# Deep Atrous Guided Filter for Image Restoration in Under Display Cameras

Source: [Deep Atrous Guided Filter for Image Restoration in Under Display Cameras](<https://www.sota2.com/research/paper/68fa684b1e209131cea1b2e3>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T22:33:00.328Z

Authors: Varun Sundar, Sumanth Hegde, Divya Kothandaraman, Kaushik Mitra

Published: 2020-08-14T07:54:52.000Z

## Abstract

Under Display Cameras present a promising opportunity for phone manufacturers to achieve bezel-free displays by positioning the camera behind semi-transparent OLED screens. Unfortunately, such imaging systems suffer from severe image degradation due to light attenuation and diffraction effects. In this work, we present Deep Atrous Guided Filter (DAGF), a two-stage, end-to-end approach for image restoration in UDC systems. A Low-Resolution Network first restores image quality at low-resolution, which is subsequently used by the Guided Filter Network as a filtering input to produce a high-resolution output. Besides the initial downsampling, our low-resolution network uses multiple, parallel atrous convolutions to preserve spatial resolution and emulates multi-scale processing. Our approach's ability to directly train on megapixel images results in significant performance improvement. We additionally propose a simple simulation scheme to pre-train our model and boost performance. Our overall framework ranks 2nd and 5th in the RLQ-TOD'20 UDC Challenge for POLED and TOLED displays, respectively.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Under-Display Camera Image Restoration on POLED 49 (val)](<https://www.sota2.com/research/sota/under-display-camera-image-restoration-on-poled-49-val.md>): Under-display camera image restoration; POLED 49 (val); PSNR: 33.79.

- [UDC Restoration on POLED (test)](<https://www.sota2.com/research/sota/udc-restoration-on-poled-test.md>): UDC Restoration; POLED (test); PSNR: 32.29.

- [Under-Display Camera Image Restoration on TOLED 49 (val)](<https://www.sota2.com/research/sota/under-display-camera-image-restoration-on-toled-49-val.md>): Under-display camera image restoration; TOLED 49 (val); PSNR: 37.87.

- [UDC Restoration on TOLED (test)](<https://www.sota2.com/research/sota/udc-restoration-on-toled-test.md>): UDC Restoration; TOLED (test); PSNR: 36.49.

- [Hardware-induced Blur Removal on T-OLED](<https://www.sota2.com/research/sota/hardware-induced-blur-removal-on-t-oled.md>): Hardware-induced Blur Removal; T-OLED; PSNR: 36.49.

- [UDC Image Restoration on Synthetic UDC dataset HDRI Haven (test)](<https://www.sota2.com/research/sota/udc-image-restoration-on-synthetic-udc-dataset-hdri-haven-test.md>): UDC Image Restoration; Synthetic UDC dataset HDRI Haven (test); PSNR: 27.12.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2008.06229>)

[PDF](<https://arxiv.org/pdf/2008.06229.pdf>)
