The main goal of the MoleVision project is to develop an application based on deep machine learning algorithms that can be used by patients for the early detection of melanoma as distinct from benign nevi/moles. Ease of use in application and accuracy in diagnosis are the main features of MoleVision. Histopathologically validated image datasets are initially provided by the University Hospital Düsseldorf (UKD) and the University Hospital Bonn (UKB) as well as by two major skin centers (as associated partners). A deep neural convolutional network will be trained with the collected dataset to classify different types of benign nevi/moles and melanomas. One of the main features of MoleVision is online emergency support to rule out malignant skin cancers. To achieve this, a secure and encrypted communication channel between patients and dermatologists is also being implemented. Key highlights of MoleVision are (1) its higher accuracy through deep learning, (2) its functionality with less data volume through generative algorithms, (3) its secure communication channel between patients and medical professionals combined with a secure private cloud solution, and (4) secure and high-quality teledermatology through video consultation support, (5) innovative and smart dermoscopy hardware for home use to provide high-quality skin images and additional information for image classification.
The deep learning algorithms used in MoleVision provide higher accuracy in diagnosis compared to traditional methods.
The application is designed to be user-friendly, making it easy for patients to use and interpret results.
MoleVision includes a secure communication channel between patients and dermatologists to provide online emergency support and rule out malignant skin cancers.
MoleVision provides secure and high-quality teledermatology through video consultations and innovative dermoscopy hardware for home use.
MoleVision implements a secure private cloud solution to protect patient data privacy.