[A new method of automatic analysis of tongue deviation using self-correction].

The article analyzes the old analysis method of tongue deviation and introduces a new analysis method of it with self-correction avoiding the shortcomings of the old method. In this article, comparisons and analyses are made to current central axis extraction methods, and the facts proved that these methods were not suitable for central axis extraction of tongue images. To overcome the shortcoming that the old method utilized area symmetry to extract central axis so that it would lead to a failure to find central axis, we introduced a kind of shape symmetry analysis method to extract the central axis. This method was capable of correcting the edge of tongue root automatically, and it improved the accuracy of central axis extraction. In additional, in this article, a kind of mouth corner analysis method by analysis of variational hue of tongue images was introduced. In the experiment for comparison, this method was more accurate than the old one and its efficiency was higher than that of the old one.