EndoCV2021-polyp_det_seg_gen

Maintained by: Sharib Ali

About:

This repo is designated for the evaluation of methods developed in EndoCV2021 (3rd International Endoscopy Computer Vision Challenge and Workshop in conjuenction to IEEE ISBI 2021) challenge, "Addressing generalisability in polyp detection and segmentation". Several tools that may help participants to assess their methods are provided.

Where to find what?

Evaluating detection methods:

• Convert voc format to COCO format (to single json file)

• Evaluate your prediction (COCO .json format) output with GT (.json COCO format)

• Evaluate your prediction (VOC .txt format) with GT (.txt VOC format) Deprecated!!! but participants can use if they wish

Evaluating generalisability in detection methods:

• Evaluation for testing generalisability between two detection results on multiple datasets

Evaluating segmentation methods:

• Evaluate segmentation methods for each test set separately

Evaluating generalisability in segmentation methods:

• Evaluate method generalisability

References (details on generalisation tests can be found here):

Please cite these papers

[1] 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

[2] Ali et al. Deep learning for detection and segmentation of artefact and disease instances in gastrointestinal endoscopy. Medical Image Analysis 2021.

[3] Ali, S. et al. A multi-centre polyp detection and segmentation dataset for generalisability assessment. Sci Data 10, 75 (2023). https://doi.org/10.1038/s41597-023-01981-y