Flow-Guided Hair Removal for Automated Skin Lesion Identification
January 2018
in “
Multimedia Tools and Applications
”
TLDR The new method removes hair from skin images quickly and accurately to help identify skin lesions better.
In 2018, Kang, Kim, and Park developed a novel method for the automated removal of hair from skin lesion images to enhance the accuracy of lesion identification. Their technique involved an edge-tangent flow-based edge-detection method to identify hair-like structures and a texture synthesis approach for restoring hair-occluded areas with minimal artifacts. The method was designed to maintain the integrity of lesion information and performed well in terms of detection accuracy and restoration quality. Implemented on a PC, hair detection took 2-3 seconds and restoration 1-2 minutes per image. The study aimed to contribute to early skin cancer diagnosis, particularly benefiting patients in remote areas by improving automated skin care services. The number of images or participants involved in the study was not specified.
