The dynamic factor of tongue diagnosis is an important parameter during the diagnosis of the traditional Chinese medicine (TCM). To accurately analyze the properties of the dynamic tongue, we must track the tongue. Automatic tracking method for dynamic tongue has been proposed. We improve a new term of external forces for original Snake. Multi-resolution wavelet, Gabor decomposition is used to generate edge potential field of the external forces. The contour of tongue in initial frame is initialized through the feature of tongue in HSV color space, and the displacement vectors between the two frames are got by Gabor jets. The rough contour of tongue in the next frame is attained by the contour of the initial frame plus the displacement vectors. In order to find the accurate object in the next frame, the accurate contour of the tongue is obtained by the iterations of improved snake. The experimental results using different types of video sequences are presented to demonstrate the efficiency of our proposed algorithm.
Published in:
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
(Volume:9
)
Date of Conference: 18-21 Aug. 2005