A Study Of The Techniques For Thrombocyte Image Analysis

Analysing Thrombocyte image with computer is important in both medical study and clinical diagnosis. It is difficult to recognize thrombocyte by computer due to the following facts: First, it has no shape character; Second, the black spots in the cells, the uneven luminance of the background, the touch of adjacent cells, etc. make the segmenting more difficult. The main part of this paper is to study the segmenting techniques for Thrombocyte image. Based on the theory of the image segmentation using combined edge and gray-level information, a segmenting procedure is derived in which a number of the traditional image processing techniques are used: We first execute the edge detection procedure and do a particular work with the edge image in order to eliminate the noise and black spots edge, then calculate the local threshold based on the edge information and segment the image, and finally extract the object region with the interaction of operator(s). The result of experimentation affirms that this method is effective to segmenting of Thrombocyte image.