Image Feature Extraction and Recognition of Chinese Herbal Medicine Based on Pulse Coupled Neural Networks

Abstract

In order to effectively extract the characteristic information of microscopic image feature to Chinese herbal medicines (CHM), and improve the recognition accuracy automatically, a novel algorithm using Pulse Coupled Neural Networks (PCNN) is put forward. Firstly, the PCNN model is introduced from suitable for processing image of biological tissue. Secondly, the characteristic of time series with PCNN image processing is formed, and transformed into the feature of one dimensional entropy series, which can behalf the image inherent characteristics. Finally, the automatic identification is taken to the extracted image entropy sequence feature. The experimental results show that the entropy sequence feature has the ability of anti-geometric distortions, the novel method have characteristics of simple extraction approach, little extraction parameter, easy implementation, higher accurate recognition ratio and strong robustness.