Doing More With Moire Pattern Detection in Digital Photos

Detecting moire patterns in digital photographs is meaningful as it provides priors towards image quality evaluation and demoireing tasks. In this paper, we present a simple yet efficient framework to extract moire edge maps from images with moire patterns. The framework includes a strategy for training triplet (natural image, moire layer, and their synthetic mixture) generation, and a Moire Pattern Detection Neural Network (MoireDet) for moire edge map estimation. This strategy ensures consistent pixel-level alignments during training, accommodating characteristics of a diverse set of camera-captured screen images and real-world moire patterns from natural images. The design of three encoders in MoireDet exploits both high-level contextual and low-level structural features of various moire patterns. Through comprehensive experiments, we demonstrate the advantages of MoireDet: better identification precision of moire images on two datasets, and a marked improvement over state-of-the-art demoireing methods.