there is a two task in this project:
Task 1: You need to build a classification model which classify fingerprint image belong to which person. As you know everyone has unique image so u need to implement one-shooting technigue or synthesize training data.
Task2: Task 1 the model would be constructed. So Task 2 PART: ON TEST TIME we deliberately corrup some part of the image. (Some part is missing). Using GAN, we will complete the missing part (fiill up) and we test first with corripted images and images which we would complete using Gan. We would expect that the accuracy of the classification model incease using GAN.