×

The method of identifying biological tissue lesion based on Otsu clustering. (Chinese. English summary) Zbl 1463.92030

Summary: In the process of high intensity focused ultrasound (HIFU) treatment of tumors, the damage of biological tissue is an important manifestation of therapeutic effect. Therefore, it is important for researching on the damage threshold of tissues based on the method of B-ultrasound monitoring in efficacy evaluation. Firstly, statistical features were extracted from ultrasonic images before and after HIFU irradiation of the fresh pork tissue in vitro. The inter-distance was employed to evaluate the ability of several statistical features to recognize whether biological tissues had been denatured. Then, the feature with the biggest inter-distance was considered as the best statistical feature. Finally, the method of maximal inter-cluster variance was used to automatically determine the damage threshold of tissues. The results show that the inter-distance of skewness was the biggest feature. Comparing with the other features, the skewness had better ability to distinguish the damage tissues. And biological tissues had been denatured when the skewness value of HIFU treatment region was around 0.575. It was indicated that the method can identify effectively the HIFU treatment region, when the classified threshold was applied to the sample data.

MSC:

92C55 Biomedical imaging and signal processing
62P10 Applications of statistics to biology and medical sciences; meta analysis
92C50 Medical applications (general)
PDFBibTeX XMLCite