HomeWhy Challenges?All ChallengesCreate Your Own ChallengeContributorsForum
Sign in / Register

MICCAI 2017 Satellite Event

RETOUCH results were announced on Sep 14th, 2017 at a joint OMIA-RETOUCH workshop at MICCAI 2017 in Quebec City, Canada, and are summarized below. The challenge consisted of 70 training datasets (OCT scans with reference annotations) and 42 test datasets (OCT scans, 14 per Cirrus/Spectralis/Topcon device). Eight teams participated in the RETOUCH challenge by submitting the results on the test set and providing a paper describing their algorithm.

 

The 1st place was awarded to team SFU-ENSC (School of Engineering Science, Simon Fraser University, Canada). 
         Donghuan Lu, Morgan Heisler, Sieun Lee, Gavin Ding, Marinko V. Sarunic, and Mirza Faisal Beg:
         "Retinal Fluid Segmentation and Detection in Optical Coherence Tomography Images using Fully Convolutional Neural Network".

               

 

Participating Teams

Name (alphabetically) Contact Paper
Helios IIIT Hyderabad, Hyderabad, India Shivin Yadav PDF
MABIC National Institute for Mathematical Sciences, Daejeon, Korea Sungho Kang PDF
NJUST Nanjing University of Science & Technology, China Qiang Chen PDF
RetinAI RetinAI Medical GmbH and University of Bern, Switzerland Stefanos Apostolopoulos PDF
RMIT RMIT University, Melbourne, Australia Ruwan Tennakoon PDF (suppl)
SFU Simon Fraser University, Burnaby, Canada Donghuan Lu PDF
UCF University of Central Florida, Orlando, US Dustin Morley PDF
UMN University of Minnesota, Minneapolis, US Abdolreza Rashno PDF

 


RESULTS

The average rank-score across the two tasks (detection and segmentation) determined the final RETOUCH ranking. In case of a tie, the segmentation performance had the preference.

     

 

 

Segmentation Task Details

Each team received a rank (1=best) for each combination of: Fluid type x OCT device x Error measure, based on the mean error measure value over the corresponding set of test images. The segmentation task score was determined by adding the 18 individual ranks. The team with the lowest segmentation score was ranked #1 on the segmentation task leaderboard.

                                                                             

Dice Score (DSC)

 

Absolute Volume Difference (AVD)

 

Detection Task Details

For each fluid type (IRF, SRF and PED): the Receiver Operating Curve (ROC) was created across all the test set images and an area under the curve (AUC) was calculated. Each team received a rank (1=best) for each of the fluid types based on the obtained AUC value. The detection task score was determined by adding the three fluid type ranks. The team with the lowest detection score was ranked #1 on the detection task leaderboard.
 

 

      

 


Consortium for Open Medical Image Computing © 2012-