Entropy Estimation for Segmentation of Multi-Spectral Chromosome Images

In the early 1990s, the state-of-the-art in commercial chromosome image acquisition was grayscale. Automated chromosome classification was based on the grayscale image and boundary information obtained during segmentation. Multi-spectral image acquisition was developed in 1990 and commercialized in the mid1990s. One acquisition method, multiplex fluorescence in-situ hybridization (M-FISH), uses five color dyes. We previously introduced a segmentation algorithm for M-FISH images that minimizes the entropy of classified pixels within possible chromosomes. In this paper, we extend this entropy-minimization algorithm to work on raw image data, which removes the requirement for pixel classification. This method works by estimating entropy from raw image data rather than calculating entropy from classified pixels. A successful example image is given to illustrate the algorithm. Finally, it is determined that entropy estimation for minimum entropy segmentation adds a heavy computational burden without contributing any significant increase in classification performance, and thus not worth the effort.