Retrieval of Medical Ultrasound Image Based on Texture Feature
Yong Wang · Journal of Tianjin University Science and Technology · 2005
In medical informatics, the study of medical images′ physical features contributes to implement auto-analysis and content-based retrieval. Therefore, a kind of texture feature extracting algorithm named improved-Gabor filter is put forward.After the Gabor wavelet analysis of medical ultrasound images, the texture frequency and direction performance can be got and with them the texture feature vector is constituted and dimension of the vector is also reduced.Applying these features in content-based image retrieval system can avoid dimension crisis made by using Gabor wavelet.Experimental results of five kinds of echocardiography show that, the precision and recall of retrieval are high, and the retrieval time is short, the results of these experiments are satisfying.