Lesion area assessment in psoriasis patients

Abstract

Psoriasis is an autoimmune skin disease, afflicting skin with red plaques that are usually accompanied by silvery-white scales. Various medical treatments are used, with different impacts on the patients, but there is no definite cure for the disease. The PASI standard is employed to measure the performance of the treatments. It includes four parameters, namely area, erythema, scaliness and skin thickness. The PASI parameters are usually measured manually by physicians through subjective clinical observations which are imprecise, time consuming and in some cases lead to diverse results. This paper presents a computer-based automatic method to measure the area parameter in the PASI standard. In the proposed method, the YCbCr colour space is used to differentiate the plaques from the skin by applying an optimal threshold method. Performance evaluation results indicate that the proposed method is able to determine lesion areas with accuracy higher than 96% for 18 out of 20 cases and higher than 92% for another case. As well as high accuracy the proposed method has another advantage over previous methods: it can automatically detect plaques with silvery-white scales, plaques on hairy skins and tiny plaques, as well as simple (scale-less) plaques.