobank body MRI ⬇️
The UK Biobank study has successfully imaged more than 32,000 volunteer participants with neck-to-knee body MRI. Each scan is linked to extensive metadata, providing a comprehensive survey of imaged anatomy and related health states. Despite its potential for research, this vast amount of data presents a challenge to established methods of evaluation, which often rely on manual input. To date, the range of reference values for cardiovascular and metabolic risk factors is therefore incomplete. In this work, neural networks were trained for regression to infer various biological metrics from the neck-to-knee body MRI automatically. The approach requires no manual intervention or ground truth segmentations for training. The examined fields span 64 variables derived from anthropometric measurements, dual-energy X-ray absorptiometry (DXA), atlas-based segmentations, and dedicated liver scans. The standardized framework achieved a close fit to the target values (median R^2 > 0.97) in 7-fold cross-validation with the ResNet50. Interpretation of aggregated saliency maps suggests that the network correctly targets specific body regions and limbs, and learned to emulate different modalities. On several body composition metrics, the quality of the predictions is within the range of variability observed between established gold standard techniques.
49.Class-Imbalanced Semi-Supervised Learning ⬇️
Semi-Supervised Learning (SSL) has achieved great success in overcoming the difficulties of labeling and making full use of unlabeled data. However, SSL has a limited assumption that the numbers of samples in different classes are balanced, and many SSL algorithms show lower performance for the datasets with the imbalanced class distribution. In this paper, we introduce a task of class-imbalanced semi-supervised learning (CISSL), which refers to semi-supervised learning with class-imbalanced data. In doing so, we consider class imbalance in both labeled and unlabeled sets. First, we analyze existing SSL methods in imbalanced environments and examine how the class imbalance affects SSL methods. Then we propose Suppressed Consistency Loss (SCL), a regularization method robust to class imbalance. Our method shows better performance than the conventional methods in the CISSL environment. In particular, the more severe the class imbalance and the smaller the size of the labeled data, the better our method performs.
50.Reinforcement learning for the manipulation of eye tracking data ⬇️
In this paper, we present an approach based on reinforcement learning for eye tracking data manipulation. It is based on two opposing agents, where one tries to classify the data correctly and the second agent looks for patterns in the data, which get manipulated to hide specific information. We show that our approach is successfully applicable to preserve the privacy of a subject. In addition, our approach allows to evaluate the importance of temporal, as well as spatial, information of eye tracking data for specific classification goals. In general, this approach can also be used for stimuli manipulation, making it interesting for gaze guidance. For this purpose, this work provides the theoretical basis, which is why we have also integrated a section on how to apply this method for gaze guidance.
51.Unraveling Meta-Learning: Understanding Feature Representations for Few-Shot Tasks ⬇️
Meta-learning algorithms produce feature extractors which achieve state-of-the-art performance on few-shot classification. While the literature is rich with meta-learning methods, little is known about why the resulting feature extractors perform so well. We develop a better understanding of the underlying mechanics of meta-learning and the difference between models trained using meta-learning and models which are trained classically. In doing so, we develop several hypotheses for why meta-learned models perform better. In addition to visualizations, we design several regularizers inspired by our hypotheses which improve performance on few-shot classification.
52.Gaussian Smoothen Semantic Features (GSSF) -- Exploring the Linguistic Aspects of Visual Captioning in Indian Languages (Bengali) Using MSCOCO Framework ⬇️
In this work, we have introduced Gaussian Smoothen Semantic Features (GSSF) for Better Semantic Selection for Indian regional language-based image captioning and introduced a procedure where we used the existing translation and English crowd-sourced sentences for training. We have shown that this architecture is a promising alternative source, where there is a crunch in resources. Our main contribution of this work is the development of deep learning architectures for the Bengali language (is the fifth widely spoken language in the world) with a completely different grammar and language attributes. We have shown that these are working well for complex applications like language generation from image contexts and can diversify the representation through introducing constraints, more extensive features, and unique feature spaces. We also established that we could achieve absolute precision and diversity when we use smoothened semantic tensor with the traditional LSTM and feature decomposition networks. With better learning architecture, we succeeded in establishing an automated algorithm and assessment procedure that can help in the evaluation of competent applications without the requirement for expertise and human intervention.
53.Coresets for the Nearest-Neighbor Rule ⬇️
The problem of nearest-neighbor condensation deals with finding a subset R from a set of labeled points P such that for every point p in R the nearest-neighbor of p in R has the same label as p. This is motivated by applications in classification, where the nearest-neighbor rule assigns to an unlabeled query point the label of its nearest-neighbor in the point set. In this context, condensation aims to reduce the size of the set needed to classify new points. However, finding such subsets of minimum cardinality is NP-hard, and most research has focused on practical heuristics without performance guarantees. Additionally, the use of exact nearest-neighbors is always assumed, ignoring the effect of condensation in the classification accuracy when nearest-neighbors are computed approximately.
In this paper, we address these shortcomings by proposing new approximation-sensitive criteria for the nearest-neighbor condensation problem, along with practical algorithms with provable performance guarantees. We characterize sufficient conditions to guarantee correct classification of unlabeled points using approximate nearest-neighbor queries on these subsets, which introduces the notion of coresets for classification with the nearest-neighbor rule. Moreover, we prove that it is NP-hard to compute subsets with these characteristics, whose cardinality approximates that of the minimum cardinality subset. Additionally, we propose new algorithms for computing such subsets, with tight approximation factors in general metrics, and improved factors for doubling metrics and l_p metrics with p >= 2. Finally, we show an alternative implementation scheme that reduces the worst-case time complexity of one of these algorithms, becoming the first truly subquadratic approximation algorithm for the nearest-neighbor condensation problem.
54.Hold me tight! Influence of discriminative features on deep network boundaries ⬇️
Important insights towards the explainability of neural networks and their properties reside in the formation of their decision boundaries. In this work, we borrow tools from the field of adversarial robustness and propose a new framework that permits to relate the features of the dataset with the distance of data samples to the decision boundary along specific directions. We demonstrate that the inductive bias of deep learning has the tendency to generate classification functions that are invariant along non-discriminative directions of the dataset. More surprisingly, we further show that training on small perturbations of the data samples are sufficient to completely change the decision boundary. This is actually the characteristic exploited by the so-called adversarial training to produce robust classifiers. Our general framework can be used to reveal the effect of specific dataset features on the macroscopic properties of deep models and to develop a better understanding of the successes and limitations of deep learning.
55.3D Dynamic Scene Graphs: Actionable Spatial Perception with Places, Objects, and Humans ⬇️
We present a unified representation for actionable spatial perception: 3D Dynamic Scene Graphs. Scene graphs are directed graphs where nodes represent entities in the scene (e.g. objects, walls, rooms), and edges represent relations (e.g. inclusion, adjacency) among nodes. Dynamic scene graphs (DSGs) extend this notion to represent dynamic scenes with moving agents (e.g. humans, robots), and to include actionable information that supports planning and decision-making (e.g. spatio-temporal relations, topology at different levels of abstraction). Our second contribution is to provide the first fully automatic Spatial PerceptIon eNgine(SPIN) to build a DSG from visual-inertial data. We integrate state-of-the-art techniques for object and human detection and pose estimation, and we describe how to robustly infer object, robot, and human nodes in crowded scenes. To the best of our knowledge, this is the first paper that reconciles visual-inertial SLAM and dense human mesh tracking. Moreover, we provide algorithms to obtain hierarchical representations of indoor environments (e.g. places, structures, rooms) and their relations. Our third contribution is to demonstrate the proposed spatial perception engine in a photo-realistic Unity-based simulator, where we assess its robustness and expressiveness. Finally, we discuss the implications of our proposal on modern robotics applications. 3D Dynamic Scene Graphs can have a profound impact on planning and decision-making, human-robot interaction, long-term autonomy, and scene prediction. A video abstract is available at this https URL
56.Learning representations of irregular particle-detector geometry with distance-weighted graph networks ⬇️
We explore the use of graph networks to deal with irregular-geometry detectors in the context of particle reconstruction. Thanks to their representation-learning capabilities, graph networks can exploit the full detector granularity, while natively managing the event sparsity and arbitrarily complex detector geometries. We introduce two distance-weighted graph network architectures, dubbed GarNet and GravNet layers, and apply them to a typical particle reconstruction task. The performance of the new architectures is evaluated on a data set of simulated particle interactions on a toy model of a highly granular calorimeter, loosely inspired by the endcap calorimeter to be installed in the CMS detector for the High-Luminosity LHC phase. We study the clustering of energy depositions, which is the basis for calorimetric particle reconstruction, and provide a quantitative comparison to alternative approaches. The proposed algorithms provide an interesting alternative to existing methods, offering equally performing or less resource-demanding solutions with less underlying assumptions on the detector geometry and, consequently, the possibility to generalize to other detectors.
57.Adaptive Kernel Estimation of the Spectral Density with Boundary Kernel Analysis ⬇️
A hybrid estimator of the log-spectral density of a stationary time series is proposed. First, a multiple taper estimate is performed, followed by kernel smoothing the log-multitaper estimate. This procedure reduces the expected mean square error by
$({\pi^2 \over 4})^{.8}$ over simply smoothing the log tapered periodogram. The optimal number of tapers is$O(N^{8/15})$ . A data adaptive implementation of a variable bandwidth kernel smoother is given. When the spectral density is discontinuous, one sided smoothing estimates are used.