CNN Classifier based Low-resolution Face Recognition Algorithm
Shyam Singh Rajput, K. V. Arya · 2020
Faces captured under an uncontrolled environment usually have low-resolution (LR) in addition to poor illumination conditions and uncontrolled poses. These types of images significantly deteriorate the performance of traditional face recognition techniques. Hence, in this paper, a new face recognition algorithm is developed to minimize the problem of matching the LR images with high-resolution gallery images. In the proposed algorithm, the convolutional neural network (CNN) based classification mechanism is used to improve recognition accuracy. Investigational results performed on two standard databases of face images exposes the preeminence of the proposed algorithm over the other comparative existing face recognition models.