MoleVisionMoleVision – Precise Dermatological Skin Cancer Diagnostics through Deep Learning
AI-powered application for early melanoma detection with secure patient–physician communication.
Official Project Mark
Federal Ministry of Research, Technology and Space
Medical AI & Teledermatology Research Initiative
Duration: 42 Months
4 Partners
Universities, Industry & Research Centers
Project Description & Objectives
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.
Petanux Specific Project Responsibilities
- •Deep convolutional neural network architecture for melanoma lesion classification
- •Optical dermoscopy feature extraction & multispectral image preprocessing
- •Privacy-preserving teledermatology software architecture & secure physician-patient channel
Project partners
University of Siegen
University Hospital Düsseldorf (UKD)
University Hospital Bonn (UKB)
Petanux GmbH
Operated and supported by VDI Technologiezentrum
This federal research grant is officially managed, supervised, and supported on behalf of the Federal Ministry by VDI Technologiezentrum.
Key highlights
Higher accuracy through deep learning model optimization
Functionality with less data volume through generative data augmentation
Secure communication channel between patients and medical professionals with private cloud infrastructure
High-quality teledermatology support for remote video consultation
Get in touch with our Research and Innovation Team
Let's get in touch to explore research partnerships, grant proposals, and joint innovation initiatives.