A Bayesian framework for human body pose tracking from depth image sequences.

This paper addresses the problem of accurate and robust tracking of 3D human body pose from depth image sequences. Recovering the large number of degrees of freedom in human body movements from a depth image sequence is challenging due to the need to resolve the depth ambiguity caused by self-occlusions and the difficulty to recover from tracking failure. Human body poses could be estimated through model fitting using dense correspondences between depth data and an articulated human model (local optimization method). Although it usually achieves a high accuracy due to dense correspondences, it may fail to recover from tracking failure. Alternately, human pose may be reconstructed by detecting and tracking human body anatomical landmarks (key-points) based on low-level depth image analysis. While this method (key-point based method) is robust and recovers from tracking failure, its pose estimation accuracy depends solely on image-based localization accuracy of key-points. To address these limitations, we present a flexible Bayesian framework for integrating pose estimation results obtained by methods based on key-points and local optimization. Experimental results are shown and performance comparison is presented to demonstrate the effectiveness of the proposed approach.

Development of a controlled vocabulary and software application to analyze fruit shape variation in tomato and other plant species.

The domestication and improvement of fruit-bearing crops resulted in a large diversity of fruit form. To facilitate consistent terminology pertaining to shape, a controlled vocabulary focusing specifically on fruit shape traits was developed. Mathematical equations were established for the attributes so that objective, quantitative measurements of fruit shape could be conducted. The controlled vocabulary and equations were integrated into a newly developed software application, Tomato Analyzer, which conducts semiautomatic phenotypic measurements. To demonstrate the utility of Tomato Analyzer in the detection of shape variation, fruit from two F2 populations of tomato (Solanum spp.) were analyzed. Principal components analysis was used to identify the traits that best described shape variation within as well as between the two populations. The three principal components were analyzed as traits, and several significant quantitative trait loci (QTL) were identified in both populations. The usefulness and flexibility of the software was further demonstrated by analyzing the distal fruit end angle of fruit at various user-defined settings. Results of the QTL analyses indicated that significance levels of detected QTL were greatly improved by selecting the setting that maximized phenotypic variation in a given population. Tomato Analyzer was also applied to conduct phenotypic analyses of fruit from several other species, demonstrating that many of the algorithms developed for tomato could be readily applied to other plants. The controlled vocabulary, algorithms, and software application presented herein will provide plant scientists with novel tools to consistently, accurately, and efficiently describe two-dimensional fruit shapes.