Research Article
An Adaptive Superpixel Based Hand Gesture Tracking and Recognition System
Algorithm 2
Adaptive superpixel based hand gesture tracking.
For each frame to the end | Normal hand tracking | Input: frame , | (1) SLIC get superpixels on surrounding of . | (2) For each superpixel , Compute theYCbCr histgram , and confidence map using (6). | (3) Sample candidates around with , discard unproper samples using (7) and (8). | (4) Calculate the motion parameter for each using (9). | (5) Calculate the likelihood for each using (10). | (6) Get the best match of hand with MAP estimate on and using (11). | Output: hand detection in frame t. | Failure recovery and updating | Input: current hand detection | (1) Check the occurance of occlusion with threshold. Calculate similarity matrix using (4). Detect the hand location to recover the occlusion. | (2) Check the occurance of background confusion using (12) and re-track frames to recover . | (3) In case of background confusion, sample all detections of re-tracking () and discard all previous samples. | (4) In case of occlusion or no failure, use one sample for every frames to replace a previous sample using (13). | (5) Replace the appearance model every frames by re-train on new samples. | Output: recovered hand detection if normal hand tracking fails, and new hand appearance model . |
|