A mobile/desktop medical application for automatic differential diagnosis of psoriasis lesions

In the last few years, computer-based classification has been introduced as an additional tool to improve the clinical diagnosis of the erythemato-squamous diseases. The objectives of this study are: to demonstrate the importance of computer-based classification algorithms which have only clinical features as input in helping the physician to differentiate between psoriasis and non-psoriasis diseases and, to introduce these Machine Learning algorithms as a first stage in developing an expert system for the diagnosis and severity assessment of psoriasis lesions. From the erythemato-squamous diseases dataset taken from UCI (University of California, Irvine) machine repository, only the first ten clinical features are used as input for six state-of-the-art classification algorithms. The accuracy obtained using this set of algorithms is above 93%. The results obtained led to the development of a mobile/desktop medical application that can help the physician in differentiating psoriasis lesions from other erythemato-squamous lesions using only clinical features.