An adaptable system for RGB-D based human body detection and pose estimation

Sparse Pose Regression via Componentwise Clustering Feature Point Representation

“…Lallemand et al [11] integrate the activity information with pose regression, and formulate the problem as a joint regression-classification task which recovers the 3D body pose and classifies the performed activity. Buys et al [3] associate each pixel to a body component using randomized decision forests, and cluster all pixels into more robust component estimation. Then the best feasible kinematic model is matched with the coarse pose, and an accurate pose is obtained by estimating an appearance model and segmentation refinement.…”

“…According to the methodology, research work on pose estimation via depth data can be roughly divided into four categories: the feature point approaches, the pixel classification approaches, the probabilistic graphical model approaches and others. In general, the feature point approaches [12], [1], [13], [8], [7] detect interest points of the range image ( with the extrema of geodesic distances of depth points), and then employ a database lookup scheme (e.g., [1]), classify each interesting point as a point of a body component using its local descriptors (e.g., [12]), or detect additional feature points as the secondary feature (e.g., [1], [13]) for discriminating human poses; the pixel classification approaches [14], [15], [18], [20], [11], [3], [9] determine which body component each pixel belongs to, using the so-called depth difference feature and random decision forests; the probabilistic graphical model approaches [26], [27], [6], [23] represent a human body as a set of rigid or non-rigid parts connected with pairwise constraints and formulate the localization of body parts as an optimization of the prior probability of part configurations. Different approaches fit for different applications of visual media, i.e., the codec and reconstruction of models of human body, the retrieval and analysis of human motion.…”

“…Typically, Shotton et al [14], [15] integrate a simple but efficient feature: the depth difference of pairwise offset pixels, with the powerful randomized decision forests (RDF), and produce realtime and accurate pose recognition. Afterwards, several pixel classification approaches [11], [3] follow the idea and make various improvements. However, those approaches require large, varied offline training data for obtaining the decision forests, as well as many parameters (e.g., the number and depth of trees, the size of sampling window, and distribution of pairwise offsets within the window).…”

An ultra-fast human detection method for color-depth camera

“…An effective way to counter the challenges is to employ depth information [14,15,16,17]. The depth image produced by stereo rigs or depth cameras significantly simplifies the task of real-time people detection, but there are still difficulties in dealing with occluded subjects in horizontal front view images.…”

Toward an affordable and user-friendly visual motion capture system

“…Their general conclusion is that the KS embedded marker-less methods are not reliable and accurate enough for quantitative evaluation of human motion. Other groups have developed their own algorithms using one or multiple KSs [6]. Marker-less pose estimation from multi-view video has been a long-standing problem in computer Research supported by the Japan Society for Promotion of Science and partially by the author's respective institutions.…”