Finger-Vein Image Inpainting Based on an Encoder-Decoder Generative Network

Finger-vein patterns are usually used for biometric recognition. There may be spots or stains on fingers when capturing finger-vein images. Therefore, the obtained finger-vein images may have irregular incompleteness. In addition, due to light attenuation in biological tissue, the collected finger-vein images are often seriously degraded. It is essential to establish an image inpainting and enhancement model for the finger-vein recognition scheme. In this paper, we proposed a novel image restoration mechanism for finger-vein image including three main steps. First, the finger-vein images are enhanced by the combination of Gabor filter and Weber's low. Second, an encoder-decoder generative network is employed to make image inpainting. Finally, different loss functions are taken into consideration for the model optimization. In the simulation part, we carry out some comparative experiments, which demonstrates the effectiveness and practicality of the proposed finger-vein image restoration mechanism.