A Novel Strategy for Gender Identification from Hand Dorsal Images using Computer Vision
Navoneel Chakrabarty · 2019
Criminal Investigations and Archaeological Studies often involve Gender Determination from different body parts when dismembered human remains are encountered. The body parts may include hair or hand in criminal investigations and hand bone length or hand to index finger length ratio in archaeological research. Often in crimes or even in accidents, Forensic Science comes into play. Forensic Science, by definition, is the scientific application of criminal and civil laws for collection of scientific evidences for prosecution or defense or both. So, in Forensic Research, Gender Determination is a very integral part especially from hair or hand, which becomes difficult for anyone other than Forensic Scientists, to conclude. This study aims at classification of the gender of an individual into male and female from the image of his or her dorsal side of hand. A Deep Learning-Machine Learning Hybrid Model is developed, yielding a Validation Accuracy of 100% and Recall of 1.0. Such an Automated System is highly essential in making Forensic Research, easy and hence swift analysis of scientific evidences.