Preprocessing Method Flow of Under-screen Fingerprint Image

摘要

With the development of the full-screen technology of smart phones, under-screen fingerprint recognition technology is used by more and more smart phones. The under-screen fingerprint sensor is placed under the screen glass and the TFT array, and receives reflected light through them. Therefore, the image collected by the under-screen fingerprint sensor has poor image quality, and is disturbed by serious grid-like noise. Usually, image preprocessing method flow is needed for improving the image quality before we extract the feature points of the fingerprint image. However, due to the different imaging principles, the existing fingerprint image preprocessing method flows are not suitable for the under-screen fingerprint image. In this paper, we analyze the specific degradation and noise type of the optical-based under-screen fingerprint images, and design a special preprocessing method flow for optical-based under-screen fingerprint image. Specifically, in our proposed image preprocessing method flow, we add notch filter and low-pass filter to remove periodic grid-like noise, and we use homomorphic filter to enhance image contrast. Experimental results show that the image quality of under-screen fingerprint images is significantly improved after using our proposed preprocessing method flow.