2nd Data-Centric Land Cover Classification Challenge: Ranking Samples with AI-Based Systems

The 2nd Data-Centric Land Cover Classification Challenge is here! Organized in conjunction with the British Machine Vision Conference (BMVC) 2024, this challenge focuses on developing an AI-based ranking system to rank samples based on their levels of label noise. Participants will be provided with a semantic segmentation dataset consisting of 5,000 256×256 images and their corresponding (noisy) labels. Success will be measured by comparing the submitted ranking with an undisclosed one and calculating the Kendall Tau score.

The top-ranked methods’ authors will be invited to present their approaches at the Workshop in Glasgow, UK, on 28 November 2024, and co-author a journal paper summarizing the challenge’s outcome. The challenge deadline is 10 November 2024. Register and learn more at the official challenge website: https://mveo.github.io/challenge.html