Wound characterization is important task in chronic wounds treatment, because changes of the wound size and tissue types are indicators of the healing progress. Developed color image processing software analyze digital wound image and based on learned tissue samples performs tissue classification. Implemented statistical pattern recognition algorithm classifies individual pixels of the wound image based on color information. Classification parameters were learned from examples presented to the application during the learning process. Results of the analysis contain the wound image represented in pseudo colors as well as percentage of tissue types within the wound area. |
Sadržaj objavljen na bib.irb.hr (osim ako je drugačaije navedeno) podložan je pravilima Creative Commons licence
