Open AccessEXTRACTION OF STANDING HUMAN BODIES FROM IMAGES WITH MULTI LEVEL SEGMENTATION AND SPLINE REGRESSION
Author(s) -
Rotti Srinivasamurthy Swathi,
K. Ramesh
Publication year - 2016
Publication title -
zenodo (cern european organization for nuclear research)
Language(s) - English
DOI - 10.5281/zenodo.163304
Subject(s) - spline (mechanical) , segmentation , artificial intelligence , extraction (chemistry) , regression , computer vision , computer science , mathematics , pattern recognition (psychology) , statistics , engineering , chromatography , chemistry , structural engineering
The ability to extract and detection of human activities by computer vision is very important with many potential applications. Extraction of human bodies from images from respective digital image has accomplished consideration in recent times and extensive variety of research is carried on to meet the sought result. In this paper proposed a novel strategy to extract human bodies from images where the scene density, the highly dimensional pose space, and various human appearances are handled in better way compared to conventional state of art methods. The proposed approach is ordered into five distinct strides (i) face detection, (ii) Multiple level segmentation, (iii) skin detection, (iv) upper body segmentation and (v) lower body segmentation respectively. In this paper we propose another method spline regression to extraction of human bodies from single images. Finally the simulation results have achieved better performance and high proficiency over traditional state of art methods