CVPR 2015 论文列表

Spatial verification is a key step in boosting the performance of object-based image retrieval. It serves to eliminate unreliable correspondences between salient points in a given pair of images and is typically performed by analyzing the consistency of spatial transformations between the image regions involved in individual correspondences. In this paper, we consider the pairwise geometric relations between correspondences and propose a strategy to incorporate these relations at significantly reduced computational cost, which makes it suitable for large-scale object retrieval. In addition, we combine the information on geometric relations from both the individual correspondences and pairs of correspondences to further improve the verification accuracy. Experimental results on three reference datasets show that the proposed approach results in a substantial performance improvement compared to the existing methods, without making concessions regarding computational efficiency.

Large-scale Structure-from-Motion systems typically spend major computational effort on pairwise image matching and geometric verification in order to discover connected components in large-scale, unordered image collections. In recent years, the research community has spent significant effort on improving the efficiency of this stage. In this paper, we present a comprehensive overview of various state-of-the-art methods, evaluating and analyzing their performance. Based on the insights of this evaluation, we propose a learning-based approach, the PAirwise Image Geometry Encoding (PAIGE), to efficiently identify image pairs with scene overlap without the need to perform exhaustive putative matching and geometric verification. PAIGE achieves state-of-the-art performance and integrates well into existing Structure-from-Motion pipelines.

It is well known that the problem of camera pose estimation with unknown focal length has 7 degrees of freedom. Since each image point gives 2 constraints, solving this problem requires a minimum of 3.5 image points of 4 known 3D points, where 0.5 means either x or y coordinate of an image point. We refer to this minimal problem as P3.5P. However, the existing methods require 4 full image points to solve the camera pose and focal length. In this paper, we present a general solution to the true minimal P3.5P problem with up to 10 solutions. The remaining image coordinate is then used to filter the candidate solutions, which typically results in a single solution for good data or no solution for outliers. Experiments show the proposed method significantly improves the efficiency over the state of the art methods while maintaining a high accuracy.

We present a simple, efficient model for learning boundary detection based on a random forest classifier. Our approach combines (1) efficient clustering of training examples based on a simple partitioning of the space of local edge orientations and (2) scale-dependent calibration of individual tree output probabilities prior to multiscale combination. The resulting model outperforms published results on the challenging BSDS500 boundary detection benchmark. Further, on large datasets our model requires substantially less memory for training and speeds up training time by a factor of 10 over the structured forest model.

In multimedia annotation, due to the time constraints and the tediousness of manual tagging, it is quite common to utilize both tagged and untagged data to improve the performance of supervised learning when only limited tagged training data are available. This is often done by adding a geometrically based regularization term in the objective function of a supervised learning model. In this case, a similarity graph is indispensable to exploit the geometrical relationships among the training data points, and the graph construction scheme essentially determines the performance of these graph-based learning algorithms. However, most of the existing works construct the graph empirically and are usually based on a single feature without using the label information. In this paper, we propose a semi-supervised annotation approach by learning an optimal graph (OGL) from multi-cues (i.e., partial tags and multiple features) which can more accurately embed the relationships among the data points. We further extend our model to address out-of-sample and noisy label issues. Extensive experiments on four public datasets show the consistent superiority of OGL over state-of-the-art methods by up to 12% in terms of mean average precision.

Modern search engines receive large numbers of business related, local aware queries. Such queries are best answered using accurate, up-to-date, business listings, that contain representations of business categories. Creating such listings is a challenging task as businesses often change hands or close down. For businesses with street side locations one can leverage the abundance of street level imagery, such as Google Street View, to automate the process. However, while data is abundant, labeled data is not; the limiting factor is creation of large scale labeled training data. In this work, we utilize an ontology of geographical concepts to automatically propagate business category information and create a large, multi label, training data for fine grained storefront classification. Our learner, which is based on the GoogLeNet/inception Deep Convolutional Network architecture and classifies 208 categories, achieves human level accuracy.

Recently, hashing based approximate nearest neighbor (ANN) search has attracted much attention. Extensive new algorithms have been developed and successfully applied to different applications. However, two critical problems are rarely mentioned. First, in real-world applications, the data often comes in a streaming fashion but most of existing hashing methods are batch based models. Second, when the dataset becomes huge, it is almost impossible to load all the data into memory to train hashing models. In this paper, we propose a novel approach to handle these two problems simultaneously based on the idea of data sketching. A sketch of one dataset preserves its major characters but with significantly smaller size. With a small size sketch, our method can learn hash functions in an online fashion, while needs rather low computational complexity and storage space. Extensive experiments on two large scale benchmarks and one synthetic dataset demonstrate the efficacy of the proposed method.

We present an outdoor photometric stereo method using images captured in a single day. We simulate a sky hemisphere for each image according to its GPS and timestamp, and parameterize the obtained sky hemisphere into a quadratic skylight and a Gaussian sunlight distribution. Unlike previous works which usually model outdoor illumination as a sum of constant ambient light and a distant point light, our method models natural illumination according to a popular sky model and thus provides sufficient constraints for shape reconstruction from one day images. We generate pixel profiles of uniformly sampled unit vectors for the corresponding time of captures and evaluate them using correlation with the actual pixel profiles. The estimated surface normal is refined by MRF optimization. We have tested our method to recover objects and scenes of various sizes in real-world outdoor daylight.

One of the cornerstone principles of deep models is their abstraction capacity, i.e. their ability to learn abstract concepts from `simpler' ones. Through extensive experiments, we characterize the nature of the relationship between abstract concepts (specifically objects in images) learned by popular and high performing convolutional networks (conv-nets) and established mid-level representations used in computer vision (specifically semantic visual attributes). We focus on attributes due to their impact on several applications, such as object description, retrieval and mining, and active (and zero-shot) learning. Among the findings we uncover, we show empirical evidence of the existence of Attribute Centric Nodes (ACNs) within a conv-net, which is trained to recognize objects (not attributes) in images. These special conv-net nodes (1) collectively encode information pertinent to visual attribute representation and discrimination, (2) are unevenly and sparsely distribution across all layers of the conv-net, and (3) play an important role in conv-net based object recognition.

Low-rank matrix factorization is an essential problem in many areas including computer vision, with applications in e.g. affine structure-from-motion, photometric stereo, and non-rigid structure from motion. However, very little attention has been drawn to minimal cases for this problem or to using the minimal configuration of observations to find the solution. Minimal problems are useful when either outliers are present or the observation matrix is sparse. In this paper, we first give some theoretical insights on how to generate all the minimal problems of a given size using Laman graph theory. We then propose a new parametrization and a building-block scheme to solve these minimal problems by extending the solution from a small sized minimal problem. We test our solvers on synthetic data as well as real data with outliers or a large portion of missing data and show that our method can handle the cases when other iterative methods, based on convex relaxation, fail.

Edges in images of translucent objects are very different from edges in images of opaque objects. The physical causes for these differences are hard to characterize analytically and are not well understood. This paper considers one class of translucency edges---those caused by a discontinuity in surface orientation---and describes the physical causes of their appearance. We simulate thousands of translucency edge profiles using many different scattering material parameters, and we explain the resulting variety of edge patterns by qualitatively analyzing light transport. We also discuss the existence of shape and material metamers, or combinations of distinct shape or material parameters that generate the same edge profile. This knowledge is relevant to visual inference tasks that involve translucent objects, such as shape or material estimation.

Multi-object tracking has been recently approached with the min-cost network flow optimization techniques. Such methods simultaneously resolve multiple object tracks in a video and enable modeling of dependencies among tracks. Min-cost network flow methods also fit well within the "tracking-by-detection" paradigm where object trajectories are obtained by connecting per-frame outputs of an object detector. Object detectors, however, often fail due to occlusions and clutter in the video. To cope with such situations, we propose to add pairwise costs to the min-cost network flow framework. While integer solutions to such a problem become NP-hard, we design a convex relaxation solution with an efficient rounding heuristic which empirically gives certificates of small suboptimality. We evaluate two particular types of pairwise costs and demonstrate improvements over recent tracking methods in real-world video sequences.

For several decades, image restoration remains an active research topic in low-level computer vision and hence new approaches are constantly emerging. However, many recently proposed algorithms achieve state-of-the-art performance only at the expense of very high computation time, which clearly limits their practical relevance. In this work, we propose an effective approach with both high computational efficiency and high restoration quality. We extend conventional nonlinear reaction diffusion models by several parametrized linear filters as well as several parametrized influence functions. We propose to train the parameters of the filters and the influence functions through a loss based approach. Experiments show that our trained nonlinear reaction diffusion models largely benefit from the training of the parameters and finally lead to the best reported performance on common test datasets for image restoration. Due to their structural simplicity, our trained models are highly efficient and are also well-suited for parallel computation on GPUs.

We present an object-based co-segmentation method that takes advantage of depth data and is able to correctly handle noisy images in which the common foreground object is missing. With RGBD images, our method utilizes the depth channel to enhance identification of similar foreground objects via a proposed RGBD co-saliency map, as well as to improve detection of object-like regions and provide depth-based local features for region comparison. To accurately deal with noisy images where the common object appears more than or less than once, we formulate co-segmentation in a fully-connected graph structure together with mutual exclusion (mutex) constraints that prevent improper solutions. Experiments show that this object-based RGBD co-segmentation with mutex constraints outperforms related techniques on an RGBD co-segmentation dataset, while effectively processing noisy images. Moreover, we show that this method also provides performance comparable to state-of-the-art RGB co-segmentation techniques on regular RGB images with depth maps estimated from them.

This paper proposes a novel model and dataset for 3D scene flow estimation with an application to autonomous driving. Taking advantage of the fact that outdoor scenes often decompose into a small number of independently moving objects, we represent each element in the scene by its rigid motion parameters and each superpixel by a 3D plane as well as an index to the corresponding object. This minimal representation increases robustness and leads to a discrete-continuous CRF where the data term decomposes into pairwise potentials between superpixels and objects. Moreover, our model intrinsically segments the scene into its constituting dynamic components. We demonstrate the performance of our model on existing benchmarks as well as a novel realistic dataset with scene flow ground truth. We obtain this dataset by annotating 400 dynamic scenes from the KITTI raw data collection using detailed 3D CAD models for all vehicles in motion. Our experiments also reveal novel challenges which cannot be handled by existing methods.

Hierarchical segmentation based object proposal methods have become an important step in modern object detection paradigm. However, standard single-way hierarchical methods are fundamentally flawed in that the errors in early steps cannot be corrected and accumulate. In this work, we propose a novel multi-branch hierarchical segmentation approach that alleviates such problems by learning multiple merging strategies in each step in a complementary manner, such that errors in one merging strategy could be corrected by the others. Our approach achieves the state-of-the-art performance for both object proposal and object detection tasks, comparing to previous object proposal methods.

Object detection is always conducted by object proposal generation and classification sequentially. This paper handles object detection in a superpixel oriented manner instead of the proposal oriented. Specially, this paper takes object detection as a multi-label superpixel labeling problem by minimizing an energy function. It uses the data cost term to capture the appearance, smooth cost term to encode the spatial context and label cost term to favor compact detection. The data cost is learned through a convolutional neural network and the parameters in the labeling model are learned through a structural SVM. Compared with proposal generation and classification based methods, the proposed superpixel labeling method can naturally detect objects missed by proposal generation step and capture the global image context to infer the overlapping objects. The proposed method shows its advantage in Pascal VOC and ImageNet. Notably, it performs better than the ImageNet ILSVRC2014 winner GoogLeNet (45.0% V.S. 43.9% in mAP) with much shallower and fewer CNNs.

The following message is relayed from an update made on IEEE Xplore. Notice of Violation of IEEE Publication Principles "Dual Domain Filters Based Texture and Structure Preserving Image Non-Blind Deconvolution" by Hang Yang, Ming Zhu, Yan Niu, Yujing Guan, and Zhongbo Zhang in the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2015, pp. 705-713 After careful and considered review of the content and authorship of this paper by a duly constituted expert committee, this paper has been found to be in violation of IEEE’s Publication Principles. This paper copied portions of text from the papers cited below. The original text was copied without attribution (including appropriate references to the original author(s) and/or paper titles) and without permission. "Group-Based Sparse Representation for Image Restoration" by Jian Zhang, Debin Zhao, and Wen Gao in the IEEE Transactions on Image Processing, Vol 23, No 8, August 2014, pp. 3336-3351 "Dual-domain Image Denoising" by Claude Knaus, Matthias Zwicker in the Proceedings of the IEEE International Conference on Image Processing, (ICIP), September 2013, pp. 440-444 "A Machine Learning Approach for Non-blind Image Deconvolution" by Christian Schuler, Harold Christopher Burger, Stefan Harmeling, and Bernhard Scholkopf in the Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2013, pp. 1067-1074 Image deconvolution continues to be an active research topic of recovering a sharp image, given a blurry one generated by a convolution. One of the most challenging problems in image deconvolution is how to preserve the fine scale texture structures while removing blur and noise. Various methods have been implemented in both spatial and transform domains, such as gradient based methods, nonlocal self-similarity methods, sparsity based methods. However, each domain has its advantages and shortcomings, which can be complemented by each other. In this work we propose a new approach for efficient image deconvolution based on dual domain filters. In the deblurring process, we offer a hybrid method that a novel rolling guidance filter is used to ensure proper texture/structure separation, and then in the transform domain, we use the short-time Fourier transform to recover the textures while removing noise with energy shrinkage. Our hybrid algorithm that is surprisingly easy to implement, and experimental results clearly show that the proposed algorithm outperforms many state-of-the-art deconvolution algorithms in terms of both quantitative measure and visual perception quality.

Registration between images taken with different cameras, from different viewpoints or under different lighting conditions is a challenging problem. It needs to solve not only the geometric registration problem but also the photometric matching problem. In this paper, we propose to estimate the integrated geometric and photometric transformations between two images based on a local affine Fourier-moment matching framework, which is developed to achieve deformable registration. We combine the local Fourier moment constraints with the smoothness constraints to determine the local affine transforms in a hierarchal block model. Our experimental results on registering some real images related by large color and geometric transformations show the proposed registration algorithm provides superior image registration results compared to the state-of-the-art image registration methods.