Real-Time Under-Display Cameras Image Restoration and HDR on Mobile Devices

The new trend of full-screen devices implies positioning the camera behind screen to bring a larger display-to-body ratio, enhance eye contact, and provide notch-free viewing experience on smartphones, TV or tablets. On other hand, images captured by under-display cameras (UDCs) are degraded in front them. Deep learning methods for image restoration can significantly reduce degradation images, providing satisfying results human eyes. However, most proposed solutions unreliable efficient enough be used real-time mobile devices. In this paper, we aim solve problem using deep capable processing FHD commercial smartphones while high-quality results. We propose lightweight model blind UDC Image Restoration HDR, also benchmark comparing performance runtime different smartphones. Our models competitive benchmarks $$\times 4$$ less operations than others. To best our knowledge, first work approach analyze real-world single from efficiency production point view.