A Threshold Segmentation Method of Chromosome Microscope Image
The study of chromosome image segmentation and recognition methods involves the knowledge in multiple disciplines such as computer image processing and image pattern recognition. In general, the features of the object region of chromosome image are bad in similarity and uniformity and it is easy to mix the fake boundary formed by the surrounding parts with the boundary of chromosome; so, it is rather difficult to accurately segment and calculate chromosome and the traditional methods are generally not high in accurately segmenting chromosome image. This paper mainly with the microscope image of chromosome as the object of study, performs feature extraction, classification and identification of chromosome image and extracts the shape features of cells. The core of algorithm is to segment the image into several classes by seeking different feature thresholds. The feature threshold here generally refers to the gray-scale features of pixels and in essence, it is to divide the pixels into different gray-scale collections according to gray levels. Adopt the erosion operation of binary image in morphological image processing, refine the chromosome and label the connected chromosome image component for statistics of numbers of chromosome. The testing experiment proves that the method of this paper is effective.