41 datasets found

Simulated data from five standard data generating processes (DGPs) to evaluate when SGTs and Booging outperform their counterparts relying on their traditionally associated...

Image quality is the basis of image communication and understanding tasks. Due to the blur and noise effects caused by imaging, transmission and other processes, the image...

Perceptual image restoration seeks for high-fidelity images that most likely degrade to given images.

The dataset used for testing the proposed network.

The proposed network is used for image restoration with unknown degradation factors.

A large-scale, high-quality dataset for image restoration, containing 50,000 high-quality images with rich texture details and semantic diversity.

The Flare7K++ dataset is an extension of the Flare7K dataset, consisting of 5,000 images of scattering flare and 2,000 images of reflective flare.

The Flare7K dataset is a synthetic dataset for lens flare removal, consisting of 5,000 images of scattering flare and 2,000 images of reflective flare.

The SnowKITTI dataset is a real-world dataset for single image desnowing.

The SnowCityScapes dataset is a real-world dataset for single image desnowing.

The Snow100K dataset is a real-world dataset for single image desnowing.

The SRRS dataset is a synthetic dataset for single image desnowing.

The CSD dataset is a synthetic dataset for single image desnowing.

The dataset used in this paper is Poisson inverse problems involving the Poisson data-fidelity term f introduced in (2).

The CU-Mamba model is used for image restoration, and the authors tested it on the SIDD and DND datasets.

The Open Turbulent Image Set (OTIS) dataset contains images of turbulent scenes.

The TMT dataset consists of static and dynamic parts. The static part is synthesized using the place dataset [58] for static scenes. The dynamic part is generated using the...

Real-world image restoration and enhancement

The dataset used in the Generative Diffusion Prior for unified image restoration and enhancement.

The dataset used in the paper is a synthetic dataset for blind super-resolution, consisting of low-resolution images and their corresponding blur kernels.