基于模糊CMAC神经网络的医学舌苔图像检索  

The Retrieval of the Medical Tongue-coating Images Using Self-organized Map Neural Network

作　　者：秦鉴[1] 吴正治[2] 张灵[3] 孟君[1] 刘红健[1] 金明华[1] 吴国珍[1] 

机构地区：[1]中山大学附属第一医院,广东广州510080 [2]南方医科大学附属深圳医院,广东深圳518035 [3]广东工业大学计算机学院,广东广州510090

出　　处：《深圳中西医结合杂志》2007年第6期329-332,F0003,共5页Shenzhen Journal of Integrated Traditional Chinese and Western Medicine

基　　金：国家自然科学基金(30371717);国家中医药管理局基金(02-03JP37);广东省自然科学基金(031688)资助。

摘　　要：笔者提出了一种基于模糊CMAC神经网络的医学舌苔图像检索方法。该法利用CMAC网络,实现彩色图像直方图的模糊相关求和,并在此过程中自动调整加权算法中特征色的加权因子,使舌苔图像的检索能针对几大类标准进行,提高了检索效率。实验结果表明,与过去根据经验设定参数的方法相比,本法具有高效、准确,对阴影、反光现象不敏感等特点。A retrieval way for medical tongue-coating images based on fuzzy CMAC neural network is presented in this paper. With the weight strength of the CMAC network, the number and the regarded feature-colors are adjusted self-adaptively, during the process of calculating the confused relative factors＇ summary of the color-images, so that the retrieval for tongue-coating images concentrates efficiently on several distinct kinds. The experimental results show, comparing with the manual way, the new way is more accurate and effective, and less relative with the illumination of the images.

关 键 词：模糊CMAC网络 舌苔 特征色 模糊相关 

分 类 号：R241[医药卫生—中医诊断学]