Enhanced hand part classification from a single depth image using random decision forests. Issue 8 (1st July 2016)

• Record Type: Journal Article
• Authors: Sohn, Myoung‐Kyu; Lee, Sang‐Heon; Kim, Hyunduk; Park, Hyeyoung
• Is Part Of: IET computer vision. Volume 10:Issue 8(2016)
• Page Start: 861
• Page End: 867
• Publication Date: 2016-07-01
• Subjects: palmprint recognition -- feature extraction -- pose estimation -- gesture recognition -- image classification -- trees (mathematics); enhanced hand part classification -- single-depth image -- random decision forests -- hand pose recognition -- human-computer interaction -- low-cost three-dimensional cameras -- low-cost 3D cameras -- natural gestures -- RDF -- foreground pixel -- robust feature extraction method -- per-pixel classification -- tree selection algorithm -- legacy-trained RDF -- unseen test data classification -- label image synthesis -- 3D hand mesh model; Computer vision -- Periodicals; Pattern recognition systems -- Periodicals; 006.37
• DOI: 10.1049/iet-cvi.2015.0239
• Languages: English
• ISSNs: 1751-9632

Abstract

Abstract : Hand pose recognition has received increasing attention in an area of human–computer interaction. With the recent spread of many low‐cost three‐dimensional (3D) cameras, research into understanding more natural gestures has increased. In this study, the authors present a method for hand part classification and joint estimation from a single depth image. They apply random decision forests (RDFs) for hand part classification. Foreground pixels in the hand image are estimated by RDF. Then hand joints are estimated based on the classified hand parts. They suggest a robust feature extraction method for per‐pixel classification, which enhances the accuracy of hand part classification. They also propose a tree selection algorithm using legacy trained RDF to classify unseen test data. Selecting trees using the proposed method show better performance than using all the trees as in conventional method. Depth images and label images synthesised by 3D hand mesh model were used for training forests and algorithm verification. The authors' experiments show that the enhanced algorithm outperforms the state‐of‐the‐art method in accuracy.