Abstract

This paper introduces a novel dataset, called LCDMoire, which was created forthe first-ever image demoireing challenge that was part of the Advances inImage Manipulation (AIM) workshop, held in conjunction with ICCV 2019. Thedataset comprises 10,200 synthetically generated image pairs (consisting of animage degraded by moire and a clean ground truth image). In addition todescribing the dataset and its creation, this paper also reviews the challengetracks, competition, and results, the latter summarizing the currentstate-of-the-art on this dataset.

 

Quick Read (beta)