The field of image processing is moving towards more advanced and controllable methods for image restoration and fusion. Current research focuses on developing techniques that can handle complex degradations and accommodate user-specific requirements. Noteworthy papers in this area include those that propose novel frameworks for image fusion, such as methods that utilize degradation and semantic dual-prior guidance or controllable image fusion with language-vision prompts. Other significant contributions include region-customized image restoration with human instructions and methods for multispectral demosaicing via dual cameras. Notable papers:

• indiSplit brings severity cognizance to image decomposition in fluorescence microscopy, enabling more accurate unmixing of cellular structures.
• ControlFusion introduces a controllable image fusion framework with language-vision prompts, allowing for adaptive neutralization of composite degradations.
• DSPFusion presents a degradation and semantic dual-prior guided framework for degraded image fusion, utilizing degradation priors and high-quality scene semantic priors to guide information recovery and fusion.