This repository is a survey / code collection centered on deep learning–based image fusion (e.g. fusing infrared + visible light images, multi-modal fusion) methods. It catalogs many fusion algorithms (e.g. DenseFuse, FusionGAN, NestFuse, etc.), links to code implementations, and describes evaluation metrics. The repository includes a “General Evaluation Metric” subfolder containing objective fusion metrics. It is not a single monolithic tool, but rather a curated reference and aggregation of methods, code and performance comparisons in the domain of image fusion. Survey style description of method taxonomy, architectures, loss types. Compilation of many state-of-the-art image fusion methods (infrared + visible, multi-focus, multi-exposure). Survey style description of method taxonomy, architectures, loss types.
Features
- Compilation of many state-of-the-art image fusion methods (infrared + visible, multi-focus, multi-exposure)
- Links to implementation code for each method
- General objective evaluation metrics (e.g. for fusion quality)
- Survey style description of method taxonomy, architectures, loss types
- Up-to-date list of new fusion methods and variants
- Language predominantly MATLAB (as code base)