Fingerprint analysis for age estimation using deep learning models (ResNet50 and VGG-16)

Abstract

Automatic detection of age from biometric traits is gaining attention from researchers due to its application in many fields. Even though substantial work has been carried out for finding the age using face, gait and speech, very few efforts have been made using fingerprint. This is because of the complexity of extracting distinguishable features. In our proposed work, a ResNet50 model is trained and tested by a fingerprint database namely, house database of digital fingerprint, which contains 1000 images of size 103×96 where 900 are training images, and 400 are testing images. By this model, we have estimated the age of a person under four groups such as 1-8, 9-15, 16-25, 25-60. Experimental results reveal the efficiency of the proposed approach.