Psoriasis is a chronic, immune-mediated, relapsing, inflammatory skin disease with variable morphology, distribution, severity and course. It creates raised, reddish patches covered with silvery whitish scales termed as Psoriatic Plaque. The prevalence of this disease varies from 1%-12% among different populations worldwide. It is an eye estimation procedure and suffers from inter and intraobserver variability. To get rid of that, the development of automatic image analysis will be valuable work. The result of those systems requires the development of an accurate and robust psoriasis plaque segmentation procedure.

This project concerns automatic psoriatic plaque segmentation in skin images. Feature Pyramid Network (FPN) model based on "se_resnext50_32x4d" encoder is used. The model outperforms with 94% Intersection over Union (IoU), 97% FScore, and 97% Accuracy.