Deep learning-based finger vein recognition and security: A review

The recognition system implies development that passes through the various stages. The finger vein recognition (FVR) is the lead over the other biological modalities like finger print, face iris, etc., this paper reviews the worked done in the area of FVR. The pre-processing is needed to enhance the images for better results. The feature extraction module provides the collection of the best features in the finger vein images, which is used for template generation. The template-based schemes are in fact the best and most appropriate for security purpose because it only preserved the scrambled information rather than the original features of the human beings. According to this review the convolutional neural network (CNN)-based models are best for FVR but still there have been some challenges faced by it so those would be improved in the experimental work of this review.