The Study on Multi-exposure Image Fusion of Finger Vein

For the finger vein recognition technology, the image of the finger vein is seriously missing, and the brightness contrast is not very obvious. A multi-exposure image fusion algorithm based on the camera response curve is proposed. Gaussian pyramid decomposition is performed on multiple finger vein images of different exposures in the same scene, the amount of information of the pixels is measured according to the camera response curve, and the Gaussian equation constrains the brightness of the fused image. Then a weight correction function is proposed to avoid the loss of image detail in overexposed or underexposed regions, and finally multi-scale and multi-resolution fusion of the image through Laplacian pyramid. Finally, according to the algorithm proposed in this paper, several sets of actual finger vein images were collected and analyzed. The effects of the algorithm were analyzed and evaluated from subjective and objective aspects. Experiments show that the proposed algorithm is simple and feasible, which not only preserves the image details of overexposed and under-exposed areas in the source image set, but also minimizes distortion.