CVPR 2017 论文列表

Deep neural networks require a large amount of labeled training data during supervised learning. However, collecting and labeling so much data might be infeasible in many cases. In this paper, we introduce a deep transfer learning scheme, called selective joint fine-tuning, for improving the performance of deep learning tasks with insufficient training data. In this scheme, a target learning task with insufficient training data is carried out simultaneously with another source learning task with abundant training data. However, the source learning task does not use all existing training data. Our core idea is to identify and use a subset of training images from the original source learning task whose low-level characteristics are similar to those from the target learning task, and jointly fine-tune shared convolutional layers for both tasks. Specifically, we compute descriptors from linear or nonlinear filter bank responses on training images from both tasks, and use such descriptors to search for a desired subset of training samples for the source learning task. Experiments demonstrate that our deep transfer learning scheme achieves state-of-the-art performance on multiple visual classification tasks with insufficient training data for deep learning. Such tasks include Caltech 256, MIT Indoor 67, and fine-grained classification problems (Oxford Flowers 102 and Stanford Dogs 120). In comparison to fine-tuning without a source domain, the proposed method can improve the classification accuracy by 2% - 10% using a single model. Codes and models are available at https://github.com/ZYYSzj/Selective-Joint-Fine-tuning.

In recent years, there has been a renewed interest in jointly modeling perception and action. At the core of this investigation is the idea of modeling affordances. However, when it comes to predicting affordances, even the state of the art approaches still do not use any ConvNets. Why is that? Unlike semantic or 3D tasks, there still does not exist any large-scale dataset for affordances. In this paper, we tackle the challenge of creating one of the biggest dataset for learning affordances. We use seven sitcoms to extract a diverse set of scenes and how actors interact with different objects in the scenes. Our dataset consists of more than 10K scenes and 28K ways humans can interact with these 10K images. We also propose a two-step approach to predict affordances in a new scene. In the first step, given a location in the scene we classify which of the 30 pose classes is the likely affordance pose. Given the pose class and the scene, we then use a Variational Autoencoder (VAE) to extract the scale and deformation of the pose. The VAE allows us to sample the distribution of possible poses at test time. Finally, we show the importance of large-scale data in learning a generalizable and robust model of affordances.

Graph matching is a fundamental problem in computer vision and pattern recognition area. In general, it can be formulated as an Integer Quadratic Programming (IQP) problem. Since it is NP-hard, approximate relaxations are required. In this paper, a new graph matching method has been proposed. There are three main contributions of the proposed method: (1) we propose a new graph matching relaxation model, called Binary Constraint Preserving Graph Matching (BPGM), which aims to incorporate the discrete binary mapping constraints more in graph matching relaxation. Our BPGM is motivated by a new observation that the discrete binary constraints in IQP matching problem can be represented (or encoded) exactly by a l2-norm constraint. (2) An effective projection algorithm has been derived to solve BPGM model. (3) Using BPGM, we propose a path-following strategy to optimize IQP matching problem and thus obtain a desired discrete solution at convergence. Promising experimental results show the effectiveness of the proposed method.

Traditional action recognition methods aim to recognize actions with complete observations/executions. However, it is often difficult to capture fully executed actions due to occlusions, interruptions, etc. Meanwhile, action prediction/recognition in advance based on partial observations is essential for preventing the situation from deteriorating. Besides, fast spotting human activities using partially observed data is a critical ingredient for retrieval systems. Inspired by the recent success of data binarization in efficient retrieval/recognition, we propose a novel approach, named Partial Reconstructive Binary Coding (PRBC), for action analysis based on limited frame glimpses during any period of the complete execution. Specifically, we learn discriminative compact binary codes for partial actions via a joint learning framework, which collaboratively tackles feature reconstruction as well as binary coding. We obtain the solution to PRBC based on a discrete alternating iteration algorithm. Extensive experiments on four realistic action datasets in terms of three tasks (i.e., partial action retrieval, recognition and prediction) clearly show the superiority of PRBC over the state-of-the-art methods, along with significantly reduced memory load and computational costs during the online test.

This paper studies visual pattern discovery in large-scale image collections via binarized mode seeking, where images can only be represented as binary codes for efficient storage and computation. We address this problem from the perspective of binary space mode seeking. First, a binary mean shift (bMS) is proposed to discover frequent patterns via mode seeking directly in binary space. The binomial-based kernel and binary constraint are introduced for binarized analysis. Second, we further extend bMS to a more general form, namely contrastive binary mean shift (cbMS), which maximizes the contrastive density in binary space, for finding informative patterns that are both frequent and discriminative for the dataset. With the binarized algorithm and optimization, our methods demonstrate significant computation (50X) and storage (32X) improvement compared to standard techniques operating in Euclidean space, while the performance does not largely degenerate. Furthermore, cbMS discovers more informative patterns by suppressing low discriminative modes. We evaluate our methods on both annotated ILSVRC (1M images) and un-annotated blind Flickr (10M images) datasets with million scale images, which demonstrates both the scalability and effectiveness of our algorithms for discovering frequent and informative patterns in large scale collection.

In this paper we introduce a large-scale hand pose dataset, collected using a novel capture method. Existing datasets are either generated synthetically or captured using depth sensors: synthetic datasets exhibit a certain level of appearance difference from real depth images, and real datasets are limited in quantity and coverage, mainly due to the difficulty to annotate them. We propose a tracking system with six 6D magnetic sensors and inverse kinematics to automatically obtain 21-joints hand pose annotations of depth maps captured with minimal restriction on the range of motion. The capture protocol aims to fully cover the natural hand pose space. As shown in embedding plots, the new dataset exhibits a significantly wider and denser range of hand poses compared to existing benchmarks. Current state-of-the-art methods are evaluated on the dataset, and we demonstrate significant improvements in cross-benchmark performance. We also show significant improvements in egocentric hand pose estimation with a CNN trained on the new dataset.

Despite the recent success of neural networks in image feature learning, a major problem in the video domain is the lack of sufficient labeled data for learning to model temporal information. In this paper, we propose an unsupervised temporal modeling method that learns from untrimmed videos. The speed of motion varies constantly, e.g., a man may run quickly or slowly. We therefore train a Multirate Visual Recurrent Model (MVRM) by encoding frames of a clip with different intervals. This learning process makes the learned model more capable of dealing with motion speed variance. Given a clip sampled from a video, we use its past and future neighboring clips as the temporal context, and reconstruct the two temporal transitions, i.e., present->past transition and present->future transition, reflecting the temporal information in different views. The proposed method exploits the two transitions simultaneously by incorporating a bidirectional reconstruction which consists of a backward reconstruction and a forward reconstruction. We apply the proposed method to two challenging video tasks, i.e., complex event detection and video captioning, in which it achieves state-of-the-art performance. Notably, our method generates the best single feature for event detection with a relative improvement of 10.4% on the MEDTest-13 dataset and achieves the best performance in video captioning across all evaluation metrics on the YouTube2Text dataset.

We develop the first approximate inference algorithm for 1-Best (and M-Best) decoding in bidirectional neural sequence models by extending Beam Search (BS) to reason about both forward and backward time dependencies. Beam Search (BS) is a widely used approximate inference algorithm for decoding sequences from unidirectional neural sequence models. Interestingly, approximate inference in bidirectional models remains an open problem, despite their significant advantage in modeling information from both the past and future. To enable the use of bidirectional models, we present Bidirectional Beam Search (BiBS), an efficient algorithm for approximate bidirectional inference. To evaluate our method and as an interesting problem in its own right, we introduce a novel Fill-in-the-Blank Image Captioning task which requires reasoning about both past and future sentence structure to reconstruct sensible image descriptions. We use this task as well as the Visual Madlibs dataset to demonstrate the effectiveness of our approach, consistently outperforming all baseline methods.

Person re-identification (ReID) is an important task in wide area video surveillance which focuses on identifying people across different cameras. Recently, deep learning networks with a triplet loss become a common framework for person ReID. However, the triplet loss pays main attentions on obtaining correct orders on the training set. It still suffers from a weaker generalization capability from the training set to the testing set, thus resulting in inferior performance. In this paper, we design a quadruplet loss, which can lead to the model output with a larger inter-class variation and a smaller intra-class variation compared to the triplet loss. As a result, our model has a better generalization ability and can achieve a higher performance on the testing set. In particular, a quadruplet deep network using a margin-based online hard negative mining is proposed based on the quadruplet loss for the person ReID. In extensive experiments, the proposed network outperforms most of the state-of-the-art algorithms on representative datasets which clearly demonstrates the effectiveness of our proposed method.

Querying with an example image is a simple and intuitive interface to retrieve information from a visual database. Most of the research in image retrieval has focused on the task of instance-level image retrieval, where the goal is to retrieve images that contain the same object instance as the query image. In this work we move beyond instance-level retrieval and consider the task of semantic image retrieval in complex scenes, where the goal is to retrieve images that share the same semantics as the query image. We show that, despite its subjective nature, the task of semantically ranking visual scenes is consistently implemented across a pool of human annotators. We also show that a similarity based on human-annotated region-level captions is highly correlated with the human ranking and constitutes a good computable surrogate. Following this observation, we learn a visual embedding of the images where the similarity in the visual space is correlated with their semantic similarity surrogate. We further extend our model to learn a joint embedding of visual and textual cues that allows one to query the database using a text modifier in addition to the query image, adapting the results to the modifier. Finally, our model can ground the ranking decisions by showing regions that contributed the most to the similarity between pairs of images, providing a visual explanation of the similarity.

Lacking realistic ground truth data, image denoising techniques are traditionally evaluated on images corrupted by synthesized i.i.d. Gaussian noise. We aim to obviate this unrealistic setting by developing a methodology for benchmarking denoising techniques on real photographs. We capture pairs of images with different ISO values and appropriately adjusted exposure times, where the nearly noise-free low-ISO image serves as reference. To derive the ground truth, careful post-processing is needed. We correct spatial misalignment, cope with inaccuracies in the exposure parameters through a linear intensity transform based on a novel heteroscedastic Tobit regression model, and remove residual low-frequency bias that stems, e.g., from minor illumination changes. We then capture a novel benchmark dataset, the Darmstadt Noise Dataset (DND), with consumer cameras of differing sensor sizes. One interesting finding is that various recent techniques that perform well on synthetic noise are clearly outperformed by BM3D on photographs with real noise. Our benchmark delineates realistic evaluation scenarios that deviate strongly from those commonly used in the scientific literature.

Among learning based hashing methods, supervised hashing seeks compact binary representation of the training data to preserve semantic similarities. Recent years have witnessed various problem formulations and optimization methods for supervised hashing. Most of them optimize a form of loss function with a regulization term, which can be viewed as a maximum a posterior (MAP) estimation of the hashing codes. However, these approaches are prone to overfitting unless hyperparameters are tuned carefully. To address this problem, we present a novel fully Bayesian treatment for supervised hashing problem, named Bayesian Supervised Hashing (BSH), in which hyperparameters are automatically tuned during optimization. Additionally, by utilizing automatic relevance determination (ARD), we can figure out relative discriminating ability of different hashing bits and select most informative bits among them. Experimental results on three real-world image datasets with semantic information show that BSH can achieve superior performance over state-of-the-art methods with comparable training time.

In this paper, we develop an interest point detector and binary feature descriptor for spherical images. We take as inspiration a recent framework developed for planar images, BRISK (Binary Robust Invariant Scalable Keypoints), and adapt the method to operate on spherical images. All of our processing is intrinsic to the sphere and avoids the distortion inherent in storing and indexing spherical images in a 2D representation. We discretise images on a spherical geodesic grid formed by recursive subdivision of a triangular mesh. This leads to a multiscale pixel grid comprising mainly hexagonal pixels that lends itself naturally to a spherical image pyramid representation. For interest point detection, we use a variant of the Accelerated Segment Test (AST) corner detector which operates on our geodesic grid. We estimate a continuous scale and location for features and descriptors are built by sampling onto a regular pattern in the tangent space. We evaluate repeatability, precision and recall on both synthetic spherical images with known ground truth correspondences and real images.

This paper presents BIND (Binary Integrated Net Descriptor), a texture-less object detector that encodes multi-layered binary-represented nets for high precision edge-based description. Our proposed concept aligns layers of object-sized patches (nets) onto highly fragmented occlusion resistant line-segment midpoints (linelets) to encode regional information into efficient binary strings. These lightweight nets encourage discriminative object description through their high-spatial resolution, enabling highly precise encoding of the object's edges and internal texture-less information. BIND achieved various invariant properties such as rotation, scale and edge-polarity through its unique binary logical-operated encoding and matching techniques, while performing remarkably well in occlusion and clutter. Apart from yielding efficient computational performance, BIND also attained remarkable recognition rates surpassing recent state-of-the-art texture-less object detectors such as BORDER, BOLD and LINE2D.

In this work, we explore the problem of generating fantastic special-effects for the typography. It is quite challenging due to the model diversities to illustrate varied text effects for different characters. To address this issue, our key idea is to exploit the analytics on the high regularity of the spatial distribution for text effects to guide the synthesis process. Specifically, we characterize the stylized patches by their normalized positions and the optimal scales to depict their style elements. Our method first estimates these two features and derives their correlation statistically. They are then converted into soft constraints for texture transfer to accomplish adaptive multi-scale texture synthesis and to make style element distribution uniform. It allows our algorithm to produce artistic typography that fits for both local texture patterns and the global spatial distribution in the example. Experimental results demonstrate the superiority of our method for various text effects over conventional style transfer methods. In addition, we validate the effectiveness of our algorithm with extensive artistic typography library generation.

There is more to images than their objective physical content: for example, advertisements are created to persuade a viewer to take a certain action. We propose the novel problem of automatic advertisement understanding. To enable research on this problem, we create two datasets: an image dataset of 64,832 image ads, and a video dataset of 3,477 ads. Our data contains rich annotations encompassing the topic and sentiment of the ads, questions and answers describing what actions the viewer is prompted to take and the reasoning that the ad presents to persuade the viewer ("What should I do according to this ad, and why should I do it?"), and symbolic references ads make (e.g. a dove symbolizes peace). We also analyze the most common persuasive strategies ads use, and the capabilities that computer vision systems should have to understand these strategies. We present baseline classification results for several prediction tasks, including automatically answering questions about the messages of the ads.

Attribute-based recognition models, due to their impressive performance and their ability to generalize well on novel categories, have been widely adopted for many computer vision applications. However, usually both the attribute vocabulary and the class-attribute associations have to be provided manually by domain experts or large number of annotators. This is very costly and not necessarily optimal regarding recognition performance, and most importantly, it limits the applicability of attribute-based models to large scale data sets. To tackle this problem, we propose an end-to-end unsupervised attribute learning approach. We utilize online text corpora to automatically discover a salient and discriminative vocabulary that correlates well with the human concept of semantic attributes. Moreover, we propose a deep convolutional model to optimize class-attribute associations with a linguistic prior that accounts for noise and missing data in text. In a thorough evaluation on ImageNet, we demonstrate that our model is able to efficiently discover and learn semantic attributes at a large scale. Furthermore, we demonstrate that our model outperforms the state-of-the-art in zero-shot learning on three data sets: ImageNet, Animals with Attributes and aPascal/aYahoo. Finally, we enable attribute-based learning on ImageNet and will share the attributes and associations for future research.

We present a novel visual attention tracking technique based on Shared Attention modeling. By considering the viewer as a participant in the activity occurring in the scene, our model learns the loci of attention of the scene actors and use it to augment image salience. We go beyond image salience and instead of only computing the power of image regions to pull attention, we also consider the strength with which the scene actors push attention to the region in question, thus the term Attentional Push. We present a convolutional neural network (CNN) which augments standard saliency models with Attentional Push. Our model contains two pathways: an Attentional Push pathway which learns the gaze location of the scene actors and a saliency pathway. These are followed by a shallow augmented saliency CNN which combines them and generates the augmented saliency. For training, we use transfer learning to initialize and train the Attentional Push CNN by minimizing the classification error of following the actors' gaze location on a 2-D grid using a large-scale gaze-following dataset. The Attentional Push CNN is then fine-tuned along with the augmented saliency CNN to minimize the Euclidean distance between the augmented saliency and ground truth fixations using an eye-tracking dataset, annotated with the head and the gaze location of the scene actors. We evaluate our model on three challenging eye fixation datasets, SALICON, iSUN and CAT2000, and illustrate significant improvements in predicting viewers' fixations in social scenes.

We propose a new tracking framework with an attentional mechanism that chooses a subset of the associated correlation filters for increased robustness and computational efficiency. The subset of filters is adaptively selected by a deep attentional network according to the dynamic properties of the tracking target. Our contributions are manifold, and are summarised as follows: (i) Introducing the Attentional Correlation Filter Network which allows adaptive tracking of dynamic targets. (ii) Utilising an attentional network which shifts the attention to the best candidate modules, as well as predicting the estimated accuracy of currently inactive modules. (iii) Enlarging the variety of correlation filters which cover target drift, blurriness, occlusion, scale changes, and flexible aspect ratio. (iv) Validating the robustness and efficiency of the attentional mechanism for visual tracking through a number of experiments. Our method achieves similar performance to non real-time trackers, and state-of-the-art performance amongst real-time trackers.