In this paper, we propose a low-cost posture recognition scheme using a single\nwebcam for the signaling hand with nature sways and possible occlusions.\nIt goes for developing the untouchable low-complexity utility based on\nfriendly hand-posture signaling. The scheme integrates the dominant temporal-\ndifference detection, skin color detection and morphological filtering\nfor efficient cooperation in constructing the hand profile molds. Those molds\nprovide representative hand profiles for more stable posture recognition than\naccurate hand shapes with in effect trivial details. The resultant bounding box\nof tracking the signaling molds can be treated as a regular-type object-matched\nROI to facilitate the stable extraction of robust HOG features. With such\ncommonly applied features on hand, the prototype SVM is adequately capable\nof obtaining fast and stable hand postures recognition under natural hand\nmovement and non-hand object occlusion. Experimental results demonstrate\nthat our scheme can achieve hand-posture recognition with enough accuracy\nunder background clutters that the targeted hand can be allowed with medium\nmovement and palm-grasped object. Hence, the proposed method can\nbe easily embedded in the mobile phone as application software.
Loading....