OCELOT 2023: Cell Detection from Cell-Tissue Interaction Banner
Join

OCELOT 2023:

Cell Detection from Cell-Tissue Interaction


Notice❗

Four test cases (586, 589, 609, 615) are excluded due to under-annotated issue. Note that test scores will be re-computed after excluding such cases. See more information in the forum.

News 📣

  • [23.09.11] OCELOT online event was held. Feel free to access the slide and recording. 
  • [23.08.07] The full dataset is now available on Zenodo!
  • [23.07.31] The test phase is opened!
  • [23.06.05] The validation phase is opened!
  • [23.04.14] Registration has started.
  • [23.04.10] The training set is released on Zenodo.
  • [23.03.04] The OCELOT 2023 challenge has been accepted to MICCAI 2023 challenge.
  • [23.02.28] The OCELOT paper has been accepted to CVPR 2023.

Introduction ✏️

Cell detection in histology images is one of the most important tasks in computational pathology. Recently, the OCELOT dataset was released in [1] which provides overlapping cell and tissue annotations on images acquired from multiple organs stained with H&E. [1] showed that understanding the relationship between the surrounding tissue structures and individual cells can boost cell detection performance.

With the newly released OCELOT dataset, we host the OCELOT 2023: Cell Detection from Cell-Tissue Interaction challenge in MICCAI 2023, to promote research on how to utilize cell-tissue relationships for better cell detection. Unlike typical cell detection challenges, participants can utilize tissue patches and annotation for the purpose of boosting cell detection performance.

Please refer to the following links for FAQs:

References:
[1] J Ryu, A V Puche and JW Shin et al. OCELOT: Overlapped Cell on Tissue Dataset for Histopathology. CVPR 2023

Challenge Proceedings 📖

OCELOT challenge will have Springer proceedings as part of the MICCAI 2023 satellite events. Participants who submitted on the test phase (7/31 ~ 8/4) will be invited to submit an LNCS Springer paper (minimum 4 pages of main content) after the MICCAI events. Below is a tentative timeline.

  • MICCAI Conference → 10/8 ~ 10/12
  • Paper submission → 10/16~11/5
  • Review and decision → 11/6 ~ 11/19
  • Final revision → 11/20 ~ 11/26

Prizes 🏆

    After the challenge ends, the OCELOT organizers will prepare a summary of the challenge and submit one paper to a top-tier medical imaging journal. A maximum of 3 members of the top-performing teams in the test leaderboard will be invited as co-authors of the manuscript. Teams to be included in the manuscript will be notified after the OCELOT 2023 challenge event. Furthermore, the top-3 teams will receive the following cash prizes (USD), as per their ranking:

    🥇1st place: $1000
    🥈2nd place: $500
    🥉3rd place: $200

    Lunit may deduct any required tax withholdings from the prize amount based on the domicile of the winning team.

    Contact 📭

    To contact the organizers, please send an e-mail to oncology-ai-research@lunit.io