Digital Image Segmentation for Cardiac Medical Images
Sagar Bhalerao, Kumar Kapoor, Rohit Tatar, Mandar Gade · 2014
Abstract—The digital image processing has proved beneficial in different areas. It is used as an appliance for extraction of a desired part of an image. At beginning, segmentation is performed on interested part of an image whose result can be used as a reference for different study. Different segmentation methods are present to perform this task, but the project aims at finding such a method which can be abundant for better segmentation. The proposed method revolves around selection of threshold value using techniques like histogram quantization, analysing of histogram slope percentage and determining of maximum entropy value. The result are luminous when a noise reduced image is used for segmentation. Keywords—Entropy, cardiac Images, segmentation, threshold, histogram. However, this method is ineffective when the image is distorted with elements like noise, white-spots, etc. Thus, in this context, if we previously perform noise reduction technique on the image and then provide it as an input, then a better result is obtained. To calculate multilevel threshold values, the techniques used are grouping of gray-level values, plotting of the histogram and calculation of maximum entropy. These technique clearly identifies insignificant thresholds and enables us to have control over the feature extraction process. II. ARCHITECTURE I.