Convolutional neural networks in skin cancer detection using spatial and spectral domain

摘要

Skin cancers are world wide deathly health problem, where signi cant life and cost savings could be achieved ifdetection of cancer can be done in early phase. Hypespectral imaging is prominent tool for non-invasive screening.In this study we compare how use of both spectral and spatial domain increase classi cation performance ofconvolutional neural networks. We compare ve di erent neural network architectures for real patient data. Ourmodels gain same or slightly better positive predictive value as clinicians. Towards more general and reliablemodel more data is needed and collection of training data should be systematic.