Skin cancer is one of the most prevalent forms of cancer globally, reaching epidemic proportions.
Automating its diagnosis could have a significant impact on facilitating early detection and triage, particularly in areas with limited access to specialized care. The detection of suspicious lesions is a critical first step before any further analysis can be made. However, to the best of our knowledge, no publicly available datasets exist to support the development and benchmarking of AI models for this important task.
To support this goal, we have created a novel dataset using anonymized 3D body scans, offering a comprehensive, high-resolution view of nearly the entire skin surface of each patient. The images in this dataset are crops of skin regions extracted from such scans of about a 100 patients. The dataset also provides additional context in the form of metadata such as anatomical location, age group, and sun damage score for each image.
As such, with the goal of advancing research in automated skin lesion detection, we are excited to launch the iToBoS Skin Lesion Detection Challenge on Kaggle and invite participants from around the world to join us. The challenge is already live and will remain open until February 9, 2025. The dataset and all competition details can be found on the challenge page.
Additionally, the iToBoS consortium will host an online-only workshop on February 28, 2025. This event will feature a recap of the competition, presentations from leading researchers and industry experts, and an opportunity for winning teams and selected participants to showcase their solutions. The workshop is free to attend and open to all interested researchers, students, and enthusiasts. The detailed schedule and format will be updated on the challenge page and on the iToBoS website shortly. We look forward to your participation!
Learn more at iToBoS 2024 - Skin Lesion Detection with 3D-TBP.