Hanover, Germany - Lennart Jütte, a nanotechnologist and PhD candidate at the Hanover Center for Optical Technologies (HOT) at Leibniz University Hanover, has launched an innovative project called Melasome, which combines Generative Adversarial Networks (GANs) and art therapy for skin cancer patients.
Lennart Jütte usually works on developing an intelligent full-body scanner for early detection of melanomas as part of the Horizon 2020 EU project iToBoS (itobos.eu). His research focuses on innovative approaches such as Mueller Matrix Polarimetry and imaging for non-contact dermoscopy and whole-body dermoscopy. Artificial intelligence (AI) algorithms are also used to improve medical diagnostics and make them more understandable for physicians and patients.
The Melasome project now extends the applications of artificial intelligence algorithms with a completely new approach. It uses cyclic Generative Adversarial Networks (cycleGANs) to transform dermoscopy images of melanoma into unique, personalized artworks. Cycle GANs are a type of artificial neural networks that allow images to be transferred from one domain to another without the need for paired training data by learning the style of the target domain and applying it to the input images. This novel approach aims to potentially alleviate the often very high psychological stress associated with a skin cancer diagnosis. Current studies show that every third melanoma diagnosis is accompanied by mental health issues that can directly and indirectly affect treatment outcomes. Art therapy has been working on methods to alleviate this burden and support skin cancer treatments for several years. The idea of using generative AI for art therapy in medicine now represents a promising new approach that, for the first time, harnesses the great potential of artificial intelligence algorithms in the field of image recognition and classification to transform potentially traumatizing images of skin cancer sites created during medical diagnostics into artful, mental health-supporting images. The motifs can be individually adapted to the patients, also a task ideally suited for AI algorithms. The approach was presented to the patient community in a lecture and exhibition at a conference of the Melanoma Patient Networks Europe (MPNE) and received very positive feedback.
Lennart Jütte said, "I firmly believe that by combining cutting-edge technology like Generative Adversarial Networks with the real experiences of melanoma patients during diagnosis and therapy, we can make a positive contribution to the mental health of patients. I am highly motivated to further develop the Melasome project and explore new ways to support patients through technology and art in their treatment." Prof. Dr. Bernhard Roth, whose group at HOT is conducting research on skin cancer detection, also emphasized the importance of innovation and technology for progress in diagnostics and therapy, highlighting that new ideas often emerge from initial scientific curiosity and the combination of knowledge from different disciplines.
For more information about the Melasome project, please visit the website www.melasome.com.
