On Monday, February 17, the VICOROB Institute at the University of Girona hosted an insightful workshop on “Advancing Skin Lesion Detection with 3D-TBP” as part of the iToBoS project. The event highlighted groundbreaking advancements in skin lesion detection using 3D Total Body Photography (3D-TBP) and brought together researchers, clinicians, and data science enthusiasts.
Distinguished Speakers
The workshop featured esteemed experts in the field, including:
- Professor Rafael Garcia, Director of VICOROB and the iToBoS Project, University of Girona
- Dr. Josep Malvehy, Director of the Melanoma Unit at Hospital Clinic of Barcelona (Challenge Director)
- Hayat Hussein Rajani, VICOROB Research Team
- Solomon Chibuzo Nwafor and Muhammad Faran Akram, winners from Team NinjaX, who presented their winning solution
The iToBoS 2024 Skin Lesion Detection Competition, hosted on Kaggle, was a major component of the workshop. Participants developed state-of-the-art machine learning techniques to detect multiple skin lesions from anonymized 3D avatars generated using the Canfield VECTRA WB360 system. The dataset included images from diverse populations, offering a robust foundation for AI-driven dermatological research.
This workshop provided a valuable learning opportunity for students from the Intelligent Field Robotic Systems (IFROS), Master’s in Intelligent Robotic Systems (MIRS), and Medical Imaging and Applications (MAIA), who gained valuable insights in AI applications beyond robotics, including healthcare and other technological sectors. Many students from these programs participated in the competition, expanding their expertise in artificial intelligence, machine learning, and image processing. Notably, Team NinjaX, the winning team, was composed of IFROS students, further demonstrating the impact of AI-driven innovation and the interdisciplinary applications of these cutting-edge techniques.
This achievement highlights the strong AI foundation that the IFROS Master’s provides, enabling students to develop solutions that impact multiple domains beyond robotics.
Participants were encouraged to publish their findings on arXiv and open-source their code to contribute to future advancements in AI-driven dermatology.
Revolutionizing Skin Cancer Diagnosis with AI
Melanoma, responsible for 60% of lethal skin neoplasia, is one of the most aggressive cancers, with its incidence steadily rising. Early detection is crucial, but conventional total body skin examinations are time-consuming, especially for patients with numerous pigmented lesions.
The iToBoS (Intelligent Total Body Scanner for Early Detection of Melanoma) project is developing an AI-powered diagnostic platform to enhance early melanoma detection. This platform integrates a total body scanner and a Computer-Aided Diagnostics (CAD) tool capable of combining medical records, genomics data, and imaging for a highly personalized diagnosis.
The next-generation total body scanner, based on an existing prototype, will utilize high-resolution cameras with liquid lenses to achieve unparalleled image quality. Combining these images with machine learning algorithms, the system will serve as a new dermoscopic diagnostic tool, ensuring fast, reliable, and patient-specific diagnostics.