A-AFMA Ultrasound Challenge: Automatic amniotic fluid measurement and analysis from ultrasound video
The A-AFMA challenge was part of ISBI 2021, with the workshop taking place on April 13, 2021. We are no longer accepting submissions but are busy working on the next stage of the A-AFMA Challenge, which will be running in 2022.
Thank you to all the participants and congratulations to the winning teams:
Task 1 Detection:
1st Sen
2nd Golden Retriever
3rd MUSIC_SZU
Task 2 Localization:
1st DEER-AI
2nd MUSIC_SZU
3rd DEBUT_KELE
***
The A-AFMA Challenge Workshop is on 13th April, 1.30pm - 3.00pm (CET)
Workshop timetable
- Welcome
- Clinical context of amniotic fluid measurement
- The A-AFMA Challenge tasks
- Algorithm presentation: Task 1 winning team
- Algorithm presentation: Task 2 winning team
- Algorithm presentation: Chair’s choice winning team
- Analysis of results from different metrics
- Discussion
Introduction to the Challenge video
Background
Prenatal ultrasound (US) measurement of amniotic fluid is an important part of monitoring whether a fetus is thriving in-utero, as it provides a non-invasive way of assessing whether there is too much or too little amniotic fluid. These are each associated with increased medical problems and in the worst-case, death; whether due to intrauterine growth restriction, maternal diabetes, or another condition. Sonographic amniotic fluid assessment can be performed in various ways; the method currently recommended is a semi-quantitative estimate determined by measuring a maximum vertical pocket (MVP).
Manually estimating the MVP can be time-consuming and requires training and experience to perform, both in terms of identifying which areas of the US scan are amniotic fluid, and, choosing the appropriate markers on ultrasound frames on which to base the measurement. Automating these steps would be a major step towards enabling clinicians without sonography training to perform this aspect of point-of-care fetal surveillance and widen access to areas where pregnancy scans in a hospital are not possible, in both high income and low-and-middle incoming country settings.
Challenge
In this Challenge, we ask participants to consider the problem of automated measurement of the MVP using a predefined ultrasound video clip based on a linear-sweep protocol. Participants are welcome to join both or either of the two sub-challenges.
The first task is to automatically detect amniotic fluid and the maternal bladder.
The second task is to identify the appropriate landmarks for MVP measurement given the selected frame of the video clip.
Each team needs to send a registration email including the team name, team members' names, and affiliations to the organizer via email: noble_pm@eng.ox.ac.uk before accessing the data
Timeline
Registration: opens on December 7th 2020 and closes on 12th March 2021
December 7th 2020: Training dataset has been released.
March 5th 2021: Test dataset has been released
Submission closes: 19th March 2021
Results announced: 30th March 2021