CVPR 2016 Papers

Zero-shot recognition (ZSR) deals with the problem of predicting class labels for target domain instances based on source domain side information (e.g. attributes) of unseen classes. We formulate ZSR as a binary prediction problem. Our resulting classifier is class-independent. It takes an arbitrary pair of source and target domain instances as input and predicts whether or not they come from the same class, i.e. whether there is a match. We model the posterior probability of a match since it is a sufficient statistic and propose a latent probabilistic model in this context. We develop a joint discriminative learning framework based on dictionary learning to jointly learn the parameters of our model for both domains, which ultimately leads to our class-independent classifier. Many of the existing embedding methods can be viewed as special cases of our probabilistic model. On ZSR our method shows 4.90% improvement over the state-of-the-art in accuracy averaged across four benchmark datasets. We also adapt ZSR method for zero-shot retrieval and show 22.45% improvement accordingly in mean average precision (mAP).

We present YOLO, a new approach to object detection. Prior work on object detection repurposes classifiers to perform detection. Instead, we frame object detection as a regression problem to spatially separated bounding boxes and associated class probabilities. A single neural network predicts bounding boxes and class probabilities directly from full images in one evaluation. Since the whole detection pipeline is a single network, it can be optimized end-to-end directly on detection performance. Our unified architecture is extremely fast. Our base YOLO model processes images in real-time at 45 frames per second. A smaller version of the network, Fast YOLO, processes an astounding 155 frames per second while still achieving double the mAP of other real-time detectors. Compared to state-of-the-art detection systems, YOLO makes more localization errors but is less likely to predict false positives on background. Finally, YOLO learns very general representations of objects. It outperforms other detection methods, including DPM and R-CNN, when generalizing from natural images to other domains like artwork.

Video concept learning often requires a large set of training samples. In practice, however, acquiring noise-free training labels with sufficient positive examples is very expensive. A plausible solution for training data collection is by sampling from the vast quantities of images and videos on the Web. Such a solution is motivated by the assumption that the retrieved images or videos are highly correlated with the query. Still, a number of challenges remain. First, Web videos are often untrimmed. Thus, only parts of the videos are relevant to the query. Second, the retrieved Web images are always highly relevant to the issued query. However, thoughtlessly utilizing the images in the video domain may even hurt the performance due to the well known semantic drift and domain gap problems. As a result, a valid question is how Web images and videos interact for video concept learning. In this paper, we propose a Lead--Exceed Neural Network (LENN), which reinforces the training on Web images and videos in a curriculum manner. Specifically, the training proceeds by inputting frames of Web videos to obtain a network. The Web images are then filtered by the learnt network and the selected images are additionally fed into the network to enhance the architecture and further trim the videos. In addition, Long Short-Term Memory (LSTM) can be applied on the trimmed videos to explore temporal information. Encouraging results are reported on UCF101, TRECVID 2013 and 2014 MEDTest in the context of both action recognition and event detection. Without using human annotated exemplars, our proposed LENN can achieve 74.4% accuracy on UCF101 dataset.

The complex compositional structure of language makes problems at the intersection of vision and language challenging. But language also provides a strong prior that can result in good superficial performance, without the underlying models truly understanding the visual content. This can hinder progress in pushing state of art in the computer vision aspects of multi-modal AI. In this paper, we address binary Visual Question Answering (VQA) on abstract scenes. We formulate this problem as visual verification of concepts inquired in the questions. Specifically, we convert the question to a tuple that concisely summarizes the visual concept to be detected in the image. If the concept can be found in the image, the answer to the question is "yes", and otherwise "no". Abstract scenes play two roles (1) They allow us to focus on the high-level semantics of the VQA task as opposed to the low-level recognition problems, and perhaps more importantly, (2) They provide us the modality to balance the dataset such that language priors are controlled, and the role of vision is essential. In particular, we collect fine-grained pairs of scenes for every question, such that the answer to the question is "yes" for one scene, and "no" for the other for the exact same question. Indeed, language priors alone do not perform better than chance on our balanced dataset. Moreover, our proposed approach matches the performance of a state-of-the-art VQA approach on the unbalanced dataset, and outperforms it on the balanced dataset.

We present a method that learns to answer visual questions by selecting image regions relevant to the text-based query. Our method maps textual queries and visual features from various regions into a shared space where they are compared for relevance with an inner product. Our method exhibits significant improvements in answering questions such as "what color," where it is necessary to evaluate a specific location, and "what room," where it selectively identifies informative image regions. Our model is tested on the recently released VQA dataset, which features free-form human-annotated questions and answers.

In many challenging visual recognition tasks where training data is limited, Vectors of Locally Aggregated Descriptors (VLAD) have emerged as powerful image/video representations that compete with or outperform state-of-the-art approaches. In this paper, we address two fundamental limitations of VLAD: its requirement for the local descriptors to have vector form and its restriction to linear classifiers due to its high-dimensionality. To this end, we introduce a kernelized version of VLAD. This not only lets us inherently exploit more sophisticated classification schemes, but also enables us to efficiently aggregate non-vector descriptors (e.g., manifold-valued data) in the VLAD framework. Furthermore, we propose an approximate formulation that allows us to accelerate the coding process while still benefiting from the properties of kernel VLAD. Our experiments demonstrate the effectiveness of our approach at handling manifold-valued data, such as covariance descriptors, on several classification tasks. Our results also evidence the benefits of our nonlinear VLAD descriptors against the linear ones in Euclidean space using several standard benchmark datasets.

Since Convolutional Neural Networks (CNNs) have become the leading learning paradigm in visual recognition, Naive Bayes Nearest Neighbor (NBNN)-based classifiers have lost momentum in the community. This is because (1) such algorithms cannot use CNN activations as input features; (2) they cannot be used as final layer of CNN architectures for end-to-end training , and (3) they are generally not scalable and hence cannot handle big data. This paper proposes a framework that addresses all these issues, thus bringing back NBNNs on the map. We solve the first by extracting CNN activations from local patches at multiple scale levels, similarly to [13]. We address simultaneously the second and third by proposing a scalable version of Naive Bayes Non-linear Learning (NBNL, [7]). Results obtained using pre-trained CNNs on standard scene and domain adaptation databases show the strength of our approach, opening a new season for NBNNs.

This paper studies the challenging problem of identifying unusual instances of known objects in images within an "open world" setting. That is, we aim to find objects that are members of a known class, but which are not typical of that class. Thus the "unusual object" should be distinguished from both the "regular object" and the "other objects". Such unusual objects may be of interest in many applications such as surveillance or quality control. We propose to identify unusual objects by inspecting the distribution of object detection scores at multiple image regions. The key observation motivating our approach is that "regular object" images, "unusual object" images and "other objects" images exhibit different region-level scores in terms of both the score values and the spatial distributions. To model these distributions we propose to use Gaussian Processes (GP) to construct two separate generative models, one for the "regular object" and the other for the "other objects". More specifically, we design a new covariance function to simultaneously model the detection score at a single location and the score dependencies between multiple regions. We demonstrate that the proposed approach outperforms comparable methods on a new large dataset constructed for the purpose.

Much recent progress in Vision-to-Language (V2L) problems has been achieved through a combination of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). This approach does not explicitly represent high-level semantic concepts, but rather seeks to progress directly from image features to text. In this paper we investigate whether this direct approach succeeds due to, or despite, the fact that it avoids the explicit representation of high-level information. We propose a method of incorporating high-level concepts into the successful CNN-RNN approach, and show that it achieves a significant improvement on the state-of-the-art in both image captioning and visual question answering. We also show that the same mechanism can be used to introduce external semantic information and that doing so further improves performance. We achieve the best reported results on both image captioning and VQA on several benchmark datasets, and provide an analysis of the value of explicit high-level concepts in V2L problems.

In this paper, we propose a novel method for depth estimation in light fields which employs a specifically designed sparse decomposition to leverage the depth-orientation relationship on its epipolar plane images. The proposed method learns the structure of the central view and uses this information to construct a light field dictionary for which groups of atoms correspond to unique disparities. This dictionary is then used to code a sparse representation of the light field. Analysing the coefficients of this representation with respect to the disparities of their corresponding atoms yields an accurate and robust estimate of depth. In addition, if the light field has multiple depth layers, such as for reflective or transparent surfaces, statistical analysis of the coefficients can be employed to infer the respective depth of the superimposed layers.

Tracking the ball is critical for video-based analysis of team sports. However, it is difficult, especially in low-resolution images, due to the small size of the ball, its speed that creates motion blur, and its often being occluded by players. In this paper, we propose a generic and principled approach to modeling the interaction between the ball and the players while also imposing appropriate physical constraints on the ball's trajectory. We show that our approach, formulated in terms of a Mixed Integer Program, is more robust and more accurate than several state-of-the-art approaches on real-life volleyball, basketball, and soccer sequences.

This paper tackles the problem of spatio-temporal action localization in a video without assuming the availability of multiple videos or any prior annotations. Action is localized by employing images downloaded from internet using action name. Given web images, we first mitigate image noise using random walk framework and evade distracting backgrounds within images using image action proposals. Then, given a video, we generate multiple spatio-temporal action proposals. We suppress camera and background generated proposals by exploiting optical flow gradients within proposal. To obtain the most action representative proposal, we propose to reconstruct action proposals in the video by leveraging the action proposal in images. Moreover, we preserve the temporal smoothness of the video by introducing consensus regularization. Consensus regularization enforces consistency among coefficients vectors of multiple frames within proposal. %We reconstruct video action proposals from image action proposals while enforcing consistency across coefficient vectors of multiple frames by consensus regularization. Finally, the video proposal that have the lowest reconstruction cost and is motion salient is considered as final action localization. Our extensive experiments on trimmed as well as untrimmed datasets validate the effectiveness of proposed approach.

We address the problem of weakly supervised object localization where only image-level annotations are available for training. Many existing approaches tackle this problem through object proposal mining. However, a substantial amount of noise in object proposals causes ambiguities for learning discriminative object models. Such approaches are sensitive to model initialization and often converge to an undesirable local minimum. In this paper, we address this problem by progressive domain adaptation with two main steps: classification adaptation and detection adaptation. In classification adaptation, we transfer a pre-trained network to our multi-label classification task for recognizing the presence of a certain object in an image. In detection adaptation, we first use a mask-out strategy to collect class-specific object proposals and apply multiple instance learning to mine confident candidates. We then use these selected object proposals to fine-tune all the layers, resulting in a fully adapted detection network. We extensively evaluate the localization performance on the PASCAL VOC and ILSVRC datasets and demonstrate significant performance improvement over the state-of-the-art methods.

State-of-the-art learning based boundary detection methods require extensive training data. Since labelling object boundaries is one of the most expensive types of annotations, there is a need to relax the requirement to carefully annotate images to make both the training more affordable and to extend the amount of training data. In this paper we propose a technique to generate weakly supervised annotations and show that bounding box annotations alone suffice to reach high-quality object boundaries without using any object-specific boundary annotations. With the proposed weak supervision techniques we achieve the top performance on the object boundary detection task, outperforming by a large margin the current fully supervised state-of-the-art methods.

Weakly supervised learning of object detection is an important problem in image understanding that still does not have a satisfactory solution. In this paper, we address this problem by exploiting the power of deep convolutional neural networks pre-trained on large-scale image-level classification tasks. We propose a weakly supervised deep detection architecture that modifies one such network to operate at the level of image regions, performing simultaneously region selection and classification. Trained as an image classifier, the architecture implicitly learns object detectors that are better than alternative weakly supervised detection systems on the PASCAL VOC data. The model, which is a simple and elegant end-to-end architecture, outperforms standard data augmentation and fine-tuning techniques for the task of image-level classification as well.

Training object class detectors typically requires a large set of images in which objects are annotated by bounding-boxes. However, manually drawing bounding-boxes is very time consuming. We propose a new scheme for training object detectors which only requires annotators to verify bounding-boxes produced automatically by the learning algorithm. Our scheme iterates between re-training the detector, re-localizing objects in the training images, and human verification. We use the verification signal both to improve re-training and to reduce the search space for re-localisation, which makes these steps different to what is normally done in a weakly supervised setting. Extensive experiments on PASCAL VOC 2007 show that (1) using human verification to update detectors and reduce the search space leads to the rapid production of high-quality bounding-box annotations; (2) our scheme delivers detectors performing almost as good as those trained in a fully supervised setting, without ever drawing any bounding-box; (3) as the verification task is very quick, our scheme substantially reduces total annotation time by a factor 6x-9x.

Humor is an integral part of human lives. Despite being tremendously impactful, it is perhaps surprising that we do not have a detailed understanding of humor yet. As interactions between humans and AI systems increase, it is imperative that these systems are taught to understand subtleties of human expressions such as humor. In this work, we are interested in the question - what content in a scene causes it to be funny? As a first step towards understanding visual humor, we analyze the humor manifested in abstract scenes and design computational models for them. We collect two datasets of abstract scenes that facilitate the study of humor at both the scene-level and the object-level. We analyze the funny scenes and explore the different types of humor depicted in them via human studies. We model two tasks that we believe demonstrate an understanding of some aspects of visual humor. The tasks involve predicting the funniness of a scene and altering the funniness of a scene. We show that our models perform well quantitatively, and qualitatively through human studies. Our datasets are publicly available.

We present an approach to matching images of objects in fine-grained datasets without using part annotations, with an application to the challenging problem of weakly supervised single-view reconstruction. This is in contrast to prior works that require part annotations, since matching objects across class and pose variations is challenging with appearance features alone. We overcome this challenge through a novel deep learning architecture, WarpNet, that aligns an object in one image with a different object in another. We exploit the structure of the fine-grained dataset to create artificial data for training this network in an unsupervised-discriminative learning approach. The output of the network acts as a spatial prior that allows generalization at test time to match real images across variations in appearance, viewpoint and articulation. On the CUB-200-2011 dataset of bird categories, we improve the AP over an appearance-only network by 13.6%. We further demonstrate that our WarpNet matches, together with the structure of fine-grained datasets, allow single-view reconstructions with quality comparable to using annotated point correspondences.

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.