DATA-DRIVEN HOMOLOG MATCHING FOR CHROMOSOME IDENTIFICATION

Karyotyping involves the visualization and classification of chromosom es into standard classes. In ''normal'' human metaphase spreads, chrom osomes occur in homologous pairs for the autosomal classes 1-22, and X chromosome for females. Many existing approaches for performing autom ated human chromosome image analysis presuppose cell normalcy, contain ing 46 chromosomes within a metaphase spread with two chromosomes per class. This is an acceptable assumption for routine automated chromoso me image analysis, However, many genetic abnormalities are directly li nked to structural or numerical aberrations of chromosomes within the metaphase spread. Thus, two chromosomes per class cannot be assumed fo r anomaly analysis. This paper presents the development of image analy sis techniques which are extendible to detecting numerical aberrations evolving from structural abnormalities, Specifically, an approach to identifying ''normal'' chromosomes from selected class(es) within a me taphase spread is presented. Chromosome assignment to a specific class is initially based on neural networks, followed by banding pattern an d centromeric index criteria checking, and concluding with homologue m atching, Experimental results are presented comparing neural networks as the sole classifier to our homologue matcher for identifying class 17 within normal and abnormal metaphase spreads.