Psoriatic Disease Type Prediction and Analysis Using Deep Feature Learning Model

This work focus on the classification of skin surface images to identify the psoriatic type. To learn and analysis the deep insight of the psoriatic images a custom Convolutional Neural Network (CNN) developed as a prediction model. Before get into the learning process, the input images are involved with segmentation operation. For this purpose, color and texture feature-based segmentation is utilized. The custom architecture of the CNN is formulated to deliver the superior psoriatic disease type prediction result. The model has experimented with native collected data set and performance measures are analyzed. The results shows that the proposed method has high contribute in terms of psoriasis classification and severity grading with an accuracy of 98.94%.