Awesome ImageHarmonization

This project is a comprehensive, community-maintained collection of resources focused on image harmonization—the process of making composite images appear realistic by adjusting visual properties like color, lighting, and texture. It serves as a valuable reference for researchers, developers, and enthusiasts working in computer vision and graphics.

Key Features

• Structured Categorization — Resources are organized into sections: Non-DL (traditional methods), DL (deep learning approaches), Datasets, and Other Resources.
• Academic Focus — Includes key papers from top conferences (CVPR, ICCV, TOG) and journals, with publication years and links.
• Code Availability — Many entries link to open-source implementations, facilitating experimentation and reproducibility.
• Dataset Links — Provides direct access to datasets used for training and evaluating harmonization models.
• Community Contributions — Encourages users to suggest new items or report issues to keep the list up-to-date.

Philosophy

The project emphasizes accessibility and curation, aiming to centralize scattered research materials into a single, well-organized repository to accelerate progress in image harmonization.