![\"MD.ai](\"https://raw.githubusercontent.com/mdai/ml-lessons/master/images/kaggle-project.png\")
For further data exploration and to create your own additional annotations, clone the MD.ai project at https://public.md.ai/annotator/project/LxR6zdR2.
\n\nYou can then create your own team, add new labels and additional annotations.
\n\nThe “Users“ tab will allow you to create teams and assign exams to team members. You can track progress and export your new annotations in JSON format.
\n\nFurther instructions and videos are available at https://docs.md.ai.
\n\nGood luck, everyone!
\n\n
![\"Clone](\"https://raw.githubusercontent.com/mdai/ml-lessons/master/images/kaggle-project-clone-small.gif\")
For further data exploration and to create your own additional annotations, clone the MD.ai project at https://public.md.ai/annotator/project/LxR6zdR2.
\n\nYou can then create your own team, add new labels and additional annotations.
\n\nThe “Users“ tab will allow you to create teams and assign exams to team members. You can track progress and export your new annotations in JSON format.
\n\nFurther instructions and videos are available at https://docs.md.ai.
\n\nGood luck, everyone!
\n\n
Very cool and useful feature Leon ! It is working very well. Thanks !
\n\nIt could be useful and easy to draw bounding boxes to train a detection model to find the boundaries of lungs. This model could then be used to automatically crop the image on lungs to optimize pneumonia detection.
\n\n![\"Bounding](\"https://preview.ibb.co/gZ7ETU/lung_crop.png\")
After you clone the MD.ai project at https://public.md.ai/annotator/project/LxR6zdR2 and add new labels and additional annotations, you can use the MD.ai python client library to download and export these additional labels and annotations to the competition labels format (i.e., stage_1_train_labels.csv) to use in training.
\n\nSee the example notebook we created here: \nhttps://github.com/mdai/ml-lessons/blob/master/pneumonia-detection-clone-project-data.ipynb
", "rawMarkdown": "After you clone the MD.ai project at https://public.md.ai/annotator/project/LxR6zdR2 and add new labels and additional annotations, you can use the MD.ai python client library to download and export these additional labels and annotations to the competition labels format (i.e., stage_1_train_labels.csv) to use in training.\n\nSee the example notebook we created here: \nhttps://github.com/mdai/ml-lessons/blob/master/pneumonia-detection-clone-project-data.ipynb", "votes": 3 }, { "id": 377651, "postDate": "2018-08-29T14:43:00.280Z", "content": "Yes annotations on your own on the train set only is always allowed. Here it seems that some annotations already exists on the train dataset.
", "rawMarkdown": "Yes annotations on your own on the train set only is always allowed. Here it seems that some annotations already exists on the train dataset.", "replies": [ { "id": 377823, "postDate": "2018-08-29T19:53:09.753Z", "content": "Thanks! I always thought one cannot manually add annotations unless shared in public
", "rawMarkdown": "Thanks! I always thought one cannot manually add annotations unless shared in public" }, { "id": 377857, "postDate": "2018-08-29T21:09:53.757Z", "content": "As far as I know, sharing the annotations on train set is a tradition on Kaggle but not an obligation.
\n\nOn NOAA Fisheries Steller Sea Lion Population Count, 3rd place solution comprised some custom manual annotations which weren't shared.
", "rawMarkdown": "As far as I know, sharing the annotations on train set is a tradition on Kaggle but not an obligation.\n\nOn NOAA Fisheries Steller Sea Lion Population Count, 3rd place solution comprised some custom manual annotations which weren't shared. " } ] }, { "id": 377649, "postDate": "2018-08-29T14:40:37.480Z", "content": "Is this allowed in a kaggle competition?
", "rawMarkdown": "Is this allowed in a kaggle competition?\n" }, { "id": 377639, "postDate": "2018-08-29T14:17:37.963Z", "content": "@Leon,\nCan we have access to the annotations in your previous link: https://public.md.ai/annotator/project/aGq4k6NW/workspace ?
", "rawMarkdown": "@Leon,\nCan we have access to the annotations in your previous link: https://public.md.ai/annotator/project/aGq4k6NW/workspace ?" }, { "id": 377677, "postDate": "2018-08-29T15:20:31.893Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 377397, "author_name": "Alexandre Cadrin-Chênevert", "author_url": "", "post_date": "2018-08-29T05:00:24.760000", "content": "Very cool and useful feature Leon ! It is working very well. Thanks !
\n\nIt could be useful and easy to draw bounding boxes to train a detection model to find the boundaries of lungs. This model could then be used to automatically crop the image on lungs to optimize pneumonia detection.
\n\n
After you clone the MD.ai project at https://public.md.ai/annotator/project/LxR6zdR2 and add new labels and additional annotations, you can use the MD.ai python client library to download and export these additional labels and annotations to the competition labels format (i.e., stage_1_train_labels.csv) to use in training.
\n\nSee the example notebook we created here: \nhttps://github.com/mdai/ml-lessons/blob/master/pneumonia-detection-clone-project-data.ipynb
", "votes": 3, "replies": [] }, { "id": 377651, "author_name": "eagle4", "author_url": "", "post_date": "2018-08-29T14:43:00.280000", "content": "Yes annotations on your own on the train set only is always allowed. Here it seems that some annotations already exists on the train dataset.
", "votes": 0, "replies": [ { "id": 377823, "author_name": "Wudi Wang", "author_url": "", "post_date": "2018-08-29T19:53:09.753000", "content": "Thanks! I always thought one cannot manually add annotations unless shared in public
", "votes": 0, "replies": [] }, { "id": 377857, "author_name": "eagle4", "author_url": "", "post_date": "2018-08-29T21:09:53.757000", "content": "As far as I know, sharing the annotations on train set is a tradition on Kaggle but not an obligation.
\n\nOn NOAA Fisheries Steller Sea Lion Population Count, 3rd place solution comprised some custom manual annotations which weren't shared.
", "votes": 0, "replies": [] } ] }, { "id": 377649, "author_name": "Wudi Wang", "author_url": "", "post_date": "2018-08-29T14:40:37.480000", "content": "Is this allowed in a kaggle competition?
", "votes": 0, "replies": [] }, { "id": 377639, "author_name": "eagle4", "author_url": "", "post_date": "2018-08-29T14:17:37.963000", "content": "@Leon,\nCan we have access to the annotations in your previous link: https://public.md.ai/annotator/project/aGq4k6NW/workspace ?
", "votes": 0, "replies": [] }, { "id": 377677, "author_name": "", "author_url": "", "post_date": "2018-08-29T15:20:31.893000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "376687": "For further data exploration and to create your own additional annotations, clone the MD.ai project at https://public.md.ai/annotator/project/LxR6zdR2. \n\nYou can then create your own team, add new labels and additional annotations. \n\nThe “Users“ tab will allow you to create teams and assign exams to team members. You can track progress and export your new annotations in JSON format. \n\nFurther instructions and videos are available at https://docs.md.ai.\n\nGood luck, everyone!\n\n![MD.ai Annotator Project][1]\n\n![Clone Project][2]\n\n\n [1]: https://raw.githubusercontent.com/mdai/ml-lessons/master/images/kaggle-project.png\n [2]: https://raw.githubusercontent.com/mdai/ml-lessons/master/images/kaggle-project-clone-small.gif", "377397": "Very cool and useful feature Leon ! It is working very well. Thanks !\n\nIt could be useful and easy to draw bounding boxes to train a detection model to find the boundaries of lungs. This model could then be used to automatically crop the image on lungs to optimize pneumonia detection.\n\n![Bounding box crop on lung volume][1]\n\n [1]: https://preview.ibb.co/gZ7ETU/lung_crop.png", "377798": "After you clone the MD.ai project at https://public.md.ai/annotator/project/LxR6zdR2 and add new labels and additional annotations, you can use the MD.ai python client library to download and export these additional labels and annotations to the competition labels format (i.e., stage_1_train_labels.csv) to use in training.\n\nSee the example notebook we created here: \nhttps://github.com/mdai/ml-lessons/blob/master/pneumonia-detection-clone-project-data.ipynb", "377651": "Yes annotations on your own on the train set only is always allowed. Here it seems that some annotations already exists on the train dataset.", "377649": "Is this allowed in a kaggle competition?\n", "377639": "@Leon,\nCan we have access to the annotations in your previous link: https://public.md.ai/annotator/project/aGq4k6NW/workspace ?", "377677": "" } }