Super-Resolution Reconstruction of Finger Vein Image Based on Regression Tree Model

Because the speed of clear fingerprint vein acquisition is slow, and the influence of some acquisition equipment makes the acquired vein graphics are relatively poor, this paper is based on random forest and DCT algorithm for super-resolution reconstruction of blurred images. The random forest algorithm avoids overfitting the data and solves the classification problem. DCT transform compresses the image space to further speed up the training speed of the image. The experimental result shows that the method is easy to converge during the process of training and has better image reconstruction quality.