Articulated Pose Estimation Using Discriminative Armlet Classifiers
Authors:
- Georgia GkioxariGGGeorgia Gkioxari
- Pablo ArbelaezPAPablo Arbelaez
- Lubomir BourdevLBLubomir Bourdev
- Jitendra MalikJMJitendra Malik
- 1 June 2013
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10636919,p. 3342-3349
Abstract
We propose a novel approach for human pose estimation in real-world cluttered scenes, and focus on the challenging problem of predicting the pose of both arms for each person in the image. For this purpose, we build on the notion of poselets [4] and train highly discriminative classifiers to differentiate among arm configurations, which we call armlets. We propose a rich representation which, in addition to standard HOG features, integrates the information of strong contours, skin color and contextual cues in a principled manner. Unlike existing methods, we evaluate our approach on a large subset of images from the PASCAL VOC detection dataset, where critical visual phenomena, such as occlusion, truncation, multiple instances and clutter are the norm. Our approach outperforms Yang and Ramanan [26], the state-of-the-art technique, with an improvement from 29.0% to 37.5% PCP accuracy on the arm keypoint prediction task, on this new pose estimation dataset.Keywords
This publication has 18 references indexed in Scilit:
- Exploring the Spatial Hierarchy of Mixture Models for Human Pose Estimation2012
- Unbiased look at dataset bias2011
- Learning hierarchical poselets for human parsing2011
- Learning effective human pose estimation from inaccurate annotation2011
- Articulated pose estimation with flexible mixtures-of-parts2011
- Better appearance models for pictorial structures2009
- Training Deformable Models for Localization2006
- Histograms of Oriented Gradients for Human Detection2005
- Fast pose estimation with parameter-sensitive hashing2003
- Efficient matching of pictorial structures2002