Methods Of Determining The High Intensity Focused Ultrasound Treatment Area Based On Independent Component Analysis

High intensity focused ultrasound(HIFU) technology is a new type of tumor treatment technology. It is widely used because the technology has non-invasive and no sideeffects while treating tumor. However, determining the treatment area is an essential image processing technique in this technology, it is an important basis for checking lesions and drawing the outline of the tumor. This article is mainly about determining the HIFU treatment area, putting forward the method of determining the HIFU treatment area based on combination of improved independent component analysis(ICA) and mathematical morphology, and the method of determining the HIFU treatment area based on wavelet and improved ICA.Firstly, the paper introduces the method of determining the HIFU treatment area based on combination of improved ICA and mathematical morphology. The method begins with the improvement and optimization of the algorithm. It regards the independent component of image as the combination of edge’s independent component and background’s independent component, it translates the extraction of image’s component into the extraction of edge’s component by using the improved fast ICA, then it gets the independent component of the image’s edge, then the method realizes the image segmentation and determines the treatment area by enhancing the edge of the image through the mathematical morphology transformation. Compared with the traditional methods, this method has better segmentation performance, the image contour and treatment area can be clearly observed, the segmented image has good connectivity and retains the details of original image. The results show that the method has better effect when the method is used for determining the HIFU treatment area.Secondly, the method of determining the HIFU treatment area based on wavelet and improved ICA is used for denoising the medical ultrasound image with noise. The method preprocesses the image through the improved ICA to make the edge more clearly, then by using wavelet decomposition to determine the decomposition level with the best de-noising effect. In the basis of the decomposition level the method finishes de-noising and segmentation by using improved semi-soft threshold algorithm, and then gets the edge of the treatment area. By comparing the objective parameters we can know that the method can reduce the noise effectively, the sign-to-noise ratio is improved obviously, and moreover it has good segmentation result and draws the outline of the treatment area clearly. The results show that the method can denoise as well as detect the HIFU treatment area.