Snakes, shapes, and gradient vector flow
IEEE Transactions on Image Processing
Hi-index | 0.00 |
This article presents a study about two popular segmentation and tracking techniques in computer vision, skin segmentation and snakes, and their application to a particular case. The study is centered in skin segmentation over a wooden background mostly. At first wood and skin characteristics were thought to be very similar, what would have been an inconvenient. However, this paper shows that it is possible to get a satisfactory colour segmentation applying a Bayes classifier in YCbCr images format. The use of automatically initialized open active contours let the tracking of the segmented hand and its localization in the image.