Transfer learning for endoscopy disease detection and segmentation with mask-RCNN benchmark architecture

Rezvy, Shahadate, Zebin, Tahmina ORCID: https://orcid.org/0000-0003-0437-0570, Braden, Barbara, Pang, Wei, Taylor, Stephen and Gao, Xiahong (2020) Transfer learning for endoscopy disease detection and segmentation with mask-RCNN benchmark architecture. In: 2020 IEEE 17th International Symposium on Biomedical Imaging, 2020-04-03 - 2020-04-07, Iowa City.

Preview

PDF (Author accepted manuscript EDD2020) - Accepted Version Available under License Unspecified licence. Download (2MB) | Preview

Abstract

We proposed and implemented a disease detection and semantic segmentation pipeline using a modified mask-RCNN infrastructure model on the EDD2020 dataset. On the images provided for the phase-I test dataset, for 'BE', we achieved an average precision of 51.14%, for 'HGD' and 'polyp' it is 50%. However, the detection score for 'suspicious' and 'cancer' were low. For phase-I, we achieved a dice coefficient of 0.4562 and an F2 score of 0.4508. We noticed the missed and mis-classification was due to the imbalance between classes. Hence, we applied a selective and balanced augmentation stage in our architecture to provide more accurate detection and segmentation. We observed an increase in detection score to 0.29 on phase -II images after balancing the dataset from our phase-I detection score of 0.24. We achieved an improved semantic segmentation score of 0.62 from our phase-I score of 0.52.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	deep learning,computer vision,endoscopy, gastrointestinal,sdg 3 - good health and well-being ,/dk/atira/pure/sustainabledevelopmentgoals/good_health_and_well_being
Faculty \ School:	Faculty of Science > School of Computing Sciences
UEA Research Groups:	Faculty of Science > Research Groups > Smart Emerging Technologies
Depositing User:	LivePure Connector
Date Deposited:	08 Apr 2020 00:24
Last Modified:	29 Apr 2024 23:36
URI:	https://ueaeprints.uea.ac.uk/id/eprint/74726
DOI:

Downloads

Actions (login required)

View Item