3rd International Endoscopy Computer Vision Challenge and Workshop  (EndoCV2021)

In conjunction with IEEE International Symposium on Biomedical Imaging (ISBI2021), Nice, France 

Title: Addressing generalisability in polyp detection and segmentation

Computer-aided detection, localization, and segmentation methods can help improve colonoscopy procedures. Even though many methods have been built to tackle automatic detection and segmentation of polyps, benchmarking and development of computer vision methods remains an open problem. This is mostly due to the lack of datasets or challenges that incorporate highly heterogeneous dataset appealing participants to test for generalisation abilities of the methods. we aim to build a comprehensive, well-curated, and defined dataset from 6 different centres worldwide and provide 5 datasets types that include: i) multi-centre train-test split from 5 centres ii) polyp size-based split (participants should do this by themselves if of interest), iii) data centre wise split, iv) modality split (only test) and v) one hidden centre test. Participants will be evaluated on all types to address strength and weaknesses of each participants’ method.  Both detection bounding boxes and pixel-wise segmentation of polyps will be provided. Additionally, we also release negative samples. 


--> Special Note regarding leaderboard result on detection task reported here: Please note that the leaderboard results for detection have very high mAP values due to a bug in the evaluation code that has now been fixed in our GitHub evaluation code!!! Also, the compiled paper for joint journal has a large number of sequence data for which the team ranking did change on the unseen data. Note that the winners were based on both accuracy, speed and robustness tests!!!

--> Training data release today!!! (5th February, 2021). Apologies for delay.

--> If you have not yet submitted the data and challenge consent form then please do here! (please do not send multiple copies unless asked!!!)

--> GitHub for evaluation codes

--> EndoCV2021 participation declaration form sent to participants (18th February 2021)

--> Thanks to the BMRC cloud computing infrastructure at the University of Oxford for supporting our "cloud-based inference on test data" 

--> Cloud-based inference on test data and leaderboard submission extended to 25th February 2021 (PST)

--> We are no longer taking participations. The participation for this year has closed!!!

--> If you have interesting work on computer vision/deep learning work on endoscopy or surgery then do consider to submit 4 page paper for the workshop. (Intention to submit deadline 10th March 2021, see important dates )

CMT submission: https://cmt3.research.microsoft.com/EndoCV2021

Past challenges and workshops:

2nd International Endoscopy Computer Vision Challenge and Workshop, Iowa, USA (EndoCV2020)

1st International Endoscopy Computer Vision Challenge and Workshop, Venice, Italy (EAD2019)

Blog posts:

Adrian Krenzer writes about EndoCV2020 (here) ---> "Winner of EDD2020 sub-challenge"

Gorkem writes about EndoCV2020 (here)---> "Winner of EAD2020 sub-challenge"

Travel grant winner Refika Sultan DOĞAN writes about EAD2019 (here)


Youtube Channel (EndoCV2020 workshop) 

EndoCV2021 featured on RSIP vision

Workshop proceedings (open):

EAD2019:  http://ceur-ws.org/Vol-2366/
EndoCV2020: http://ceur-ws.org/Vol-2595/

Joint Journals:

  • Ali, S., Zhou, F., Braden, B. et al. An objective comparison of detection and segmentation algorithms for artefacts in clinical endoscopy. Sci Rep 10, 2748 (2020). https://doi.org/10.1038/s41598-020-59413-5
  • Ali, S., Dmitrieva, M., Ghatwary, N., Bano, S. et al., Deep learning for detection and segmentation of artefact and disease instances in gastrointestinal endoscopy. Medical Image Analysis 2021 (in press), (Open access).

Sponsored by: