A manual examination of skin lesions with a dermatoscope is the current standard procedure and state of the art in dermatoscopy. This is a highly laborious process, often limited to a small selection of lesions which excludes registering the state of off-lesion skin regions.
In iToBoS we will design and develop cutting-edge AI-based tools for a number of subcomponents of the decision-making process leading to a novel Cognitive Assistant which integrates the required multi-modal data, machine learning and multi-temporal analytics, clinician knowledge and expertise.
iToBoS aims to provide a more holistic melanoma risk assessment via the examination of patients’ full body skin maps. In order to allow clinicians to efficiently process the abundance of skin image data recorded per patient in a short amount of time, several purposed and Artificial Intelligence (AI) -based diagnostic tools will be developed for providing clinical assistance, e.g., for detecting high-risk lesions, informing about a patient’s skin phenotype, automatically detecting new lesions or change in existing ones, and informing about the recorded history of skin lesion development for each individual lesion recorded with the iToBoS scanner, by exploiting knowledge about state of the art computer vision techniques.
Next to the assessment of melanoma development risk from individual factors, AI-systems for merging multiple data sources, and compiling cross-document reports from, e.g., text-based patient history and features extracted from skin imagery, under consideration of ongoing additions of novel data points will be developed.