Automatic PASI area scoring

Abstract

Psoriasis is a skin disorder which is caused by genetic fault. There is no cure for psoriasis; however, there are many treatment modalities to help control the disease. To evaluate treatment efficacy, PASI (Psoriasis Area and Severity Index) which is the current gold standard method is used to measure psoriasis severity by evaluating the area, erythema, scaliness and thickness of the plaques. However, the calculation of PASI can be tedious and subjective. There have been many attempts in assessing psoriasis lesion area in order to calculate PASI area score. Segmentation on hue-chroma plane of CIE L*a*b* colour space is found to be effective method to extract psoriasis lesion from normal skin. However, centroids of normal skin and psoriasis lesion in hue-chroma plane must be calculated for each patient (local centroids). In this work, we investigate the possibility of using centroids obtained from three different groups based on their skin colour to avoid calculating centroids for each patient. First, patients were grouped according to their skin colour (fair, brown, and dark). For each group, the centroids of four body regions (head, trunk, arm, and leg) are calculated from normal skin and lesion samples of all patients in the group. The study involves patients from three different ethnic origins having different skin tones. Results of segmentation using global centroids are comparable with segmentation using local centroids.