[visionlist] Edoscopic Surgeon Action Detection Challenge at MIDL

[Gurkirt Singh]

SARAS Endoscopic  Surgeon Action Detection (ESAD) challenge is being
organised at the Medical Imaging with Deep Learning (MIDL, 2020) conference.
The MIDL conference will be held from 6 to 8 July 2020 in Montreal, Canada.

Link to challenge: https://saras-esad.grand-challenge.org/

Link to conference: https://2020.midl.io/

Link to baseline code: https://github.com/Viveksbawa/SARAS-ESAD-baseline

Important dates:

Challenge opens for registration: 1 March 2020

Training/Validation data release: 20 March 2020

Test data release: 1 June 2020

Result submission deadline: 15 June 2020

Final result announcement: 18 June 2020

Keywords: Surgeon action detection, prostatectomy, endoscopic video

About:

The task presented in the competition is very challenging considering the
complexity of the surgical scene. This challenge has organised the list of
actions performed by surgeons into 21 different classes. Each frame of the
dataset can have more than one action.

To the best of our knowledge, this challenge presents the first benchmark
dataset for action detection in the surgical domain, and paves the way for
the introduction, for the first time, of partial/full autonomy in surgical
robotics. Within computer vision, other datasets for action detection
exist, but are of limited size.


Reward: # let's talk about this

Winner and runner up from the challenge will be invited to give a talk in
the challenge event which will be organised on the day of the conference.
Authors from the winner and runner up methods will also be invited as
co-authors in publication summarizing the challenge. Authors from other
methods can be made co-authors as well, based on the novelty to their
methods.

Call for participation:

SARAS-ESAD 2020: SARAS endoscopic vision challenge for surgeon action
detection invites researchers from medical and general computer vision
community to participate in the challenge. The challenge aims to detect
actions performed by the surgeon during the prostatectomy procedure.

Baseline code and data reading pipline:

The baseline model can be downloaded from SARAS-ESAD-baseline
<https://github.com/Viveksbawa/SARAS-ESAD-baseline> repository at GitHub.

Features of the baseline code:

1. Data organising instructions and helping scripts.

2. Pytorch <https://pytorch.org/> dataloader for the provided dataset.

3. Pytorch <https://pytorch.org/> code for training a baseline model.

4. Evaluating the trained model on the validation dataset.

5. Generation of submission file given set of input images

We hope this will help kick start more teams to get up to speed and allow
time for more innovative solutions.

Goodluck!
Orgnising team
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://visionscience.com/pipermail/visionlist_visionscience.com/attachments/20200416/edb829b7/attachment.html>