CVPR 2016 Papers

In this paper, we propose training very deep neural networks (DNNs) for supervised learning of hash codes. Existing methods in this context train relatively "shallow" networks limited by the issues arising in back propagation (e.g. vanishing gradients) as well as computational efficiency. We propose a novel and efficient training algorithm inspired by alternating direction method of multipliers (ADMM) that overcomes some of these limitations. Our method decomposes the training process into independent layer-wise local updates through auxiliary variables. Empirically we observe that our training algorithm always converges and its computational complexity is linearly proportional to the number of edges in the networks. Empirically we manage to train DNNs with 64 hidden layers and 1024 nodes per layer for supervised hashing in about 3 hours using a single GPU. Our proposed very deep supervised hashing (VDSH) method significantly outperforms the state-of-the-art on several benchmark datasets.

In this paper we propose a new framework to compare and classify temporal sequences. The proposed approach captures the underlying dynamics of the data while avoiding expensive estimation procedures, making it suitable to process large numbers of sequences. The main idea is to first embed the sequences into a Riemannian manifold by using positive definite regularized Gram matrices of their Hankelets. The advantages of the this approach are: 1) it allows for using non-Euclidean similarity functions on the Positive Definite matrix manifold, which capture better the underlying geometry than directly comparing the sequences or their Hankel matrices; and 2) Gram matrices inherit desirable properties from the underlying Hankel matrices: their rank measure the complexity of the underlying dynamics, and the rank and the coefficients of the associated regressive models are invariant to affine transformations and varying initial conditions. The benefits of this approach are illustrated with extensive experiments in 3D action recognition using 3D joints sequences. In spite of its simplicity, the performance of this approach is competitive or better than using state-of-art approaches for this problem. Further, these results hold across a variety of metrics, supporting the idea that the improvement stems from the embedding itself, rather than from using one of these metrics.

We tackle the problem of large-scale object detection in images, where the number of objects can be arbitrarily large, and can exhibit significant overlap/occlusion. A successful approach to modelling the large-scale nature of this problem has been via point process density functions which jointly encode object qualities and spatial interactions. But the corresponding optimisation problem is typically difficult or intractable, and many of the best current methods rely on Monte Carlo Markov Chain (MCMC) simulation, which converges slowly in a large solution space. We propose an efficient point process inference for large-scale object detection using discrete energy minimization. In particular, we approximate the solution space by a finite set of object proposals and cast the point process density function to a corresponding energy function of binary variables whose values indicate which object proposals are accepted. We resort to the local submodular approximation (LSA) based trust-region optimisation to find the optimal solution. Furthermore we analyse the error of LSA approximation, and show how to adjust the point process energy to dramatically speed up the convergence without harms in the optimality. We demonstrate the superior efficiency and accuracy of our method using a variety of large-scale object detection applications such as crowd human detection, birds, cells counting/localization.

We consider the image retrieval problem of finding the images in a dataset that are most similar to a query image. Our goal is to reduce the number of vector operations and memory for performing a search without sacrificing accuracy of the returned images. We adopt a group testing formulation and design the decoding architecture using either dictionary learning or eigendecomposition. The latter is a plausible option for small-to-medium sized problems with high-dimensional global image descriptors, whereas dictionary learning is applicable in large-scale scenarios. We evaluate our approach for global descriptors obtained from both SIFT and CNN features. Experiments with standard image search benchmarks, including the Yahoo100M dataset comprising 100 million images, show that our method gives comparable (and sometimes superior) accuracy compared to exhaustive search while requiring only 10% of the vector operations and memory. Moreover, for the same search complexity, our method gives significantly better accuracy compared to approaches based on dimensionality reduction or locality sensitive hashing.

We present a new approach for efficient approximate nearest neighbor (ANN) search in high dimensional spaces, extending the idea of Product Quantization. We propose a two level product and vector quantization tree that reduces the number of vector comparisons required during tree traversal. Our approach also includes a novel highly parallelizable re-ranking method for candidate vectors by efficiently reusing already computed intermediate values. Due to its small memory footprint during traversal the method lends itself to an efficient, parallel GPU implementation. This Product Quantization Tree approach significantly outperforms recent state of the art methods for high dimensional nearest neighbor queries on standard reference datasets. Ours is the first work that demonstrates GPU performance superior to CPU performance on high dimensional, large scale ANN problems in time-critical real-world applications, like loop-closing in videos.

In this paper, we present a new algorithm for finding all intersections of three quadrics. The proposed method is algebraic in nature and it is considerably more efficient than the Groebner basis and resultant-based solutions previously used in computer vision applications. We identify several computer vision problems that are formulated and solved as systems of three quadratic equations and for which our algorithm readily delivers considerably faster results. Also, we propose new formulations of three important vision problems: absolute camera pose with unknown focal length, generalized pose-and-scale, and hand-eye calibration with known translation. These new formulations allow our algorithm to significantly outperform the state-of-the-art in speed.

Existing billion-scale nearest neighbor search systems have mostly been compared on a single dataset of a billion of SIFT vectors, where systems based on the Inverted Multi-Index (IMI) have been performing very well, achieving state-of-the-art recall in several milliseconds. SIFT-like descriptors, however, are quickly being replaced with descriptors based on deep neural networks (DNN) that provide better performance for many computer vision tasks. In this paper, we introduce a new dataset of one billion descriptors based on DNNs and reveal the relative inefficiency of IMI-based indexing for such descriptors compared to SIFT data. We then introduce two new indexing structures, the Non-Orthogonal Inverted Multi-Index (NO-IMI) and the Generalized Non-Orthogonal Inverted Multi-Index (GNO-IMI). We show that due to additional flexibility, the new structures are able to adapt to DNN descriptor distribution in a better way. In particular, extensive experiments on the new dataset demonstrate that these data structures provide considerably better trade-off between the speed of retrieval and recall, given similar amount of memory, as compared to the standard Inverted Multi-Index.

We propose the first algorithm for non-rigid 2D-to-3D shape matching, where the input is a 2D query shape as well as a 3D target shape and the output is a continuous matching curve represented as a closed contour on the 3D shape. We cast the problem as finding the shortest circular path on the product 3-manifold of the two shapes. We prove that the optimal matching can be computed in polynomial time with a (worst-case) complexity of O(m*n^2*log(n)), where m and n denote the number of vertices on the 2D and the 3D shape respectively. Quantitative evaluation confirms that the method provides excellent results for sketch-based deformable 3D shape retrieval.

During a long period of time we are combating over-fitting in the CNN training process with model regularization, including weight decay, model averaging, data augmentation, etc. In this paper, we present DisturbLabel, an extremely simple algorithm which randomly replaces a part of labels as incorrect values in each iteration. Although it seems weird to intentionally generate incorrect training labels, we show that DisturbLabel prevents the network training from over-fitting by implicitly averaging over exponentially many networks which are trained with different label sets. To the best of our knowledge, DisturbLabel serves as the first work which adds noises on the loss layer. Meanwhile, DisturbLabel cooperates well with Dropout to provide complementary regularization functions. Experiments demonstrate competitive recognition results on several popular image recognition datasets.

RGBD scene recognition has attracted increasingly attention due to the rapid development of depth sensors and their wide application scenarios. While many research has been conducted, most work used hand-crafted features which are difficult to capture high-level semantic structures. Recently, the feature extracted from deep convolutional neural network has produced state-of-the-art results for various computer vision tasks, which inspire researchers to explore incorporating CNN learned features for RGBD scene understanding. On the other hand, most existing work combines rgb and depth features without adequately exploiting the consistency and complementary information between them. Inspired by some recent work on RGBD object recognition using multi-modal feature fusion, we introduce a novel discriminative multi-modal fusion framework for rgbd scene recognition for the first time which simultaneously considers the inter- and intra-modality correlation for all samples and meanwhile regularizing the learned features to be discriminative and compact. The results from the multimodal layer can be back-propagated to the lower CNN layers, hence the parameters of the CNN layers and multimodal layers are updated iteratively until convergence. Experiments on the recently proposed large scale SUN RGB-D datasets show that our method achieved the state-of-the-art without any image segmentation.

We present hierarchical rank pooling, a video sequence encoding method for activity recognition. It consists of a network of rank pooling functions which captures the dynamics of rich convolutional neural network features within a video sequence. By stacking non-linear feature functions and rank pooling over one another, we obtain a high capacity dynamic encoding mechanism, which is used for action recognition. We present a method for jointly learning the video representation and activity classifier parameters. Our method obtains state-of-the art results on three important activity recognition benchmarks: 76.7% on Hollywood2, 66.9% on HMDB51 and, 91.4% on UCF101.

We propose a motion-based method to discover the physical parts of an articulated object class (e.g. head/torso/leg of a horse) from multiple videos. The key is to find object regions that exhibit consistent motion relative to the rest of the object, across multiple videos. We can then learn a location model for the parts and segment them accurately in the individual videos using an energy function that also enforces temporal and spatial consistency in part motion. Unlike our approach, traditional methods for motion segmentation or non-rigid structure from motion operate on one video at a time. Hence they cannot discover a part unless it displays independent motion in that particular video. We evaluate our method on a new dataset of 32 videos of tigers and horses, where we significantly outperform a recent motion segmentation method on the task of part discovery (obtaining roughly twice the accuracy).

We propose an efficient approach to exploiting motion information from consecutive frames of a video sequence to recover the 3D pose of people. Previous approaches typically compute candidate poses in individual frames and then link them in a post-processing step to resolve ambiguities. By contrast, we directly regress from a spatio-temporal volume of bounding boxes to a 3D pose in the central frame. We further show that, for this approach to achieve its full potential, it is essential to compensate for the motion in consecutive frames so that the subject remains centered. This then allows us to effectively overcome ambiguities and improve upon the state-of-the-art by a large margin on the Human3.6m, HumanEva, and KTH Multiview Football 3D human pose estimation benchmarks.

Feature representation and object category classification are two key components of most object detection methods. While significant improvements have been achieved for deep feature representation learning, traditional SVM/softmax classifiers remain the dominant methods for final object category classification. However, SVM/softmax classifiers lack the capacity of explicitly exploiting the complex structure of deep features, as they are purely discriminative methods. The recently proposed discriminative dictionary pair learning (DPL) model involves a fidelity term to minimize the reconstruction loss and a discrimination term to enhance the discriminative capability of the learned dictionary pair, and thus is appropriate for balancing the representation and discrimination to boost object detection performance. In this paper, we propose a novel object detection system by unifying DPL with the convolutional feature learning. Specifically, we incorporate DPL as a Dictionary Pair Classifier Layer (DPCL) into the deep architecture, and develop an end-to-end learning algorithm for optimizing the dictionary pairs and the neural networks simultaneously. Moreover, we design a multi-task loss for guiding our model to accomplish the three correlated tasks: objectness estimation, categoryness computation, and bounding box regression. From the extensive experiments on PASCAL VOC 2007/2012 benchmarks, our approach demonstrates the effectiveness to substantially improve the performances over the popular existing object detection frameworks (e.g., R-CNN [13] and FRCN [12]), and achieves new state-of-the-arts.

The problem of localizing occlusions between consecutive frames of a video is important but rarely tackled on its own. In most works, it is tightly interleaved with the computation of accurate optical flows, which leads to a delicate chicken-and-egg problem. With this in mind, we propose a novel approach to occlusion detection where visibility or not of a point in next frame is formulated in terms of visual reconstruction. The key issue is now to determine how well a pixel in the first image can be "recon- structed" from co-located colors in the next image. We first exploit this reasoning at the pixel level with a new detection criterion. Contrary to the ubiquitous displaced-frame-difference and forward-backward flow vector matching, the proposed alternative does not critically depend on a precomputed, dense displacement field, while being shown to be more effective. We then leverage this local modeling within an energy-minimization framework that delivers occlusion maps. An easy-to-obtain collection of parametric motion models is exploited within the energy to provide the required level of motion information. Our approach outperforms state-of-the-art detection methods on the challenging MPI Sintel dataset.

Using fiducial markers ensures reliable detection and identification of planar features in images. Fiducials are used in a wide range of applications, especially when a reliable visual reference is needed, e.g., to track the camera in cluttered or textureless environments. A marker designed for such applications must be robust to partial occlusions, varying distances and angles of view, and fast camera motions. In this paper, we present a robust, highly accurate fiducial system, whose markers consist of concentric rings, along with its theoretical foundations. Relying on projective properties, it allows to robustly localize the imaged marker and to accurately detect the position of the image of the (common) circle center. We demonstrate that our system can detect and accurately localize these circular fiducials under very challenging conditions and the experimental results reveal that it outperforms other recent fiducial systems.

In this paper we describe a new method for detecting and counting a repeating object in an image. While the method relies on a fairly sophisticated deformable part model, unlike existing techniques it estimates the model parameters in an unsupervised fashion thus alleviating the need for a user-annotated training data and avoiding the associated specificity. This automatic fitting process is carried out by exploiting the recurrence of small image patches associated with the repeating object and analyzing their spatial correlation. The analysis allows us to reject outlier patches, recover the visual and shape parameters of the part model, and detect the object instances efficiently. In order to achieve a practical system which is able to cope with diverse images, we describe a simple and intuitive active-learning procedure that updates the object classification by querying the user on very few carefully chosen marginal classifications. Evaluation of the new method against the state-of-the-art techniques demonstrates its ability to achieve higher accuracy through a better user experience.

Bird migration is a critical indicator of environmental health, biodiversity, and climate change. Existing techniques for monitoring bird migration are either expensive (e.g., satellite tracking), labor-intensive (e.g., moon watching), indirect and thus less accurate (e.g., weather radar), or intrusive (e.g., attaching geolocators on captured birds). In this paper, we present a vision-based system for detecting migrating birds in flight at night. Our system takes stereo videos of the night sky as inputs, detects multiple flying birds and estimates their orientations, speeds, and altitudes. The main challenge lies in detecting flying birds of unknown trajectories under high noise level due to the low-light environment. We address this problem by incorporating stereo constraints for rejecting physically implausible configurations and gathering evidence from two (or more) views. Specifically, we develop a robust stereo-based 3D line fitting algorithm for geometric verification and a deformable part response accumulation strategy for trajectory verification. We demonstrate the effectiveness of the proposed approach through quantitative evaluation of real videos of birds migrating at night collected with near-infrared cameras.

In this work, we propose a multi-camera system where we combine a main high-quality camera with two low-res auxiliary cameras. The auxiliary cameras are well calibrated and act as a passive depth sensor by generating disparity maps. The main camera has an interchangeable lens and can produce good quality images at high resolution. Our goal is, given the low-res depth map from the auxiliary cameras, generate a depth map from the viewpoint of the main camera. The advantage of our system, compared to other systems such as light-field cameras or RGBD sensors, is the ability to generate a high-resolution color image with a complete depth map, without sacrificing resolution and with minimal auxiliary hardware. Since the main camera has an interchangeable lens, it cannot be calibrated beforehand, and directly applying stereo matching on it and either of the auxiliary cameras often leads to unsatisfactory results. Utilizing both the calibrated cameras at once, we propose a novel approach to better estimate the disparity map of the main camera. Then by combining the defocus cue of the main camera, the disparity map can be further improved. We demonstrate the performance of our algorithm on various scenes.