Mobile application-based computer-aided diagnosis of skin tumours from dermal images

Skin tumour is defined as the enormous growth of cells in the skin. The aim is to design standalone application for diagnosis of skin tumours. The dermal images of three different types obtained from the authorized PH2 database are used to analyse the defined image processing algorithm. In this algorithm, pre-processing was performed to remove hair cells. Contour-based level set is used to segment lesion from which clinical and morphological features are extracted for classification. Significant features are obtained with the feature selection technique, Random subset. Classification is performed with three classifiers. The efficiency of the classifier obtained with different trials of classification is analysed with the ANOVA test. With these results, the Multiclass Support vector machine was configured as a suitable classifier to categorize dermal images. Therefore, an application is developed for the analysis of images acquired through mobile with the help of a magnification setup. Thus, extracted features, segmented and original images are transferred to a database for storage.