The first step in most computer aided diagnosis systems is an accurate segmentation of breast region, which affects not only the accuracy but also the speed of the analysis because it significantly reduces the area of the image to be examined. The second step usually includes removal of pectoral muscle region, which is seen in mediolateral oblique view mammograms. This is primarily done to reduce the number of false positive breast cancer detections. In this paper, a method for the segmentation of breast region based on contrast enhancement and k-means algorithm is proposed. To extract pectoral muscle, a region of interest is found, its contrast is enhanced and the pectoral muscle is identified using k-means algorithm. Cubic polynomial fitting is used for the estimation of muscle's boundary. The method is validated with mammograms from miniMIAS database.

The method for nuclei segmentation in fluorescence in-situ hybridization (FISH) images, based on the inverse multifractal analysis (IMFA) is proposed. From the blue channel of the FISH image in RGB format, the matrix of Holder exponents, with one-by-one correspondence with the image pixels, is determined first. The following semi-automatic procedure is proposed: initial nuclei segmentation is performed automatically from the matrix of Holder exponents by applying predefined hard thresholding; then the user evaluates the result and is able to refine the segmentation by changing the threshold, if necessary. After successful nuclei segmentation, the HER2 (human epidermal growth factor receptor 2) scoring can be determined in usual way: by counting red and green dots within segmented nuclei, and finding their ratio. The IMFA segmentation method is tested over 100 clinical cases, evaluated by skilled pathologist. Testing results show that the new method has advantages compared to already reported methods.

No abstract

No abstract

No abstract