A Small-Area Fingerprint Recognition Algorithm Based on Inverted Residual Connection

Small-area fingerprints image contain less detailed point information. The traditional fingerprint recognition method is based on the detail points that cannot be applied to small-area fingerprints appropriately. To solve this problem, a small-area fingerprint recognition algorithm is proposed in this paper. The algorithm is built on a small-area fingerprint recognition network with inverse residual structure reducing the parameters and computation of the deep neural network. The network converts a fingerprint image to a 256-dimensional vector and judges the similarity of fingerprints by the Euclidean distance between the vectors. The experimental results show that the method achieves a correct rate of 96.4% on the small-area fingerprint dataset established on FVC2002 and self-collected fingerprints. And the model parameter is 4.34M satisfying the usage needs of mobile terminals.