[Corpora-List] Shared Task on Medical Video Question Answering

Demner Fushman, Dina (NIH/NLM/LHC) [E] ddemner at mail.nih.gov
Thu Jan 13 03:32:52 CET 2022

Apologies for cross-posting:

From Deepak Gupta:

Dear colleagues and friends,

We propose the Shared Task on Medical Video Question Answering as part of BioNLP 2022<https://aclweb.org/aclwiki/BioNLP_Workshop> at ACL 2022<https://www.2022.aclweb.org/>. This task aims at developing models for (1) predicting the medical instructional videos and (2) locating the visual answer in the video for the medical or health-related question.

Task Website: https://medvidqa.github.io/

Task Description

Task 1: Medical Video Classification (MVC)

Given an input video, the task is to categorize the video into one of the following classes:

* Medical Instructional: An instructional medical video for non-professionals should clearly demonstrate a medical procedure providing enough details to reproduce the procedure and achieve the desired results without prior training.

* Medical Non-instructional: A medical video on a discussion of medical-related topics without any visual answer to any medical or health-related question.

* Non-medical: A video can be categorized as non-medical if the video is neither medical instructional nor medical non-instructional.

Task 2: Medical Visual Answer Localization (MVAL)

Given a medical or health-related question and a video, the task aims to locate the temporal segments (start and end timestamps) in the video where the answer to the medical question is being shown, or the explanation is illustrated in the video. This task seeks to find a video segment with a visual answer to the natural language question. The MVAL task can be considered as finding a series of “medical instructional activity-based frame localization” where a potential solution first searches for all medical instructional activity for a given medical question and then localizes a particular activity that is aligned to a given medical or health-related question in an untrimmed medical-instructional video.

Important Dates

* First call for participation: January 10, 2022

* Release of the training and validation datasets: January 10, 2022

* Release of the test sets: February 21, 2022

* Run submission opens on CodaLab: February 21, 2022

* Run submission deadline: February 25, 2022

* Release of the official results: March 2, 2022

* System description paper due: March 11, 2022

* Notification of acceptance: March 25, 2022

* Camera-ready papers due: April 4, 2022

* BioNLP workshop @ACL22: May 26, 2022

We want to kindly invite you to participate in MedVidQA 2022, this year with the new task on medical videos, join us!

Join our Google Group<https://groups.google.com/g/medvidqa2022/> for important updates! If you have any questions, ask in our Google Group<https://groups.google.com/g/medvidqa2022/> or email us.

-------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 25900 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20220113/f6ceb947/attachment.txt>

More information about the Corpora mailing list