Learned Smartphone ISP on Mobile GPUs with Deep Learning, Mobile AI & AIM 2022 Challenge: Report

The role of mobile cameras increased dramatically over the past few years, leading to more and more research in automatic image quality enhancement and RAW photo processing. In this Mobile AI challenge, the target was to develop an efficient end-to-end AI-based image signal processing (ISP) pipeline...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2022-11
Main Authors: Ignatov, Andrey, Timofte, Radu, Liu, Shuai, Feng, Chaoyu, Bai, Furui, Wang, Xiaotao, Lei, Lei, Ziyao Yi, Yan, Xiang, Liu, Zibin, Li, Shaoqing, Shi, Keming, Kong, Dehui, Xu, Ke, Kwon, Minsu, Wu, Yaqi, Zheng, Jiesi, Fan, Zhihao, Wu, Xun, Zhang, Feng, Albert No, Cho, Minhyeok, Chen, Zewen, Zhang, Xiaze, Li, Ran, Wang, Juan, Wang, Zhiming, Conde, Marcos V, Ui-Jin, Choi, Perevozchikov, Georgy, Ershov, Egor, Zheng, Hui, Dong, Mengchuan, Lou, Xin, Zhou, Wei, Pang, Cong, Qin, Haina, Cai, Mingxuan
Format: Article
Language:English
Subjects:
Cameras
Deep learning
Image enhancement
Image quality
Signal processing
Smartphones
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The role of mobile cameras increased dramatically over the past few years, leading to more and more research in automatic image quality enhancement and RAW photo processing. In this Mobile AI challenge, the target was to develop an efficient end-to-end AI-based image signal processing (ISP) pipeline replacing the standard mobile ISPs that can run on modern smartphone GPUs using TensorFlow Lite. The participants were provided with a large-scale Fujifilm UltraISP dataset consisting of thousands of paired photos captured with a normal mobile camera sensor and a professional 102MP medium-format FujiFilm GFX100 camera. The runtime of the resulting models was evaluated on the Snapdragon's 8 Gen 1 GPU that provides excellent acceleration results for the majority of common deep learning ops. The proposed solutions are compatible with all recent mobile GPUs, being able to process Full HD photos in less than 20-50 milliseconds while achieving high fidelity results. A detailed description of all models developed in this challenge is provided in this paper.
ISSN:2331-8422