NTIRE 2022 - Image Inpainting Challenge
The 7th edition of NTIRE: New Trends in Image Restoration and Enhancement workshop will be held on June 2022 in conjunction with CVPR 2022.
Image manipulation is a key computer vision task, aiming at the restoration of degraded image content, the filling in of missing information, or the needed transformation and/or manipulation to achieve the desired target (with respect to perceptual quality, contents, or performance of apps working on such images). Recent years have witnessed an increased interest from the vision and graphics communities in these fundamental topics of research. Not only has there been a constantly growing flow of related papers, but also substantial progress has been achieved.
Recently, there has been a substantial increase in the number of published papers that directly or indirectly address Image Inpainting. Due to a lack of a standardized framework, it is difficult for a new method to perform a comprehensive and fair comparison with respect to existing solutions. This workshop aims to provide an overview of the new trends and advances in those areas. Moreover, it will offer an opportunity for academic and industrial attendees to interact and explore collaborations.
Jointly with the NTIRE workshop, we have an NTIRE challenge on Image Inpainting, that is, the task of predicting the values of missing pixels in an image so that the completed result looks realistic and coherent. This challenge has 3 main objectives:
This challenge has 2 tracks:
The aim is to obtain a mask agnostic network design/solution capable of producing high-quality results with the best perceptual quality with respect to the ground truth.
In addition to the typical strokes, with this challenge, we aim at more generalizable solutions.
Thick Strokes | Medium Strokes | Thin Strokes |
---|---|---|
Every_N_Lines | Completion | Expand |
---|---|---|
Nearest_Neighbor |
---|
Following a common practice in Image Inpainting methods, we use three popular datasets for our challenge: FFHQ, Places, and ImageNet. Additionally, to explore a new benchmark, we also use the WikiArt dataset to tackle inpainting towards art creation. See the data for more info about downloading the datasets.
The top-ranked participants will be awarded and invited to follow the CVPR submission guide for workshops to describe their solutions and to submit to the associated NTIRE workshop at CVPR 2022.
See Evaluation.
In case of any questions about the challenge or the toolkit, feel free to open an issue on Github.
The terms and conditions for participating in the challenge are provided here
Thanks to everyone who makes their code and models available. In particular,
@inproceedings{romero2022ntire,
title={NTIRE 2022 image inpainting challenge: Report},
author={Romero, Andres and Castillo, Angela and Abril-Nova, Jose and Timofte, Radu and Das, Ritwik and Hira, Sanchit and Pan, Zhihong and Zhang, Min and Li, Baopu and He, Dongliang and others},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={1150--1182},
year={2022}
}