formulate the hu-man parsing and pose estimation problem jointly within a unified framework via a tailored And-Or graph. A novel Grid Layout Feature is then designed to effectively cap-ture the spatial co-occurrence/occlusion information be-tween/within the Parselets and MJGTs. Thus the mutually

information, Kalman tracking using motion models, and face tracking. Keywords: non-rigid object tracking; target localization and representation; spatially-smooth similarity function; Bhattacharyya coefficient; face tracking. 1

A computer vision system has been implemented that can recognize threedimensional objects from unknown viewpoints in single gray-scale images. Unlike most other approaches, the recognition is accomplished without any attempt to reconstruct depth information bottom-up from the visual input. Instead

whose voxels are traversed in a fixed visibility ordering. This strategy takes full account of occlusions and allows the input cameras to be far apart and widely distributed about the environment. The algorithm identifies a special set of invariant voxels which together form a spatial and photometric

are also considered. First, the textures are modeled with volume local binary patterns (VLBP), which are an extension of the LBP operator widely used in ordinary texture analysis, combining motion and appearance. To make the approach computationally simple and easy to extend, only the co-occurrences

relatively fixed spatial configuration. We report experiments on images of side views of cars. Our experiments show that the method achieves high detection accuracy on a difficult test set of real-world images, and is highly robust to partial occlusion and background variation. In addition, we discuss

histogram-based algorithms [8, 6]. First, by robustly combining multiple patch votes, we are able to handle partial occlusions or pose change. Second, the geometric relations between the template patches allow us to take into account the spatial distribution of the pixel intensities- information which

documents. We present a technique called conceptual grouping that automatically distinguishes be-tween different meanings of a user query, given a document collection. By analysing a word co-occurrence network of a text database, we are able to form groups of words related to the query, grouped by semantic

initio at each pixel it is very CPU intensive and may contain noticeable aliasing artifacts. It is difficult to take advantage of spatial coherence because the shapes of reflections and refractions from curved surfaces are so complex. In this paper we describe an algorithm that utilizes the spatial

Abstract—Indoor place categorization is an important ca-pability for service robots working and interacting in human environments. This paper presents a new place categorization method which uses information about the spatial correlation be-tween the different image modalities provided by RGB