STS 2026 : The 4th Semi-supervised Teeth Segmentation Challenge on Metal Artifact Reduction and Beyond

MICCAI 2026, Strasbourg, France

September 27-October 1, 2026

🎉Result Release!🎉 Task 1: Segmentation Task 2: Registration Task 3: MMDental

Task 1 Winners: Metal Artifact CBCT Teeth Segmentation

Task 1 evaluates tooth and root-canal instance segmentation in CBCT scans affected by metal artifacts caused by restorations, crowns, bridges and implants. The final ranking is based on the official Docker test evaluation published on the Codabench competition page; the Codabench validation leaderboard is not used for the final ranking.

Teams are ranked by the official Challenge_Score, which combines image-level and instance-level Dice, mIoU and NSD with Identification Accuracy (IA); cases containing metal artifacts are weighted twice. All nine finalists completed the full test set (50 / 50 cases), and the top three teams are highlighted below.

Team Challenge Score (↑) Instance-level Image-level IA (↑) Final Rank Links
Dice (↑) mIoU (↑) NSD (↑) Dice (↑) mIoU (↑) NSD (↑)
Usercccc2026 0.731486 0.695186 0.627607 0.811620 0.869955 0.852618 0.952289 0.540943 🥇 1
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
Userthisxu 0.730807 0.712187 0.668078 0.762043 0.849465 0.831939 0.922946 0.597627 🥈 2
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
Userfy666 0.717407 0.702382 0.660873 0.751739 0.836681 0.814385 0.906636 0.557107 🥉 3
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
UserJLShen025 0.711879 0.683602 0.623625 0.774654 0.859501 0.820476 0.930402 0.525884 4
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
User3dvmar 0.675748 0.634744 0.553767 0.769560 0.843188 0.792146 0.930322 0.461458 5
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
Userdjw 0.647416 0.618031 0.549564 0.714338 0.864749 0.806441 0.913964 0.385807 6
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
Usersjtu426lab 0.627208 0.564304 0.477657 0.761922 0.771624 0.754306 0.906851 0.393016 7
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
Userfeelok 0.474232 0.407963 0.367084 0.460241 0.815621 0.754404 0.913516 0.360640 8
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]
Usermy311 0.349353 0.302010 0.244993 0.401544 0.557699 0.412312 0.697442 0.203030 9
  1. [ReportReport]
  2. [SlidesSlides]
  3. [CodeCode]

Challenge_Score = 0.15 × image_score + 0.75 × instance_score + 0.10 × IA, where image_score is the mean of image-level Dice, mIoU and NSD and instance_score = 0.35 × Dice + 0.25 × mIoU + 0.40 × NSD at instance level. Segmentation efficiency is not part of the 2026 ranking.

“Report” opens the team’s technical report, “Slides” the presentation deck shared for the STS 2026 winners’ session, and “Code” the team’s public repository. Greyed-out items were not released by the team.

Official results are published on the Codabench competition page. See the evaluation protocol for metric definitions and penalty values.

Copyright @ College of Media Engineering, Communication University of Zhejiang, Hangzhou, 310018, China.