The research on fingerprint recognition algorithm fused with deep learning
Nan Shuping, Wang Feng · 2020
In this paper, through deep learning, preprocessing and neural network training, the similarity of the fingerprint orientation field is finally given, which has reached the final fingerprint matching. First, calculate the direction field information of the fingerprint through a mathematical model; secondly, preprocess the collected fingerprint data, and remove the fingerprint data with the center point at the edge of the fingerprint image and the fingerprint data without the center point to find suitable data for training Set, the input of the next layer in the neural network is the splicing of the input information and output information of each layer in the neural network to achieve a tight connection between each neural network layer; through continuous training, a clear fingerprint recognition effect is finally achieved. Experiments show that the method proposed in this paper has a higher recognition rate.