Ultra High-definition Demoiréing Dataset

When photographing the contents displayed on the digital screen, an inevitable frequency aliasing between the camera’s color filter array (CFA) and the screen’s LCD subpixel widely exists. The captured images are thus mixed with colorful stripes, named moiré patterns, which severely degrade the perceptual quality of images. Currently, efficiently removing moiré patterns from a single moiré image is still challenging and receives growing attention from the research community.

Ultra High-definition Demoiréing Dataset or UHDM is the first 4K resolution demoiréing dataset, consisting of 5,000 image pairs. This dataset includes diverse scenes, such as landscapes, sports, video clips, and documents. The moiré images are generated following practical routines, with different device combinations and viewpoints to produce diverse moiré patterns.

The dataset was collected by the authors of the paper Towards Efficient and Scale-Robust Ultra-High-Definition Image Demoiréing.

To obtain the real-world 4K image pairs, the authors of the paper Towards Efficient and Scale-Robust Ultra-High-Definition Image Demoiréing first collect high-quality images with resolutions ranging from 4K to 8K from the Internet. Given that Internet resources lack document scenes, which also constitute a vital application scenario (e.g., slides, papers), the authors manually generate high-quality text images and make sure they maintain 3000 dpi (Dots Per Inch). Finally, the collected moiré-free images cover a wide range of scenes, such as landscapes, sports, video clips, and documents. Given these high-quality images, the authors generate diverse real-world moiré patterns elaborated upon below.

UHDM

1. First, in order to produce realistic moiré images and ease the difficulties of calibrations, the authors shoot the clean pictures displayed on the screen with a camera phone fixed on a DJI OM 5 smartphone gimbal, which makes it possible to conveniently and flexibly adjust the camera view through its control button.
2. Second, the authors note that the characteristics of moiré patterns highly are highly dependent upon the geometric relationship between the screen and the camera. Therefore, during the capturing process, the authors continuously adjust the viewpoint every ten shots to produce diverse moiré patterns.
3. Third, the authors adopt multiple < mobile phone, screen > (i.e., three mobile phones and three digital screens, see supplement for more details) combinations to cover various device pairs, since they will also have an impact on the styles of moiré patterns.
4. Finally, in order to obtain aligned pairs, the authors utilize the RANSAC algorithm to estimate the homography matrix between the original high-quality image and the captured moiré screen image. Since it is difficult to ensure accurate pixel-wise calibration due to the camera’s internal nonlinear distortions and perturbations of moiré artifacts, manual selection is performed to rule out severely misaligned image pairs, thereby ensuring quality.

This dataset is sourced from Kaggle.