Computer Aided Method for the Detection of Structural Abnormalities in Acute Lymphocytic Leukemia from the G-Banded Karyotypes

The automated detection of structural abnormalities in the Acute Lymphocytic Leukemia (ALL) from the karyotypes is a major technical challenge in the cytogenetic imaging developments. Structural chromosomal abnormalities are the key evidences of genetic disorders such as ALL, but the automated identification of these abnormalities is still difficult. There is no successful automated method for the classification of structural abnormalities in karyotypes and the detection of ALL from the G-banded chromosomes. Currently, the identification of the ALL specific structural abnormalities including translocations from the G-banded karyotypes are done by the cytogenetic experts in the lab manually which is labor intensive and time consuming. Mimicking the actions performed by the cytogenetic expert for abnormality detection is a challenging task. The normal and abnormal chromosomes are classified using ANN (Artificial Neural Network) which is trained using the back propagation network. The proposed template matching method which combines the correlation based similarity measures SSD (Sum of Squared Differences) block matching and NCC(Normalized Cross-Correlation) suggests an efficient automated method for the detection of ALL specific structural abnormalities from the G-banded karyotype which helps in the prognosis and treatment evaluation.