Collagen second harmonic generation image analysis for diabetes determination

A collagen second harmonic generation (SHG) and third harmonic generation (THG) image analysis algorithm based on feature extraction and classification for diabetes disease determination is developed in the paper. It is designed to find the early symptoms of diabetes by using the proposed method to calculate diabetes estimation indicator. The algorithm detects the basic diabetes featured elements on images of human dermis skin. The proposed algorithm consists of the processes of smoothing, adaptive three-region thresholding, binarization, segmentation, feature extraction, and diabetes disease determination. This feature classification criterion provides a new viewpoint to help diabetes diagnosis.