Friday, May 2025
11:20 AM - 11:40 AM
Room: LL20D
Session: Under Display Camera Systems and Algorithms
Image Restoration for Under-Display Cameras: A Review of Current Technologies
Invited
Description:
Under-Display Cameras (UDCs) are a promising technology that integrates cameras beneath the display, enabling a larger screen-to-body ratio. However, positioning the camera beneath the display presents distinct challenges, as the light must pass through multiple display layers, leading to image and video degradation. These degradations include reduced transmittance, color distortion, blurring, noise, and flare, which significantly impact the quality of captured images or videos. These degradations significantly impact applications like photography, video recording, and facial recognition, which require high image accuracy and clarity. In this paper, we provide a comprehensive survey of recent technologies to mitigate various UDC degradations. We categorize the research into two key perspectives: datasets and deep-learning models. We explore studies focusing on creating synthetic and real-world UDC datasets that capture UDC degradation characteristics, facilitating research on UDC restoration methods. Additionally, we examine advancements in deep-learning techniques designed to restore UDC-distorted image and video quality, highlighting future opportunities for improving UDC performance.