Person identification by using ResNet on hand images

The use of biometric features for surveillance and the identification of people is a very popular theme among researchers. There are various studies focused on this topic using facial features, gait features, audio features. This study aims to identify people by analyzing their hand images with deep learning model. Prior to the training step, all the images are preprocessed. They are smoothened while the edges are preserved by guided filter. The processed images are trained and tested in several models such as ResNet18, ResNet34, ResNet50, ResNet101 models. The obtained results are compared with the results of some other studies on the same dataset.