Improving Accuracy of Face Detection in ID Proofs using CNN and Comparing with DLNN

The goal of the proposed work is to upgradethe accuracy rate in face observation with Convolutional Neural Networks algorithm and comparing with Deep Learning Neural Networks. A count aggregate of 40 trails of different angles of id cards were collected along with their passport size photos for detecting the faces. These samples are divided into two groups each having 20 samples and the accuracy values were calculated to face detection in CNN. The G power for this is taken as 0.8. Convolutional Neural Networks achieved the accuracy of 93% and the same for Deep Learning Neural Networks is 96.80% in the recognition of face in ID cards. From this task it is observed that the DLNN algorithm executed significantly better than CNN algorithm in face detection in ID proofs on the basis of accuracy.