1.Variational Saccading: Efficient Inference for Large Resolution Images pdf
Image classification with deep neural networks is typically restricted to images of small dimensionality such as 224x244 in Resnet models. This limitation excludes the 4000x3000 dimensional images that are taken by modern smartphone cameras and smart devices. In this work, we aim to mitigate the prohibitive inferential and memory costs of operating in such large dimensional spaces. To sample from the high-resolution original input distribution, we propose using a smaller proxy distribution to learn the co-ordinates that correspond to regions of interest in the high-dimensional space. We introduce a new principled variational lower bound that captures the relationship of the proxy distribution's posterior and the original image's co-ordinate space in a way that maximizes the conditional classification likelihood. We empirically demonstrate on one synthetic benchmark and one real world large resolution DSLR camera image dataset that our method produces comparable results with 10x faster inference and lower memory consumption than a model that utilizes the entire original input distribution.
2.Backdooring Convolutional Neural Networks via Targeted Weight Perturbations pdf
We present a new type of backdoor attack that exploits a vulnerability of convolutional neural networks (CNNs) that has been previously unstudied. In particular, we examine the application of facial recognition. Deep learning techniques are at the top of the game for facial recognition, which means they have now been implemented in many production-level systems. Alarmingly, unlike other commercial technologies such as operating systems and network devices, deep learning-based facial recognition algorithms are not presently designed with security requirements or audited for security vulnerabilities before deployment. Given how young the technology is and how abstract many of the internal workings of these algorithms are, neural network-based facial recognition systems are prime targets for security breaches. As more and more of our personal information begins to be guarded by facial recognition (e.g., the iPhone X), exploring the security vulnerabilities of these systems from a penetration testing standpoint is crucial. Along these lines, we describe a general methodology for backdooring CNNs via targeted weight perturbations. Using a five-layer CNN and ResNet-50 as case studies, we show that an attacker is able to significantly increase the chance that inputs they supply will be falsely accepted by a CNN while simultaneously preserving the error rates for legitimate enrolled classes.
3.Kernel Transformer Networks for Compact Spherical Convolution pdf
Ideally, 360° imagery could inherit the deep convolutional neural networks (CNNs) already trained with great success on perspective projection images. However, existing methods to transfer CNNs from perspective to spherical images introduce significant computational costs and/or degradations in accuracy. In this work, we present the Kernel Transformer Network (KTN). KTNs efficiently transfer convolution kernels from perspective images to the equirectangular projection of 360° images. Given a source CNN for perspective images as input, the KTN produces a function parameterized by a polar angle and kernel as output. Given a novel 360° image, that function in turn can compute convolutions for arbitrary layers and kernels as would the source CNN on the corresponding tangent plane projections. Distinct from all existing methods, KTNs allow model transfer: the same model can be applied to different source CNNs with the same base architecture. This enables application to multiple recognition tasks without re-training the KTN. Validating our approach with multiple source CNNs and datasets, we show that KTNs improve the state of the art for spherical convolution. KTNs successfully preserve the source CNN's accuracy, while offering transferability, scalability to typical image resolutions, and, in many cases, a substantially lower memory footprint.
4.Color Constancy by GANs: An Experimental Survey pdf
In this paper, we formulate the color constancy task as an image-to-image translation problem using GANs. By conducting a large set of experiments on different datasets, an experimental survey is provided on the use of different types of GANs to solve for color constancy i.e. CC-GANs (Color Constancy GANs). Based on the experimental review, recommendations are given for the design of CC-GAN architectures based on different criteria, circumstances and datasets.
5.Graph Cut Segmentation Methods Revisited with a Quantum Algorithm pdf
The design and performance of computer vision algorithms are greatly influenced by the hardware on which they are implemented. CPUs, multi-core CPUs, FPGAs and GPUs have inspired new algorithms and enabled existing ideas to be realized. This is notably the case with GPUs, which has significantly changed the landscape of computer vision research through deep learning. As the end of Moores law approaches, researchers and hardware manufacturers are exploring alternative hardware computing paradigms. Quantum computers are a very promising alternative and offer polynomial or even exponential speed-ups over conventional computing for some problems. This paper presents a novel approach to image segmentation that uses new quantum computing hardware. Segmentation is formulated as a graph cut problem that can be mapped to the quantum approximation optimization algorithm (QAOA). This algorithm can be implemented on current and near-term quantum computers. Encouraging results are presented on artificial and medical imaging data. This represents an important, practical step towards leveraging quantum computers for computer vision.
6.Real-time Indoor Scene Reconstruction with RGBD and Inertia Input pdf
Camera motion estimation is a key technique for 3D scene reconstruction and Simultaneous localization and mapping (SLAM). To make it be feasibly achieved, previous works usually assume slow camera motions, which limits its usage in many real cases. We propose an end-to-end 3D reconstruction system which combines color, depth and inertial measurements to achieve robust reconstruction with fast sensor motions. Our framework extends Kalman filter to fuse the three kinds of information and involve an iterative method to jointly optimize feature correspondences, camera poses and scene geometry. We also propose a novel geometry-aware patch deformation technique to adapt the feature appearance in image domain, leading to a more accurate feature matching under fast camera motions. Experiments show that our patch deformation method improves the accuracy of feature tracking, and our 3D reconstruction outperforms the state-of-the-art solutions under fast camera motions.
7.Scale-aware multi-level guidance for interactive instance segmentation pdf
In interactive instance segmentation, users give feedback to iteratively refine segmentation masks. The user-provided clicks are transformed into guidance maps which provide the network with necessary cues on the whereabouts of the object of interest. Guidance maps used in current systems are purely distance-based and are either too localized or non-informative. We propose a novel transformation of user clicks to generate scale-aware guidance maps that leverage the hierarchical structural information present in an image. Using our guidance maps, even the most basic FCNs are able to outperform existing approaches that require state-of-the-art segmentation networks pre-trained on large scale segmentation datasets. We demonstrate the effectiveness of our proposed transformation strategy through comprehensive experimentation in which we significantly raise state-of-the-art on four standard interactive segmentation benchmarks.
8.Optimizing Speed/Accuracy Trade-Off for Person Re-identification via Knowledge Distillation pdf
Finding a person across a camera network plays an important role in video surveillance. For a real-world person re-identification application, in order to guarantee an optimal time response, it is crucial to find the balance between accuracy and speed. We analyse this trade-off, comparing a classical method, that comprises hand-crafted feature description and metric learning, in particular, LOMO and XQDA, with state-of-the-art deep learning techniques, using image classification networks, ResNet and MobileNets. Additionally, we propose and analyse network distillation as a learning strategy to reduce the computational cost of the deep learning approach at test time. We evaluate both methods on the Market-1501 and DukeMTMC-reID large-scale datasets.
9.SeFM: A Sequential Feature Point Matching Algorithm for Object 3D Reconstruction pdf
3D reconstruction is a fundamental issue in many applications and the feature point matching problem is a key step while reconstructing target objects. Conventional algorithms can only find a small number of feature points from two images which is quite insufficient for reconstruction. To overcome this problem, we propose SeFM a sequential feature point matching algorithm. We first utilize the epipolar geometry to find the epipole of each image. Rotating along the epipole, we generate a set of the epipolar lines and reserve those intersecting with the input image. Next, a rough matching phase, followed by a dense matching phase, is applied to find the matching dot-pairs using dynamic programming. Furthermore, we also remove wrong matching dot-pairs by calculating the validity. Experimental results illustrate that SeFM can achieve around 1,000 to 10,000 times matching dot-pairs, depending on individual image, compared to conventional algorithms and the object reconstruction with only two images is semantically visible. Moreover, it outperforms conventional algorithms, such as SIFT and SURF, regarding precision and recall.
10.Deep Energies for Estimating Three-Dimensional Facial Pose and Expression pdf
While much progress has been made in capturing high-quality facial performances using motion capture markers and shape-from-shading, high-end systems typically also rely on rotoscope curves hand-drawn on the image. These curves are subjective and difficult to draw consistently; moreover, ad-hoc procedural methods are required for generating matching rotoscope curves on synthetic renders embedded in the optimization used to determine three-dimensional facial pose and expression. We propose an alternative approach whereby these curves and other keypoints are detected automatically on both the image and the synthetic renders using trained neural networks, eliminating artist subjectivity and the ad-hoc procedures meant to mimic it. More generally, we propose using machine learning networks to implicitly define deep energies which when minimized using classical optimization techniques lead to three-dimensional facial pose and expression estimation.
11.TDAN: Temporally Deformable Alignment Network for Video Super-Resolution pdf
Video super-resolution (VSR) aims to restore a photo-realistic high-resolution (HR) video frame from both its corresponding low-resolution (LR) frame (reference frame) and multiple neighboring frames (supporting frames). Due to varying motion of cameras or objects, the reference frame and each support frame are not aligned. Therefore, temporal alignment is a challenging yet important problem for VSR. Previous VSR methods usually utilize optical flow between the reference frame and each supporting frame to wrap the supporting frame for temporal alignment. Therefore, the performance of these image-level wrapping-based models will highly depend on the prediction accuracy of optical flow, and inaccurate optical flow will lead to artifacts in the wrapped supporting frames, which also will be propagated into the reconstructed HR video frame. To overcome the limitation, in this paper, we propose a temporal deformable alignment network (TDAN) to adaptively align the reference frame and each supporting frame at the feature level without computing optical flow. The TDAN uses features from both the reference frame and each supporting frame to dynamically predict offsets of sampling convolution kernels. By using the corresponding kernels, TDAN transforms supporting frames to align with the reference frame. To predict the HR video frame, a reconstruction network taking aligned frames and the reference frame is utilized. Experimental results demonstrate the effectiveness of the proposed TDAN-based VSR model.
12.Improved Search Strategies for Determining Facial Expression pdf
It is well known that popular optimization techniques can lead to overfitting or even a lack of convergence altogether; thus, practitioners often utilize ad hoc regularization terms added to the energy functional. When carefully crafted, these regularizations can produce compelling results. However, regularization changes both the energy landscape and the solution to the optimization problem, which can result in underfitting. Surprisingly, many practitioners both add regularization and claim that their model lacks the expressivity to fit the data. Motivated by a geometric interpretation of the linearized search space, we propose an approach that ameliorates overfitting without the need for regularization terms that restrict the expressiveness of the underlying model. We illustrate the efficacy of our approach on minimization problems related to three-dimensional facial expression estimation where overfitting clouds semantic understanding and regularization may lead to underfitting that misses or misinterprets subtle expressions.
13.Star Tracking using an Event Camera pdf
Star trackers are primarily optical devices that are used to estimate the attitude of a spacecraft by recognising and tracking star patterns. Currently, most star trackers use conventional optical sensors. In this application paper, we propose the usage of event sensors for star tracking. There are potentially two benefits of using event sensors for star tracking: lower power consumption and higher operating speeds. Our main contribution is to formulate an algorithmic pipeline for star tracking from event data that includes novel formulations of rotation averaging and bundle adjustment. In addition, we also release with this paper a dataset for star tracking using event cameras. With this work, we introduce the problem of star tracking using event cameras to the computer vision community, whose expertise in SLAM and geometric optimisation can be brought to bear on this commercially important application.
14.Adversarial Defense of Image Classification Using a Variational Auto-Encoder pdf
Deep neural networks are known to be vulnerable to adversarial attacks. This exposes them to potential exploits in security-sensitive applications and highlights their lack of robustness. This paper uses a variational auto-encoder (VAE) to defend against adversarial attacks for image classification tasks. This VAE defense has a few nice properties: (1) it is quite flexible and its use of randomness makes it harder to attack; (2) it can learn disentangled representations that prevent blurry reconstruction; and (3) a patch-wise VAE defense strategy is used that does not require retraining for different size images. For moderate to severe attacks, this system outperforms or closely matches the performance of JPEG compression, with the best quality parameter. It also has more flexibility and potential for improvement via training.
15.An Attempt towards Interpretable Audio-Visual Video Captioning pdf
Automatically generating a natural language sentence to describe the content of an input video is a very challenging problem. It is an essential multimodal task in which auditory and visual contents are equally important. Although audio information has been exploited to improve video captioning in previous works, it is usually regarded as an additional feature fed into a black box fusion machine. How are the words in the generated sentences associated with the auditory and visual modalities? The problem is still not investigated. In this paper, we make the first attempt to design an interpretable audio-visual video captioning network to discover the association between words in sentences and audio-visual sequences. To achieve this, we propose a multimodal convolutional neural network-based audio-visual video captioning framework and introduce a modality-aware module for exploring modality selection during sentence generation. Besides, we collect new audio captioning and visual captioning datasets for further exploring the interactions between auditory and visual modalities for high-level video understanding. Extensive experiments demonstrate that the modality-aware module makes our model interpretable on modality selection during sentence generation. Even with the added interpretability, our video captioning network can still achieve comparable performance with recent state-of-the-art methods.
16.Towards Hiding Adversarial Examples from Network Interpretation pdf
Deep networks have been shown to be fooled rather easily using adversarial attack algorithms. Practical methods such as adversarial patches have been shown to be extremely effective in causing misclassification. However, these patches can be highlighted using standard network interpretation algorithms, thus revealing the identity of the adversary. We show that it is possible to create adversarial patches which not only fool the prediction, but also change what we interpret regarding the cause of prediction. We show that our algorithms can empower adversarial patches, by hiding them from network interpretation tools. We believe our algorithms can facilitate developing more robust network interpretation tools that truly explain the network's underlying decision making process.
17.High-Quality Face Capture Using Anatomical Muscles pdf
Muscle-based systems have the potential to provide both anatomical accuracy and semantic interpretability as compared to blendshape models; however, a lack of expressivity and differentiability has limited their impact. Thus, we propose modifying a recently developed rather expressive muscle-based system in order to make it fully-differentiable; in fact, our proposed modifications allow this physically robust and anatomically accurate muscle model to conveniently be driven by an underlying blendshape basis. Our formulation is intuitive, natural, as well as monolithically and fully coupled such that one can differentiate the model from end to end, which makes it viable for both optimization and learning-based approaches for a variety of applications. We illustrate this with a number of examples including both shape matching of three-dimensional geometry as as well as the automatic determination of a three-dimensional facial pose from a single two-dimensional RGB image without using markers or depth information.
18.Neural Image Decompression: Learning to Render Better Image Previews pdf
A rapidly increasing portion of Internet traffic is dominated by requests from mobile devices with limited- and metered-bandwidth constraints. To satisfy these requests, it has become standard practice for websites to transmit small and extremely compressed image previews as part of the initial page-load process. Recent work, based on an adaptive triangulation of the target image, has shown the ability to generate thumbnails of full images at extreme compression rates: 200 bytes or less with impressive gains (in terms of PSNR and SSIM) over both JPEG and WebP standards. However, qualitative assessments and preservation of semantic content can be less favorable. We present a novel method to significantly improve the reconstruction quality of the original image with no changes to the encoded information. Our neural-based decoding not only achieves higher PSNR and SSIM scores than the original methods, but also yields a substantial increase in semantic-level content preservation. In addition, by keeping the same encoding stream, our solution is completely inter-operable with the original decoder. The end result is suitable for a range of small-device deployments, as it involves only a single forward-pass through a small, scalable network.
19.Tri-axial Self-Attention for Concurrent Activity Recognition pdf
We present a system for concurrent activity recognition. To extract features associated with different activities, we propose a feature-to-activity attention that maps the extracted global features to sub-features associated with individual activities. To model the temporal associations of individual activities, we propose a transformer-network encoder that models independent temporal associations for each activity. To make the concurrent activity prediction aware of the potential associations between activities, we propose self-attention with an association mask. Our system achieved state-of-the-art or comparable performance on three commonly used concurrent activity detection datasets. Our visualizations demonstrate that our system is able to locate the important spatial-temporal features for final decision making. We also showed that our system can be applied to general multilabel classification problems.
20.StoryGAN: A Sequential Conditional GAN for Story Visualization pdf
In this work we propose a new task called Story Visualization. Given a multi-sentence paragraph, the story is visualized by generating a sequence of images, one for each sentence. In contrast to video generation, story visualization focuses less on the continuity in generated images (frames), but more on the global consistency across dynamic scenes and characters -- a challenge that has not been addressed by any single-image or video generation methods. Therefore, we propose a new story-to-image-sequence generation model, StoryGAN, based on the sequential conditional GAN framework. Our model is unique in that it consists of a deep Context Encoder that dynamically tracks the story flow, and two discriminators at the story and image levels, respectively, to enhance the image quality and the consistency of the generated sequences. To evaluate the model, we modified existing datasets to create the CLEVR-SV and Pororo-SV datasets. Empirically, StoryGAN outperformed state-of-the-art models in image quality, contextual consistency metrics, and human evaluation.
21.ROI-10D: Monocular Lifting of 2D Detection to 6D Pose and Metric Shape pdf
We present a deep learning method for end-to-end monocular 3D object detection and metric shape retrieval. We propose a novel loss formulation by lifting 2D detection, orientation, and scale estimation into 3D space. Instead of optimizing these quantities separately, the 3D instantiation allows to properly measure the metric misalignment of boxes. We experimentally show that our 10D lifting of sparse 2D Regions of Interests (RoIs) achieves great results both for 6D pose and recovery of the textured metric geometry of instances. This further enables 3D synthetic data augmentation via inpainting recovered meshes directly onto the 2D scenes. We evaluate on KITTI3D against other strong monocular methods and demonstrate that our approach doubles the AP on the 3D pose metrics on the official test set, defining the new state of the art.
22.Object Discovery in Videos as Foreground Motion Clustering pdf
We consider the problem of providing dense segmentation masks for object discovery in videos. We formulate the object discovery problem as foreground motion clustering, where the goal is to cluster foreground pixels in videos into different objects. We introduce a novel pixel-trajectory recurrent neural network that learns feature embeddings of foreground pixel trajectories linked in time. By clustering the pixel trajectories using the learned feature embeddings, our method establishes correspondences between foreground object masks across video frames. To demonstrate the effectiveness of our framework for object discovery, we conduct experiments on commonly used datasets for motion segmentation, where we achieve state-of-the-art performance.
23.Neural Word Search in Historical Manuscript Collections pdf
We address the problem of segmenting and retrieving word images in collections of historical manuscripts given a text query. This is commonly referred to as "word spotting". To this end, we first propose an end-to-end trainable model based on deep neural networks that we dub Ctrl-F-Net. The model simultaneously generates region proposals and embeds them into a word embedding space, wherein a search is performed. We further introduce a simplified version called Ctrl-F-Mini. It is faster with similar performance, though it is limited to more easily segmented manuscripts. We evaluate both models on common benchmark datasets and surpass the previous state of the art. Finally, in collaboration with historians, we employ the Ctrl-F-Net to search within a large manuscript collection of over 100 thousand pages, written across two centuries. With only 11 training pages, we enable large scale data collection in manuscript-based historical research. This results in a speed up of data collection and the number of manuscripts processed by orders of magnitude. Given the time consuming manual work required to study old manuscripts in the humanities, quick and robust tools for word spotting has the potential to revolutionise domains like history, religion and language.
24.Knockoff Nets: Stealing Functionality of Black-Box Models pdf
Machine Learning (ML) models are increasingly deployed in the wild to perform a wide range of tasks. In this work, we ask to what extent can an adversary steal functionality of such "victim" models based solely on blackbox interactions: image in, predictions out. In contrast to prior work, we present an adversary lacking knowledge of train/test data used by the model, its internals, and semantics over model outputs. We formulate model functionality stealing as a two-step approach: (i) querying a set of input images to the blackbox model to obtain predictions; and (ii) training a "knockoff" with queried image-prediction pairs. We make multiple remarkable observations: (a) querying random images from a different distribution than that of the blackbox training data results in a well-performing knockoff; (b) this is possible even when the knockoff is represented using a different architecture; and (c) our reinforcement learning approach additionally improves query sample efficiency in certain settings and provides performance gains. We validate model functionality stealing on a range of datasets and tasks, as well as on a popular image analysis API where we create a reasonable knockoff for as little as $30.
25.ChauffeurNet: Learning to Drive by Imitating the Best and Synthesizing the Worst pdf
Our goal is to train a policy for autonomous driving via imitation learning that is robust enough to drive a real vehicle. We find that standard behavior cloning is insufficient for handling complex driving scenarios, even when we leverage a perception system for preprocessing the input and a controller for executing the output on the car: 30 million examples are still not enough. We propose exposing the learner to synthesized data in the form of perturbations to the expert's driving, which creates interesting situations such as collisions and/or going off the road. Rather than purely imitating all data, we augment the imitation loss with additional losses that penalize undesirable events and encourage progress -- the perturbations then provide an important signal for these losses and lead to robustness of the learned model. We show that the ChauffeurNet model can handle complex situations in simulation, and present ablation experiments that emphasize the importance of each of our proposed changes and show that the model is responding to the appropriate causal factors. Finally, we demonstrate the model driving a car in the real world.
26.A High-Order Scheme for Image Segmentation via a modified Level-Set method pdf
The method is based on an adaptive "filtered" scheme recently introduced by the authors. The main feature of the scheme is the possibility to stabilize an a priori unstable high-order scheme via a filter function which allows to combine a high-order scheme in the regularity regions and a monotone scheme elsewhere, in presence of singularities. The filtered scheme considered in this paper uses the local Lax-Friedrichs scheme as monotone scheme and the Lax-Wendroff scheme as high-order scheme but other couplings are possible. Moreover, we introduce also a modified velocity function for the level-set model used in segmentation, this velocity allows to obtain more accurate results with respect to other velocities proposed in the literature. Some numerical tests on synthetic and real images confirm the accuracy of the proposed method and the advantages given by the new velocity.
27.Back to square one: probabilistic trajectory forecasting without bells and whistles pdf
We introduce a spatio-temporal convolutional neural network model for trajectory forecasting from visual sources. Applied in an auto-regressive way it provides an explicit probability distribution over continuations of a given initial trajectory segment. We discuss it in relation to (more complicated) existing work and report on experiments on two standard datasets for trajectory forecasting: MNISTseq and Stanford Drones, achieving results on-par with or better than previous methods.
28.LNEMLC: Label Network Embeddings for Multi-Label Classifiation pdf
Multi-label classification aims to classify instances with discrete non-exclusive labels. Most approaches on multi-label classification focus on effective adaptation or transformation of existing binary and multi-class learning approaches but fail in modelling the joint probability of labels or do not preserve generalization abilities for unseen label combinations. To address these issues we propose a new multi-label classification scheme, LNEMLC - Label Network Embedding for Multi-Label Classification, that embeds the label network and uses it to extend input space in learning and inference of any base multi-label classifier. The approach allows capturing of labels' joint probability at low computational complexity providing results comparable to the best methods reported in the literature. We demonstrate how the method reveals statistically significant improvements over the simple kNN baseline classifier. We also provide hints for selecting the robust configuration that works satisfactorily across data domains.
29.Privacy Partitioning: Protecting User Data During the Deep Learning Inference Phase pdf
We present a practical method for protecting data during the inference phase of deep learning based on bipartite topology threat modeling and an interactive adversarial deep network construction. We term this approach \emph{Privacy Partitioning}. In the proposed framework, we split the machine learning models and deploy a few layers into users' local devices, and the rest of the layers into a remote server. We propose an approach to protect user's data during the inference phase, while still achieve good classification accuracy.
We conduct an experimental evaluation of this approach on benchmark datasets of three computer vision tasks. The experimental results indicate that this approach can be used to significantly attenuate the capacity for an adversary with access to the state-of-the-art deep network's intermediate states to learn privacy-sensitive inputs to the network. For example, we demonstrate that our approach can prevent attackers from inferring the private attributes such as gender from the Face image dataset without sacrificing the classification accuracy of the original machine learning task such as Face Identification.
30.Learning Implicit Fields for Generative Shape Modeling pdf
We advocate the use of implicit fields for learning generative models of shapes and introduce an implicit field decoder for shape generation, aimed at improving the visual quality of the generated shapes. An implicit field assigns a value to each point in 3D space, so that a shape can be extracted as an iso-surface. Our implicit field decoder is trained to perform this assignment by means of a binary classifier. Specifically, it takes a point coordinate, along with a feature vector encoding a shape, and outputs a value which indicates whether the point is outside the shape or not. By replacing conventional decoders by our decoder for representation learning and generative modeling of shapes, we demonstrate superior results for tasks such as shape autoencoding, generation, interpolation, and single-view 3D reconstruction, particularly in terms of visual quality.