In this work we show how to combine bottom-up and topdown approaches into a single figure-ground segmentation process. This process provides accurate delineation of object boundaries that cannot be achieved by either the topdown or bottom-up approach alone. The top-down approach uses object representation learned from examples to detect an object in a given input image and provide an approximation to its figure-ground segmentation. The bottomup approach uses image-based criteria to define coherent groups of pixels that are likely to belong together to either the figure or the background part. The combination provides a final segmentation that draws on the relative merits of both approaches: The result is as close as possible to the top-down approximation, but is also constrained by the bottom-up process to be consistent with significant image discontinuities. We construct a global cost function that represents these top-down and bottom-up requirements. We then show how the global minimum of this function can be efficiently found by applying the sum-product algorithm. This algorithm also provides a confidence map that can be used to identify image regions where additional top-down or bottom-up information may further improve the segmentation. Our experiments show that the results derived from the algorithm are superior to results given by a pure top-down or pure bottom-up approach. The scheme has broad applicability, enabling the combined use of a range of existing bottom-up and top-down segmentations. 1.

Object identification is a specialized type of recognition in which the category (e.g. cars) is known and the goal is to recognize an object’s exact identity (e.g. Bob’s BMW). Two special challenges characterize object identification. First, inter-object variation is often small (many cars look alike) and may be dwarfed by illumination or pose changes. Second, there may be many different instances of the category but few or just one positive “training ” examples per object instance. Because variation among object instances may be small, a solution must locate possibly subtle object-specific salient features, like a door handle, while avoiding distracting ones such as specular highlights. With just one training example per object instance, however, standard modeling and feature selection techniques cannot be used. We describe an on-line algorithm that takes one image from a known category and builds an efficient “same ” versus “different ” classification cascade by predicting the most discriminative features for that object instance. Our method not only estimates the saliency and scoring function for each candidate feature, but also models the dependency between features, building an ordered sequence of discriminative features specific to the given image. Learned stopping thresholds make the identifier very efficient. To make this possible, category-specific characteristics are learned automatically in an off-line training procedure from labeled image pairs of the category. Our method, using the same algorithm for both cars and faces, outperforms a wide variety of other methods. 1.

We investigate the learning of the appearance of an object from a single image of it. Instead of using a large number of pictures of the object to recognize, we use a labeled reference database of pictures of other objects to learn invariance to noise and variations in pose and illumination.

Abstract not found

This paper proposes a novel approach for facial expression recognition by boosting Local Binary Patterns (LBP) based classifiers. Low-cost LBP features are introduced to effectively describle local features of face images. A novel learning procedure, Conditional Mutual Infomation based Boosting (CMIB), is proposed. CMIB learns a sequence of weak classifiers that maximize their mutual information about a candidate class, conditional to the response of any weak classifier already selected; a strong classifier is constructed by combining the learned weak classifiers using the Naive-Bayes. Extensive experiments on the Cohn-Kanade database illustrated that LBP features are effective for expression analysis, and CMIB enables much faster training than AdaBoost, and yields a classifier of improved classification performance.

We show that the discrimination between visually similar classes often depends on the detection of socalled ‘satellite features’. These are local features which are not informative by themselves, and can only be detected reliably at locations specified relative to other features. This makes satellite features difficult to extract by current classification methods. We describe a novel scheme which can extract discriminative satellite features and use them to distinguish between visually similar classes. The algorithm first searches for a set of features (“anchor features”) that can be found in all the similar classes. Such features can be detected because the classes are visually similar. The anchors are used to determine the locations of satellite features, which are extracted during learning and used in classification to distinguish between the similar classes. The algorithm is fully automatic, and is shown to work well for many categories of visually similar classes. 1.

Abstract. Cascades of boosted ensembles have become popular in the object detection community following their highly successful introduction in the face detector of Viola and Jones [1]. In this paper, we explore several aspects of this architecture that have not yet received adequate attention: decision points of cascade stages, faster ensemble learning, and stronger weak hypotheses. We present a novel strategy to determine the appropriate balance between false positive and detection rates in the individual stages of the cascade based on a probablistic model of the overall cascade’s performance. To improve the training time of individual stages, we explore the use of feature filtering before the application of Adaboost. Finally, we show that the use of stronger weak hypotheses based on CART can significantly improve upon the standard face detection results on the CMU-MIT data set. 1

Low-complexity feature selection is analyzed in the context of visual recognition. It is hypothesized that high-order dependences of bandpass features contain little information for discrimination of natural images. This hypothesis is characterized formally by the introduction of the concepts of conjunctive interference and decomposability order of a feature set. Necessary and sufficient conditions for the feasibility of low-complexity feature selection are then derived in terms of these concepts. It is shown that the intrinsic complexity of feature selection is determined by the decomposability order of the feature set and not its dimension. Feature selection algorithms are then derived for all levels of complexity and are shown to be approximated by existing information-theoretic methods, which they consistently outperform. The new algorithms are also used to objectively test the hypothesis of low decomposability order through comparison of classification performance. It is shown that, for image classification, the gain of modeling feature dependencies has strongly diminishing returns: best results are obtained under the assumption of decomposability order 1. This suggests a generic law for bandpass features extracted from natural images: that the effect, on the dependence of any two features, of observing any other feature is constant across image classes.

We introduce a new model addressing feature selection from a large dictionary of variables that can be computed from a signal or an image. Features are extracted according to an efficiency criterion, on the basis of specified classification or recognition tasks. This is done by estimating a probability distribution P on the complete dictionary, which distributes its mass over the more efficient, or informative, components. We implement a stochastic gradient descent algorithm, using the probability as a state variable and optimizing a multi-task goodness of fit criterion for classifiers based on variable randomly chosen according to P. We then generate classifiers from the optimal distribution of weights learned on the training set. The method is first tested on several pattern recognition problems including face detection, handwritten digit recognition, spam classification and micro-array analysis. We then compare our approach with other step-wise algorithms like random forests or recursive feature elimination. Keywords: stochastic learning algorithms, Robbins-Monro application, pattern recognition, classification algorithm, feature selection

In this paper, we present a methodology for recognizing seated postures using data from pressure sensors installed on a chair. Information about seated postures could be used to help avoid adverse effects of sitting for long periods of time or to predict seated activities for a human-computer interface. Our system design displays accurate near-real-time classification performance on data from subjects on which the posture recognition system was not trained by using a set of carefully designed, subject-invariant signal features. By using a near-optimal sensor placement strategy, we keep the number of required sensors low thereby reducing cost and computational complexity. We evaluated the performance of our technology using a series of empirical methods including (1) cross-validation (classification accuracy of 87 % for ten postures using data from 31 sensors), and (2) a physical deployment of our system (78 % classification accuracy using data from 19 sensors).