Abstract:In this work

we present an automated method for identifying chromosome in fluorescence stained medical cell images. The detection of the candidate cell images is based on dynamical similarity segment algorithm. For the elimination of false findings

which come from illumination

dye concentration and morphological variation

the identification method based on characteristic parameters was applied. Salient features characterizing the area

chromaticity

shape

compactness were extracted from the candidate chromosome regions and the identification procedure was accomplished with the application of fuzzy comprehensive evaluation. By setting up the membership functions in a fuzzy system

the chromosomes in cells can be identified effectively. The proposed method was evaluated on a set of cell images and compared with the methods of the classified identification and the single characteristic identification. The result showed that the proposed method produces more accurate result

and the accuracy is up to 95% with a real-time performance.