CVPR 2016 Papers

The way people look in terms of facial attributes (ethnicity, hair color, facial hair, etc.) and the clothes or accessories they wear (sunglasses, hat, hoodies, etc.) is highly dependent on geo-location and weather condition, respectively. This work explores, for the first time, the use of this contextual information, as people with wearable cameras walk across different neighborhoods of a city, in order to learn a rich feature representation for facial attribute classification, without the costly manual annotation required by previous methods. By tracking the faces of casual walkers on more than 40 hours of egocentric video, we are able to cover tens of thousands of different identities and automatically extract nearly 5 million pairs of images connected by or from different face tracks, along with their weather and location context, under pose and lighting variations. These image pairs are then fed into a deep network that preserves similarity of images connected by the same track, in order to capture identity-related attribute features, and optimizes for location and weather prediction to capture additional facial attribute features. Finally, the network is fine-tuned with manually annotated samples. We perform an extensive experimental analysis on wearable data and two standard benchmark datasets based on web images (LFWA and CelebA). Our method outperforms by a large margin a network trained from scratch. Moreover, even without using manually annotated identity labels for pre-training as in previous methods, our approach achieves results that are better than the state of the art.

Face detection is one of the most studied topics in the computer vision community. Much of the progresses have been made by the availability of face detection benchmark datasets. We show that there is a gap between current face detection performance and the real world requirements. To facilitate future face detection research, we introduce the WIDER FACE dataset, which is 10 times larger than existing datasets. The dataset contains rich annotations, including occlusions, poses, event categories, and face bounding boxes. Faces in the proposed dataset are extremely challenging due to large variations in scale, pose and occlusion, as shown in Fig. 1. Furthermore, we show that WIDER FACE dataset is an effective training source for face detection. We benchmark several representative detection systems, providing an overview of state-of-the-art performance and propose a solution to deal with large scale variation. Finally, we discuss common failure cases that worth to be further investigated.

Objects make distinctive sounds when they are hit or scratched. These sounds reveal aspects of an object's material properties, as well as the actions that produced them. In this paper, we propose the task of predicting what sound an object makes when struck as a way of studying physical interactions within a visual scene. We present an algorithm that synthesizes sound from silent videos of people hitting and scratching objects with a drumstick. This algorithm uses a recurrent neural network to predict sound features from videos and then produces a waveform from these features with an example-based synthesis procedure. We show that the sounds predicted by our model are realistic enough to fool participants in a "real or fake" psychophysical experiment, and that they convey significant information about material properties and physical interactions.

We present an approach that exploits hierarchical Recurrent Neural Networks (RNNs) to tackle the video captioning problem, i.e., generating one or multiple sentences to describe a realistic video. Our hierarchical framework contains a sentence generator and a paragraph generator. The sentence generator produces one simple short sentence that describes a specific short video interval. It exploits both temporal- and spatial-attention mechanisms to selectively focus on visual elements during generation. The paragraph generator captures the inter-sentence dependency by taking as input the sentential embedding produced by the sentence generator, combining it with the paragraph history, and outputting the new initial state for the sentence generator. We evaluate our approach on two large-scale benchmark datasets: YouTubeClips and TACoS-MultiLevel. The experiments demonstrate that our approach significantly outperforms the current state-of-the-art methods with BLEU@4 scores 0.499 and 0.305 respectively.

Data-driven approaches for edge detection have proven effective and achieve top results on modern benchmarks. However, all current data-driven edge detectors require manual supervision for training in the form of hand-labeled region segments or object boundaries. Specifically, human annotators mark semantically meaningful edges which are subsequently used for training. Is this form of strong, high-level supervision actually necessary to learn to accurately detect edges? In this work we present a simple yet effective approach for training edge detectors without human supervision. To this end we utilize motion, and more specifically, the only input to our method is noisy semi-dense matches between frames. We begin with only a rudimentary knowledge of edges (in the form of image gradients), and alternate between improving motion estimation and edge detection in turn. Using a large corpus of video data, we show that edge detectors trained using our unsupervised scheme approach the performance of the same methods trained with full supervision (within 3-5%). Finally, we show that when using a deep network for the edge detector, our approach provides a novel pre-training scheme for object detection.

We address the problem of automatically learning the main steps to complete a certain task, such as changing a car tire, from a set of narrated instruction videos. The contributions of this paper are three-fold. First, we develop a new unsupervised learning approach that takes advantage of the complementary nature of the input video and the associated narration. The method solves two clustering problems, one in text and one in video, applied one after each other and linked by joint constraints to obtain a single coherent sequence of steps in both modalities. Second, we collect and annotate a new challenging dataset of real-world instruction videos from the Internet. The dataset contains about 800,000 frames for five different tasks that include complex interactions between people and objects, and are captured in a variety of indoor and outdoor settings. Third, we experimentally demonstrate that the proposed method can automatically discover, in an unsupervised manner, the main steps to achieve the task and locate the steps in the input videos.

The field of object detection has made significant advances riding on the wave of region-based ConvNets, but their training procedure still includes many heuristics and hyperparameters that are costly to tune. We present a simple yet surprisingly effective online hard example mining (OHEM) algorithm for training region-based ConvNet detectors. Our motivation is the same as it has always been -- detection datasets contain an overwhelming number of easy examples and a small number of hard examples. Automatic selection of these hard examples can make training more effective and efficient. OHEM is a simple and intuitive algorithm that eliminates several heuristics and hyperparameters in common use. But more importantly, it yields consistent and significant boosts in detection performance on benchmarks like PASCAL VOC 2007 and 2012. Its effectiveness increases as datasets become larger and more difficult, as demonstrated by the results on the MS COCO dataset. Moreover, combined with complementary advances in the field, OHEM leads to state-of-the-art results of 78.9% and 76.3% mAP on PASCAL VOC 2007 and 2012 respectively.

We develop new representations and algorithms for three-dimensional (3D) object detection and spatial layout prediction in cluttered indoor scenes. RGB-D images are traditionally described by local geometric features of the 3D point cloud. We propose a cloud of oriented gradient (COG) descriptor that links the 2D appearance and 3D pose of object categories, and thus accurately models how perspective projection affects perceived image boundaries. We also propose a "Manhattan voxel" representation which better captures the 3D room layout geometry of common indoor environments. Effective classification rules are learned via a structured prediction framework that accounts for the intersection-over-union overlap of hypothesized 3D cuboids with human annotations, as well as orientation estimation errors. Contextual relationships among categories and layout are captured via a cascade of classifiers, leading to holistic scene hypotheses with improved accuracy. Our model is learned solely from annotated RGB-D images, without the benefit of CAD models, but nevertheless its performance substantially exceeds the state-of-the-art on the SUN RGB-D database. Avoiding CAD models allows easier learning of detectors for many object categories.

This paper proposes a novel extremely efficient, fully-parallelizable, task-specific algorithm for the computation of global point-wise correspondences in images and videos. Our algorithm, the Global Patch Collider, is based on detecting unique collisions between image points using a collection of learned tree structures that act as conditional hash functions. In contrast to conventional approaches that rely on pairwise distance computation, our algorithm isolates distinctive pixel pairs that hit the same leaf during traversal through multiple learned tree structures. The split functions stored at the intermediate nodes of the trees are trained to ensure that only visually similar patches or their geometric or photometric transformed versions fall into the same leaf node. The matching process involves passing all pixel positions in the images under analysis through the tree structures. We then compute matches by isolating points that uniquely collide with each other ie. fell in the same empty leaf in multiple trees. Our algorithm is linear in the number of pixels but can be made constant time on a parallel computation architecture as the tree traversal for individual image points is decoupled. We demonstrate the efficacy of our method by using it to perform optical flow matching and stereo matching on some challenging benchmarks. Experimental results show that not only is our method extremely computationally efficient, but it is also able to match or outperform state of the art methods that are much more complex.

This paper presents an approach for reconstruction of 4D temporally coherent models of complex dynamic scenes. No prior knowledge is required of scene structure or camera calibration allowing reconstruction from multiple moving cameras. Sparse-to-dense temporal correspondence is integrated with joint multi-view segmentation and reconstruction to obtain a complete 4D representation of static and dynamic objects. Temporal coherence is exploited to overcome visual ambiguities resulting in improved reconstruction of complex scenes. Robust joint segmentation and reconstruction of dynamic objects is achieved by introducing a geodesic star convexity constraint. Comparative evaluation is performed on a variety of unstructured indoor and outdoor dynamic scenes with hand-held cameras and multiple people. This demonstrates reconstruction of complete temporally coherent 4D scene models with improved non-rigid object segmentation and shape reconstruction.

Given semantic descriptions of object classes, zero-shot learning aims to accurately recognize objects of the unseen classes, from which no examples are available at the training stage, by associating them to the seen classes, from which labeled examples are provided. We propose to tackle this problem from the perspective of manifold learning. Our main idea is to align the semantic space that is derived from external information to the model space that concerns itself with recognizing visual features. To this end, we introduce a set of "phantom" object classes whose coordinates live in both the semantic space and the model space. Serving as bases in a dictionary, they can be optimized from labeled data such that the synthesized real object classifiers achieve optimal discriminative performance. We demonstrate superior accuracy of our approach over the state of the art on four benchmark datasets for zero-shot learning, including the full ImageNet Fall 2011 dataset with more than 20,000 unseen classes.

We propose a novel spatially continuous framework for convex relaxations based on functional lifting. Our method can be interpreted as a sublabel-accurate solution to multilabel problems. We show that previously proposed functional lifting methods optimize an energy which is linear between two labels and hence require (often infinitely) many labels for a faithful approximation. In contrast, the proposed formulation is based on a piecewise convex approximation and therefore needs far fewer labels - see Fig. 1. In comparison to recent MRF-based approaches, our method is formulated in a spatially continuous setting and shows less grid bias. Moreover, in a local sense, our formulation is the tightest possible convex relaxation. It is easy to implement and allows an efficient primal-dual optimization on GPUs. We show the effectiveness of our approach on several computer vision problems.

Building a complete 3D model of a scene, given only a single depth image, is underconstrained. To gain a full volumetric model, one needs either multiple views, or a single view together with a library of unambiguous 3D models that will fit the shape of each individual object in the scene. We hypothesize that objects of dissimilar semantic classes often share similar 3D shape components, enabling a limited dataset to model the shape of a wide range of objects, and hence estimate their hidden geometry. Exploring this hypothesis, we propose an algorithm that can complete the unobserved geometry of tabletop-sized objects, based on a supervised model trained on already available volumetric elements. Our model maps from a local observation in a single depth image to an estimate of the surface shape in the surrounding neighborhood. We validate our approach both qualitatively and quantitatively on a range of indoor object collections and challenging real scenes.

Deep Recurrent Neural Network architectures, though remarkably capable at modeling sequences, lack an intuitive high-level spatio-temporal structure. That is while many problems in computer vision inherently have an underlying high-level structure and can benefit from it. Spatio-temporal graphs are a popular tool for imposing such high-level intuitions in the formulation of real world problems. In this paper, we propose an approach for combining the power of high-level spatio-temporal graphs and sequence learning success of Recurrent Neural Networks (RNNs). We develop a scalable method for casting an arbitrary spatio-temporal graph as a rich RNN mixture that is feedforward, fully differentiable, and jointly trainable. The proposed method is generic and principled as it can be used for transforming any spatio-temporal graph through employing a certain set of well defined steps. The evaluations of the proposed approach on a diverse set of problems, ranging from modeling human motion to object interactions, shows improvement over the state-of-the-art with a large margin. We expect this method to empower new approaches to problem formulation through high-level spatio-temporal graphs and Recurrent Neural Networks.

This paper presents stacked attention networks (SANs)that learn to answer natural language questions from images. SANs use semantic representation of a question as query to search for the regions in an image that are related to the answer. We argue that image question answering (QA) often requires multiple steps of reasoning. Thus, we develop a multiple-layer SAN in which we query an image multiple times to infer the answer progressively. Experiments conducted on four image QA data sets demonstrate that the proposed SANs significantly outperform previous state-of-the-art approaches. The visualization of the attention layers illustrates the progress that the SAN locates the relevant visual clues that lead to the answer of the question layer-by-layer.

Linear Dynamical Systems (LDSs) are the fundamental tools for encoding spatio-temporal data in various disciplines. To enhance the performance of LDSs, in this paper, we address the challenging issue of performing sparse coding on the space of LDSs, where both data and dictionary atoms are LDSs. Rather than approximate the extended observability with a finite-order matrix, we represent the space of LDSs by an infinite Grassmannian consisting of the orthonormalized extended observability subspaces. Via a homeomorphic mapping, such Grassmannian is embedded into the space of symmetric matrices, where a tractable objective function can be derived for sparse coding. Then, we propose an efficient method to learn the system parameters of the dictionary atoms explicitly, by imposing the symmetric constraint to the transition matrices of the data and dictionary systems. Moreover, we combine the state covariance into the algorithm formulation, thus further promoting the performance of the models with symmetric transition matrices. Comparative experimental evaluations reveal the superior performance of proposed methods on various tasks including video classification and tactile recognition.

We investigate the problem of fine-grained sketch-based image retrieval (SBIR), where free-hand human sketches are used as queries to perform instance-level retrieval of images. This is an extremely challenging task because (i) visual comparisons not only need to be fine-grained but also executed cross-domain, (ii) free-hand (finger) sketches are highly abstract, making fine-grained matching harder, and most importantly (iii) annotated cross-domain sketch-photo datasets required for training are scarce, challenging many state-of-the-art machine learning techniques. In this paper, for the first time, we address all these challenges, providing a step towards the capabilities that would underpin a commercial sketch-based image retrieval application. We introduce a new database of 1,432 sketch-photo pairs from two categories with 32,000 fine-grained triplet ranking annotations. We then develop a deep triplet-ranking model for instance-level SBIR with a novel data augmentation and staged pre-training strategy to alleviate the issue of insufficient fine-grained training data. Extensive experiments are carried out to contribute a variety of insights into the challenges of data sufficiency and over-fitting avoidance when training deep networks for fine-grained cross-domain ranking tasks.

This paper introduces situation recognition, the problem of producing a concise summary of the situation an image depicts including: (1) the main activity (e.g., clipping), (2) the participating actors, objects, substances, and locations (e.g., man, shears, sheep, wool, and field) and most importantly (3) the roles these participants play in the activity (e.g., the man is clipping, the shears are his tool, the wool is being clipped from the sheep, and the clipping is in a field). We use FrameNet, a verb and role lexicon developed by linguists, to define a large space of possible situations and collect a large-scale dataset containing over 500 activities, 1,700 roles, 11,000 objects, 125,000 images, and 200,000 unique situations. We also introduce structured prediction baselines and show that, in activity-centric images, situation-driven prediction of objects and activities outperforms independent object and activity recognition.

Event cameras are bio-inspired vision sensors which mimic retinas to measure per-pixel intensity change rather than outputting an actual intensity image. This proposed paradigm shift away from traditional frame cameras offers significant potential advantages: namely avoiding high data rates, dynamic range limitations and motion blur. Unfortunately, however, established computer vision algorithms may not at all be applied directly to event cameras. Methods proposed so far to reconstruct images, estimate optical flow, track a camera and reconstruct a scene come with severe restrictions on the environment or on the motion of the camera, e.g. allowing only rotation. Here, we propose, to the best of our knowledge, the first algorithm to simultaneously recover the motion field and brightness image, while the camera undergoes a generic motion through any scene. Our approach employs minimisation of a cost function that contains the asynchronous event data as well as spatial and temporal regularisation within a sliding window time interval. Our implementation relies on GPU-based optimisation and runs in near real-time. In a series of examples, we demonstrate the successful operation of our framework, including in situations where conventional cameras heavily suffer from dynamic range limitations or motion blur.