Image Processing Technique for Gender Determination from Medical Microscope Image

Abstract

© 2020, Springer Nature Switzerland AG. This article proposes an image processing technique to calculate the histological variables from microscope image of occipital bones for gender determination. Occipital bones are highlighted based on 13 parameters to find out the relationship with the gender of samplings. The samples are 80 images classified into 46 males and 34 females with ages between 25 and 90 years old. Direct and stepwise discriminant functions are two methods for accuracy evaluation. The experiments show demonstrative results of gender determination accuracy based on both direct and stepwise discriminant functions. While the direct discrimination on 5 parameters represents 97.5% for both gender classification and gender prediction, the stepwise discrimination on 3 parameters is also at 97.5% for both gender classification and gender prediction.