For Medical Image Segmentation Medical Image Roi Extraction

Medical image is usually constituted by ROI(Region Of Interest) and the background. Compared with the background, ROI contains important diagnosis information, whose area may be small. But it has great significance in the doctor's diagnosis, clinical treatment and pathologic analysis. If ROI is described with mistakes, the cost will be great. Thus, the extraction of ROI of medical image becomes a pressing demand. This paper focuses on the medical image segmentation algorithm based on fuzzy connectedness, and .analyzes the principles in detail. The research work and innovations of this paper include the following areas:1. This paper reviews the medical image segmentation techniques, and introduced the theoretical basis, flow, advantages and disadvantages, relationships about traditional segmentation methods and segmentation methods based on some special theories, such as mathematical morphology, neural networks, fractal theory, wavelets, statistics, fuzzy theory and so on.2. This paper studies the flows of segmentation algorithms based on the fuzzy conncectedness and the relative fuzzy conncectedness, and compares one with the other.3. The fuzzy relations are calculated for each pair of pixels in all paths beginning seed and ending target in the traditional segmentation algorithm based on fuzzy connectedness. This process is very tedious. To solve this problem, the fuzzy connectedness algorithm is analyzed and improved. An improved fast segmentation algorithm based on the relative fuzzy connectedness is proposed. Compared with the traditional algorithm, this algorithm increases the computing efficiency without affecting the segmentation precision.