Chronic wounds multimodal image database.

Area of Science:

• Biomedical imaging
• Medical informatics
• Wound care technology

Background:

• Effective wound healing monitoring requires advanced computational tools.
• Existing datasets lack comprehensive, coregistered multimodal data for algorithm development.

Purpose of the Study:

• To introduce a novel multimodal wound image database for accelerating computer-aided wound healing monitoring.
• To provide a publicly accessible resource for researchers developing and evaluating image-based algorithms.

Main Methods:

• A specialized system capturing parallel optical axes for photo, thermal, stereo, and depth map imaging.
• Acquisition of 188 image sets from 79 patient visits, including photographs, thermal images, and 3D meshes.
• Expert-delineated wound outlines and coregistration of all images with associated registration parameters.

Main Results:

• A comprehensive multimodal database of chronic wound images, including photographs, thermal data, and 3D surface models.
• Inclusion of coregistered images and manual delineations, enhancing data utility for algorithm training and validation.
• Public release of the database at https://chronicwounddatabase.eu for research community access.

Conclusions:

• The developed database is the first publicly available resource offering coregistered multimodal data (photographs, thermal maps, 3D models) for wound healing analysis.
• This readily accessible dataset will significantly expedite the development and benchmarking of new image-based algorithms for wound monitoring.
• Facilitates faster innovation in computer-aided diagnosis and treatment assessment for chronic wounds.