舌诊是中医四诊中望诊的重要内容,其本身就体现了信息融合的思想,然而在舌象信息的全面综合识别、分析、处理等方面鲜见文献报道。本文首次采用颜色与纹理特征对舌象老嫩进行自动识别。将舌象的颜色与纹理特征融合后,主要采用基于决策层的最优线性融合方法和AdaBoost算法两种融合方案,并对其进行比较。实验结果表明,基于k近邻分类器的AdaBoost算法对舌象老嫩融合特征具有最佳的分类性能。
Tongue diagnosis is one of the essential methods of Traditional Chinese Medical diagnosis. Automatic recoguition of toughness and tenderness of tongue manifcstation based on color features and texture features is scldom reported. Two fusion methods of tongue manffestation in the decision-making layer, which arc optimal linear combination method and AdaBoost algorithm are employed and compared. The experiment results showed that AdaBoost algorithm based on k-nearest neighbor classifier presents the highest recognition rate.