{ "id": 304917, "title": "TIDE: A General Toolbox for Identifying Object Detection Errors", "url": "/competitions/tensorflow-great-barrier-reef/discussion/304917", "author_name": "Bilzard", "post_date": "2022-02-03T01:33:35.665000", "votes": 7, "comment_count": 0, "views": 0, "content": "

This tool can help us to analyze detection error of our model.
\nFor example, it breakdown the cause of FP/FN more granular category(Fig 1), and show us as a piechart(Fig 2).

\n

How to use

\n
    \n
  1. convert the GBR dataset to COCO format json file[4, 5]
    2. \n
  2. save the models prediction as COCO format json file
    3. \n
  3. evaluate with TIDE
    4. \n
\n

Reference

\n\n
\n

Fig 1.
\n\"Screen-Shot-2022-02-03-at-10-21-42\"

\n

Fig 2.
\n\"Screen-Shot-2022-02-03-at-10-38-57\"

\n

Fig 3.
\n\"Screen-Shot-2022-02-03-at-10-21-58\"

", "messages": [ { "id": 1673784, "postDate": "2022-02-03T01:33:35.667Z", "content": "

This tool can help us to analyze detection error of our model.
\nFor example, it breakdown the cause of FP/FN more granular category(Fig 1), and show us as a piechart(Fig 2).

\n

How to use

\n
    \n
  1. convert the GBR dataset to COCO format json file[4, 5]
    2. \n
  2. save the models prediction as COCO format json file
    3. \n
  3. evaluate with TIDE
    4. \n
\n

Reference

\n\n
\n

Fig 1.
\n\"Screen-Shot-2022-02-03-at-10-21-42\"

\n

Fig 2.
\n\"Screen-Shot-2022-02-03-at-10-38-57\"

\n

Fig 3.
\n\"Screen-Shot-2022-02-03-at-10-21-58\"

", "rawMarkdown": "This tool can help us to analyze detection error of our model.\nFor example, it breakdown the cause of FP/FN more granular category(Fig 1), and show us as a piechart(Fig 2).\n\n# How to use\n\n1. convert the GBR dataset to [COCO](https://cocodataset.org/#home) format json file[4, 5]\n2. save the models prediction as COCO format json file\n3. evaluate with TIDE\n\n# Reference\n\n- [1] Paper: https://dbolya.github.io/tide/paper.pdf\n- [2] GitHub: https://github.com/dbolya/tide\n- [3] Youtube Video: https://www.youtube.com/watch?v=McYFYU3PXcU\n- [4] COCO data format: https://cocodataset.org/#format-data\n- [5] COCODetectionDataset: https://voxel51.com/docs/fiftyone/user_guide/dataset_creation/datasets.html#cocodetectiondataset\n\n---\n\nFig 1.\n\"Screen-Shot-2022-02-03-at-10-21-42\"\n\nFig 2.\n\"Screen-Shot-2022-02-03-at-10-38-57\"\n\nFig 3.\n\"Screen-Shot-2022-02-03-at-10-21-58\"\n", "votes": 7 } ], "comments": [], "raw_markdown_by_id": { "1673784": "This tool can help us to analyze detection error of our model.\nFor example, it breakdown the cause of FP/FN more granular category(Fig 1), and show us as a piechart(Fig 2).\n\n# How to use\n\n1. convert the GBR dataset to [COCO](https://cocodataset.org/#home) format json file[4, 5]\n2. save the models prediction as COCO format json file\n3. evaluate with TIDE\n\n# Reference\n\n- [1] Paper: https://dbolya.github.io/tide/paper.pdf\n- [2] GitHub: https://github.com/dbolya/tide\n- [3] Youtube Video: https://www.youtube.com/watch?v=McYFYU3PXcU\n- [4] COCO data format: https://cocodataset.org/#format-data\n- [5] COCODetectionDataset: https://voxel51.com/docs/fiftyone/user_guide/dataset_creation/datasets.html#cocodetectiondataset\n\n---\n\nFig 1.\n\"Screen-Shot-2022-02-03-at-10-21-42\"\n\nFig 2.\n\"Screen-Shot-2022-02-03-at-10-38-57\"\n\nFig 3.\n\"Screen-Shot-2022-02-03-at-10-21-58\"\n" } }