ENTRep Challenge:
Advancing Vision-Language AI for
ENT Endoscopy Analysis

News

📢 We release the Public Dataset with Dataset Usage Guideline

📢 We provide the training dataset for three tasks.

📢 Public Test Submission Links for Track 1, Track 2, and Track 3 are now available on Codabench!

📢 Public Test Submission for Track 1, Track 2, and Track 3 on Codabench has now closed.

📢 Private Test Submission for Track 1, Track 2, and Track 3 is now open on Codabench!

Announcement

PRIVATE PHASE

Dear participants,

You are currently in the Private Phase of the ENTREP Challenge hosted on Codabench. Please take note of the following important information:

The current leaderboard is temporary and DOES NOT REFLECT FINAL RANKINGS.
All teams, including your own, will see “n/a” for all metrics such as Accuracy, Precision, Recall, and F1.
The leaderboard is currently sorted by submission time only — teams that submit earlier will appear higher. This does not indicate performance.
Only one submission per team will be evaluated at the end of the Private Phase. This is the submission that is displayed on the leaderboard (with “n/a”). All other submissions will be ignored and not scored.
Once the Private Phase ends, the organizers will re-evaluate the selected submissions using a hidden private test set.
The final leaderboard will be updated based on the actual results from the private evaluation and will determine the official rankings.
Please make sure that the selected submission is the best representation of your team’s solution.

Best of luck, and we look forward to seeing your final results!
ENTREP Challenge Organizing Team

ENTRep Challenge: Advancing Vision-Language AI for ENT Endoscopy Analysis

Challenge Overview: Step into the forefront of medical AI research with the ENTRep Challenge, organized as part of ACM MM 2025 in Dublin, Ireland. This initiative offers exclusive access to a high-quality, expert-annotated dataset of ENT endoscopy images collected at Thong Nhat Hospital in Ho Chi Minh City, Vietnam. Clinical specialists have meticulously labeled each image to capture various ENT pathologies and anatomical nuances. In this challenge, you will develop innovative solutions across three pivotal tasks:

Image Classification: Develop algorithms that precisely classify ENT endoscopy images by anatomical region and pathology, paving the way for more accurate diagnoses and streamlined clinical workflows.
Image-to-Image Retrieval: Design efficient systems that can swiftly retrieve visually similar images from an extensive database, enabling clinicians to conduct comparative analyses and make informed decisions.
Text-to-Image Retrieval: Create advanced methods to bridge the gap between clinical narratives and visual data, by matching detailed text descriptions with their corresponding images to enhance data accessibility and diagnostic precision.

Leaderboards

Track 1: Image Classification

Tack 2: Image-to-Image Retrieval

Track 3: Text-to-Image Retrieval

Organizers

For questions, please contact the organizers at ntthuan@selab.hcmus.edu.vn.

Thanh Dinh Le

Thong Nhat Hospital and University of Health Science,
VNU-HCM, Vietnam

Minh-Triet Tran

University of Science,
VNU-HCM, Vietnam

Tam V. Nguyen

University of Dayton,
Ohio, USA

Thao Thi Phuong Dao
Thong Nhat Hospital and University of Science,
VNU-HCM, Vietnam

Trong-Thuan Nguyen

University of Science,
VNU-HCM, Vietnam

Trung-Nghia Le

University of Science,
VNU-HCM, Vietnam

Viet-Tham Huynh

University of Science,
VNU-HCM, Vietnam

Trong-Le Do

University of Science,
VNU-HCM, Vietnam

Quang-Thuc Nguyen

University of Science,
VNU-HCM, Vietnam

Call for Papers

We would like to invite submissions of both short and long papers. Short papers (non-archived) should be formatted in the ACM MM style and are limited to 4 pages, excluding references. Long papers (archived) must also follow the ACM MM format and can be up to 8 pages, excluding references. All submissions must adhere to the ACM MM submission policies.

Topics for the workshop include, but are not limited to:

Medical Large Language Models
Information Retrieval
Multi-Modal Learning
Explainable AI in Medical Imaging
Clinical Applications and Integration
Domain Adaptation and Transfer Learning
Synthetic Data Generation
Automated Medical Report Generation
AI for Rare Pathologies
Privacy-Preserving AI
Interactive and Clinician-AI Collaboration Systems

Submission website: https://cmt3.research.microsoft.com/ENTRep2025.

Important Dates (Anywhere on Earth)

Registration opened: 26/3/2025
Public data released: 15/4/2025
Registration closed: 31/5/2025
Public session opened: 05/6/2025
Public session closed: 22/6/2025
Private data released: 23/6/2025
Private session opened: 23/6/2025
Private session closed: 30/6/2025
Results, Report, Paper submission deadline: 30/7/2025
Notification: 10/8/2025
Camera-ready submission: 20/8/2025

Page updated

Report abuse