Dear participants,

First and foremost, a heartfelt thank you for your participation in the HaN-Seg Challenge! 😊🎉

For the final analysis, we identified the best-performing method for each team based on the Dice coefficient (DC) and 95th percentile Hausdorff distance (HD95). Seven teams submitted their methods to the Final Test Phase of the HaN-Seg challenge, however, one team was excluded from further analysis due to trivial segmentation results.

The remaining six teams and their methods were evaluated using the DC and HD95. Following the challenge rules, we applied statistical ranking to each metric through a pairwise comparison of the methods using the Wilcoxon signed-rank test with Bonferroni correction. This process resulted in a significance score and metric-specific rank. Identical ranks were assigned to algorithms that showed only marginal performance differences to evaluate only statistically significant differences among methods. The final rank was determined by aggregating the ranks over both metrics.

After reviewing the Centerline DC metric (which the Metrics Reloaded framework proposes in the presence of tubular structures), we empirically found that it does not provide an objective and robust way to compare tubular organs and was thus excluded from our evaluation metrics. This decision did not affect team ranking in the Final Phase of the challenge.

Please note that the official ranking may differ from the online leaderboard, as the latter only allows ranking based on mean values. However, the challenge rules stipulated statistical ranking, which was used to determine the final rankings.

🏆 Without further delay, we are thrilled to announce that the winner of the HaNSeg challenge is Team eli1. Second place goes to Team cwlg102, and we have a tie for third place between Team CHB-QuantIF and Team Mamaa. 🏆

For more details regarding ranking and results, please refer to our challenge results paper. 📑

🎉 Congratulations to all the teams for your hard work and achievements! 🎉