Differentiating Atopic Dermatitis and Psoriasis Chronic Plaque using Convolutional Neural Network MobileNet Architecture

The use of Convolutional Neural Network (CNN) has been rising in the field of image recognition and classification. Different kinds of CNN architectures were implemented in various skin disease classification systems. In this paper, a skin disease classification for differentiating Atopic Dermatitis and Psoriasis Chronic Plaque using MobileNet architecture of Convolutional Neural Network was implemented in a Raspberry Pi. The MobileNet model was trained using the images selected from AtlasDerm, Danderm, DermIs, DermNet, DermQuest and DermatoWeb as the dataset. The network was tested using 30 captured images from patients using a Raspberry Pi camera which were verified by the Dermatologist. Confusion matrix was utilized to compute for the accuracy of the model between the two skin diseases. As a result, the system achieved an accuracy of 88% and 90% in classifying Atopic Dermatitis and Psoriasis Chronic Plaque, respectively. The use of MobileNet architecture of CNN allowed the high accuracy classification in differentiating skin diseases.