ICCV 2015 Papers

We propose a data-driven approach for intrinsic image decomposition, which is the process of inferring the confounding factors of reflectance and shading in an image. We pose this as a two-stage learning problem. First, we train a model to predict relative reflectance ordering be- tween image patches ('brighter', 'darker', 'same') from large-scale human annotations, producing a data-driven reflectance prior. Second, we show how to naturally integrate this learned prior into existing energy minimization frame- works for intrinsic image decomposition. We compare our method to the state-of-the-art approach of Bell et al. [7] on both decomposition and image relighting tasks, demonstrating the benefits of the simple relative reflectance prior, especially for scenes under challenging lighting conditions.

This paper presents our work on "SNaCK," a low-dimensional concept embedding algorithm that combines human expertise with automatic machine similarity kernels. Both parts are complimentary: human insight can capture relationships that are not apparent from the object's visual similarity and the machine can help relieve the human from having to exhaustively specify many constraints. We show that our SNaCK embeddings are useful in several tasks: distinguishing prime and nonprime numbers on MNIST, discovering labeling mistakes in the Caltech UCSD Birds (CUB) dataset with the help of deep-learned features, creating training datasets for bird classifiers, capturing subjective human taste on a new dataset of 10,000 foods, and qualitatively exploring an unstructured set of pictographic characters. Comparisons with the state-of-the-art in these tasks show that SNaCK produces better concept embeddings that require less human supervision than the leading methods.

Common sense is essential for building intelligent machines. While some commonsense knowledge is explicitly stated in human-generated text and can be learnt by mining the web, much of it is unwritten. It is often unnecessary and even unnatural to write about commonsense facts. While unwritten, this commonsense knowledge is not unseen! The visual world around us is full of structure modeled by commonsense knowledge. Can machines learn common sense simply by observing our visual world? Unfortunately, this requires automatic and accurate detection of objects, their attributes, poses, and interactions between objects, which remain challenging problems. Our key insight is that while visual common sense is depicted in visual content, it is the semantic features that are relevant and not low-level pixel information. In other words, photorealism is not necessary to learn common sense. We explore the use of human-generated abstract scenes made from clipart for learning common sense. In particular, we reason about the plausibility of an interaction or relation between a pair of nouns by measuring the similarity of the relation and nouns with other relations and nouns we have seen in abstract scenes. We show that the commonsense knowledge we learn is complementary to what can be learnt from sources of text.

Binary coding or hashing techniques are recognized to accomplish efficient near neighbor search, and have thus attracted broad interests in the recent vision and learning studies. However, such studies have rarely been dedicated to Maximum Inner Product Search (MIPS), which plays a critical role in various vision applications. In this paper, we investigate learning binary codes to exclusively handle the MIPS problem. Inspired by the latest advance in asymmetric hashing schemes, we propose an asymmetric binary code learning framework based on inner product fitting. Specifically, two sets of coding functions are learned such that the inner products between their generated binary codes can reveal the inner products between original data vectors. We also propose an alternative simpler objective which maximizes the correlations between the inner products of the produced binary codes and raw data vectors. In both objectives, the binary codes and coding functions are simultaneously learned without continuous relaxations, which is the key to achieving high-quality binary codes. We evaluate the proposed method, dubbed Asymmetric Inner-product Binary Coding (AIBC), relying on the two objectives on several large-scale image datasets. Both of them are superior to the state-of-the-art binary coding and hashing methods in performing MIPS tasks.

Analysis-by-synthesis has been a successful approach for many tasks in computer vision, such as 6D pose estimation of an object in an RGB-D image which is the topic of this work. The idea is to compare the observation with the output of a forward process, such as a rendered image of the object of interest in a particular pose. Due to occlusion or complicated sensor noise, it can be difficult to perform this comparison in a meaningful way. We propose an approach that ``learns to compare'', while taking these difficulties into account. This is done by describing the posterior density of a particular object pose with a convolutional neural network (CNN) that compares observed and rendered images. The network is trained with the maximum likelihood paradigm. We observe empirically that the CNN does not specialize to the geometry or appearance of specific objects. It can be used with objects of vastly different shapes and appearances, and in different backgrounds. Compared to state-of-the-art, we demonstrate a significant improvement on two different datasets which include a total of eleven objects, cluttered background, and heavy occlusion.

3D motion estimation is a fundamental problem with many computer vision applications. With the emergence of modern, affordable and increasingly accurate RGB-D sensors, single view approaches for estimating 3D motion, also known as scene flow, are becoming popular. In this paper we propose a novel coarse to fine correspondence-based scene flow approach to account for the effects of large displacements and to model occlusion, based on explicit geometric reasoning. Our methodology enforces piecewise motion rigidity at the level of the depth point cloud without explicitly smoothing the parameters of adjacent neighborhoods. By integrating all geometric and photometric components in a single, consistent, occlusion-aware energy model our method is able to deal with fast motions and large occlusions areas, as present in challenging datasets like MPI Sintel Flow Dataset, which have recently been augmented with depth information. By explicitly modeling large displacements and occlusion, we can now more successfully work with difficult sequences which cannot be currently processed by state of the art scene flow methods that rely on small inter-frame motion assumptions. We also show that by leveraging depth information, we can obtain superior correspondence fields compared to the best state of the art large-displacement (2D) optical flow methods.

The goal of this work is to bring semantics into the tasks of text recognition and retrieval in natural images. Although text recognition and retrieval have received a lot of attention in recent years, previous works have focused on recognizing or retrieving exactly the same word used as a query, without taking the In this paper, we ask the following question: can we predict semantic concepts directly from a word image, without explicitly trying to transcribe the word image or its characters at any point? For this goal we propose a convolutional neural network (CNN) with a weighted ranking loss objective that ensures that the concepts relevant to the query image are ranked ahead of those that are not relevant. This can also be interpreted as learning a Euclidean space where word images and concepts are jointly embedded. This model is learned in an end-to-end manner, from image pixels to semantic concepts, using a dataset of synthetically generated word images and concepts mined from a lexical database (WordNet). Our results show that, despite the complexity of the task, word images and concepts can indeed be associated with a high degree of accuracy.

We explore the problem of predicting "just noticeable differences" in a visual attribute. While some pairs of images have a clear ordering for an attribute (e.g., A is more sporty than B), for others the difference may be indistinguishable to human observers. However, existing relative attribute models are unequipped to infer partial orders on novel data. Attempting to map relative attribute ranks to equality predictions is non-trivial, particularly since the span of indistinguishable pairs in attribute space may vary in different parts of the feature space. We develop a Bayesian local learning strategy to infer when images are indistinguishable for a given attribute. On the UT-Zap50K shoes and LFW-10 faces datasets, we outperform a variety of alternative methods. In addition, we show the practical impact on fine-grained visual search.

In this paper, we revisit the joint probabilistic data association (JPDA) technique and propose a novel solution based on recent developments in finding the m-best solutions to an integer linear program. The key advantage of this approach is that it makes JPDA computationally tractable in applications with high target and/or clutter density, such as spot tracking in fluorescence microscopy sequences and pedestrian tracking in surveillance footage. We also show that our JPDA algorithm embedded in a simple tracking framework is surprisingly competitive with state-of-the-art global tracking methods in these two applications, while needing considerably less processing time.

Binary energy optimization is a popular approach for segmenting a color image into foreground/background regions. To model the appearance of the regions, color, a relatively high dimensional feature, should be handled effectively. A full color histogram is usually too sparse to be reliable. One approach is to explicitly reduce dimensionality by clustering or quantizing the color space. Another popular approach is to fit GMMs for soft implicit clustering of the color space. These approaches work well when the foreground/background are sufficiently distinct. In cases of more subtle difference in appearance, both approaches may reduce or even eliminate foreground/background distinction. This happens because either color clustering is performed completely independently from the segmentation process, as a preprocessing step (in clustering), or independently for the foreground and independently for the background (in GMM). We propose to make clustering an integral part of segmentation, by including a new clustering term in the energy function. Our energy function with a clustering term favours clusterings that make foreground/background appearance more distinct. Thus our energy function jointly optimizes over color clustering, foreground/background models, and segmentation. Exact optimization is not feasible, therefore we develop an approximate algorithm. We show the advantage of including the color clustering term into the energy function on camouflage images, as well as standard segmentation datasets.

Segmenting semantic objects from images and parsing them into their respective semantic parts are fundamental steps towards detailed object understanding in computer vision. In this paper, we propose a joint solution that tackles semantic object and part segmentation simultaneously, in which higher object-level context is provided to guide part segmentation, and more detailed part-level localization is utilized to refine object segmentation. Specifically, we first introduce the concept of semantic compositional parts (SCP) in which similar semantic parts are grouped and shared among different objects. A two-stream fully convolutional network (FCN) is then trained to provide the SCP and object potentials at each pixel. At the same time, a compact set of segments can also be obtained from the SCP predictions of the network. Given the potentials and the generated segments, in order to explore long-range context, we finally construct an efficient fully connected conditional random field (FCRF) to jointly predict the final object and part labels. Extensive evaluation on three different datasets shows that our approach can mutually enhance the performance of object and part segmentation, and outperforms the current state-of-the-art on both tasks.

Temporal information has useful features for recognizing facial expressions. However, to manually design useful features requires a lot of effort. In this paper, to reduce this effort, a deep learning technique, which is regarded as a tool to automatically extract useful features from raw data, is adopted. Our deep network is based on two different models. The first deep network extracts temporal appearance features from image sequences, while the other deep network extracts temporal geometry features from temporal facial landmark points. These two models are combined using a new integration method in order to boost the performance of the facial expression recognition. Through several experiments, we show that the two models cooperate with each other. As a result, we achieve superior performance to other state-of-the-art methods in the CK+ and Oulu-CASIA databases. Furthermore, we show that our new integration method gives more accurate results than traditional methods, such as a weighted summation and a feature concatenation method.

In this paper, we propose an optimal decomposition approach to large-scale multi-view stereo from an initial sparse reconstruction. The success of the approach depends on the introduction of surface-segmentation-based camera clustering rather than sparse-point-based camera clustering, which suffers from the problems of non-uniform reconstruction coverage ratio and high redundancy. In details, we introduce three criteria for camera clustering and surface segmentation for reconstruction, and then we formulate these criteria into an energy minimization problem under constraints. To solve this problem, we propose a joint optimization in a hierarchical framework to obtain the final surface segments and corresponding optimal camera clusters. On each level of the hierarchical framework, the camera clustering problem is formulated as a parameter estimation problem of a probability model solved by a General Expectation-Maximization algorithm and the surface segmentation problem is formulated as a Markov Random Field model based on the probability estimated by the previous camera clustering process. The experiments on several Internet datasets and aerial photo datasets demonstrate that the proposed approach method generates more uniform and complete dense reconstruction with less redundancy, resulting in more efficient multi-view stereo algorithm.

We propose an Active Learning approach to training a segmentation classifier that exploits geometric priors to streamline the annotation process in 3D image volumes. To this end, we use these priors not only to select voxels most in need of annotation but to guarantee that they lie on 2D planar patch, which makes it much easier to annotate than if they were randomly distributed in the volume. A simplified version of this approach is effective in natural 2D images. We evaluated our approach on Electron Microscopy and Magnetic Resonance image volumes, as well as on natural images. Comparing our approach against several accepted baselines demonstrates a marked performance increase.

In this paper, we address the problem of computing an intrinsic decomposition of the colors of a surface into an albedo and a shading term. The surface is reconstructed from a single or multiple RGB-D images of a static scene obtained from different views. We thereby extend and improve existing works in the area of intrinsic image decomposition. In a variational framework, we formulate the problem as a minimization of an energy composed of two terms: a data term and a regularity term. The first term is related to the image formation process and expresses the relation between the albedo, the surface normals, and the incident illumination. We use an affine shading model, a combination of a Lambertian model, and an ambient lighting term. This model is relevant for Lambertian surfaces. When available, multiple views can be used to handle view-dependent non-Lambertian reflections. The second term contains an efficient combination of l2 and l1-regularizers on the illumination vector field and albedo respectively. Unlike most previous approaches, especially Retinex-like techniques, these terms do not depend on the image gradient or texture, thus reducing the mixing shading/reflectance artifacts and leading to better results. The obtained non-linear optimization problem is efficiently solved using a cyclic block coordinate descent algorithm. Our method outperforms a range of state-of-the-art algorithms on a popular benchmark dataset.

We formulate the estimation of dense depth maps from video sequences as a problem of intrinsic image estimation. Our approach synergistically integrates the estimation of multiple intrinsic images including depth, albedo, shading, optical flow, and surface contours. We build upon an example-based framework for depth estimation that uses label transfer from a database of RGB and depth pairs. We combine this with a method that extracts consistent albedo and shading from video. In contrast to raw RGB values, albedo and shading provide a richer, more physical, foundation for depth transfer. Additionally we train a new contour detector to predict surface boundaries from albedo, shading, and pixel values and use this to improve the estimation of depth boundaries. We also integrate sparse structure from motion with our method to improve the metric accuracy of the estimated depth maps. We evaluate our Intrinsic Depth method quantitatively by estimating depth from videos in the NYU RGB-D and SUN3D datasets. We find that combining the estimation of multiple intrinsic images improves depth estimation relative to the baseline method.

We present a method for intrinsic image decomposition, which aims to decompose images into reflectance and shading layers. Our input is a sequence of images with varying illumination acquired by a static camera, e.g. an indoor scene with a moving light source or an outdoor timelapse. We leverage the local color variations observed over time to infer constraints on the reflectance and solve the ill-posed image decomposition problem. In particular, we derive an adaptive local energy from the observations of each local neighborhood over time, and integrate distant pairwise constraints to enforce coherent decomposition across all surfaces with consistent shading changes. Our method is solely based on multiple observations of a Lambertian scene under varying illumination and does not require user interaction, scene geometry, or an explicit lighting model. We compare our results with several intrinsic decomposition methods on a number of synthetic and captured datasets.

Probability density functions (PDFs) are fundamental "objects" in mathematics with numerous applications in computer vision, machine learning and medical imaging. The feasibility of basic operations such as computing the distance between two PDFs and estimating a mean of a set of PDFs is a direct function of the representation we choose to work with. In this paper, we study the Gaussian mixture model (GMM) representation of the PDFs motivated by its numerous attractive features. (1) GMMs are arguably more interpretable than, say, square root parameterizations (2) the model complexity can be explicitly controlled by the number of components and (3) they are already widely used in many applications. The main contributions of this paper are numerical algorithms to enable basic operations on such objects that strictly respect their underlying geometry. For instance, when operating with a set of k component GMMs, a first order expectation is that the result of simple operations like interpolation and averaging should provide an object that is also a k component GMM. The literature provides very little guidance on enforcing such requirements systematically. It turns out that these tasks are important internal modules for analysis and processing of a field of ensemble average propagators (EAPs), common in diffusion weighted magnetic resonance imaging. We provide proof of principle experiments showing how the proposed algorithms for interpolation can facilitate statistical analysis of such data, essential to many neuroimaging studies. Separately, we also derive interesting connections of our algorithm with functional spaces of Gaussians, that may be of independent interest.

In this paper we present a method of using standard multi-view images for 3D surface reconstruction of non-Lambertian objects. We extend the original visual hull concept to incorporate 3D cues presented by internal occluding contours, i.e., occluding contours that are inside the object's silhouettes. We discovered that these internal contours, which are results of convex parts on an object's surface, can lead to a tighter fit than the original visual hull. We formulated a new visual hull refinement scheme - Locally Convex Carving that can completely reconstruct concavity caused by two or more intersecting convex surfaces. In addition we develop a novel approach for contour tracking given labeled contours in sparse key frames. It is designed specifically for highly specular or transparent objects, for which assumptions made in traditional contour detection/tracking methods, such as highest gradient and stationary texture edges, are no longer valid. It is formulated as an energy minimization function where several novel terms are developed to increase robustness. Based on the two core algorithms, we have developed an interactive system for 3D modeling. We have validated our system, both quantitatively and qualitatively, with four datasets of different object materials. Results show that we are able to generate visually pleasing models for very challenging cases.