I took the dcm files for segmentation and by taking the coordinates, I created boxes of 10 units in size and labeled them. However, the sizes of these boxes are the same, wouldn't this cause the model to be trained incorrectly?What is the solution to this?
", "messages": [ { "id": "2967210", "postDate": "08/22/2024 16:21:49", "content": "I took the dcm files for segmentation and by taking the coordinates, I created boxes of 10 units in size and labeled them. However, the sizes of these boxes are the same, wouldn't this cause the model to be trained incorrectly?What is the solution to this?
", "rawMarkdown": "I took the dcm files for segmentation and by taking the coordinates, I created boxes of 10 units in size and labeled them. However, the sizes of these boxes are the same, wouldn't this cause the model to be trained incorrectly?What is the solution to this?", "votes": null }, { "id": "2967280", "postDate": "08/22/2024 17:49:07", "content": "It's a valid approach. A part from tune size I think you can't do much more about it. Just try. If it doesn't work you'll find other options on section code.
\nI would recommend:
\n\n\n\n", "rawMarkdown": "It's a valid approach. A part from tune size I think you can't do much more about it. Just try. If it doesn't work you'll find other options on section code.\n\nI would recommend:\n\n[from Bartley](https://www.kaggle.com/code/brendanartley/lumbar-coordinate-dataset-code)\n\n[from Guilherme Levi]( https://www.kaggle.com/code/guilhermelevi/rsna-lsdc-disc-detection-baseline-faster-r-cnn/notebook)\n\n[from Anouk Stein](https://www.kaggle.com/code/anoukstein/2d-segmentation-of-sagittal-lumbar-spine-mri)\n\n[Mine](https://www.kaggle.com/code/sacuscreed/foraminal-unet-vit-train)", "votes": null }, { "id": "2970407", "postDate": "08/26/2024 05:13:46", "content": "Hey! I wanted to confirm one thing
\nFor the testing dataset, since the coordinates (x, y) required to segment L1-L2 are not provided, do we need to train our model to detect these coordinates as well?
\nThanks
If you plane to explicitely use them yes.
", "rawMarkdown": "If you plane to explicitely use them yes.", "votes": null } ], "comments": [ { "id": 2967280, "author_name": "sacuscreed", "author_url": "", "post_date": "08/22/2024 17:49:07", "content": "It's a valid approach. A part from tune size I think you can't do much more about it. Just try. If it doesn't work you'll find other options on section code.
\nI would recommend:
\n\n\n\n", "votes": null, "replies": [] }, { "id": 2970407, "author_name": "vparh7", "author_url": "", "post_date": "08/26/2024 05:13:46", "content": "Hey! I wanted to confirm one thing
\nFor the testing dataset, since the coordinates (x, y) required to segment L1-L2 are not provided, do we need to train our model to detect these coordinates as well?
\nThanks
If you plane to explicitely use them yes.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2967210": "I took the dcm files for segmentation and by taking the coordinates, I created boxes of 10 units in size and labeled them. However, the sizes of these boxes are the same, wouldn't this cause the model to be trained incorrectly?What is the solution to this?", "2967280": "It's a valid approach. A part from tune size I think you can't do much more about it. Just try. If it doesn't work you'll find other options on section code.\n\nI would recommend:\n\n[from Bartley](https://www.kaggle.com/code/brendanartley/lumbar-coordinate-dataset-code)\n\n[from Guilherme Levi]( https://www.kaggle.com/code/guilhermelevi/rsna-lsdc-disc-detection-baseline-faster-r-cnn/notebook)\n\n[from Anouk Stein](https://www.kaggle.com/code/anoukstein/2d-segmentation-of-sagittal-lumbar-spine-mri)\n\n[Mine](https://www.kaggle.com/code/sacuscreed/foraminal-unet-vit-train)", "2970407": "Hey! I wanted to confirm one thing\nFor the testing dataset, since the coordinates (x, y) required to segment L1-L2 are not provided, do we need to train our model to detect these coordinates as well?\nThanks", "2970459": "If you plane to explicitely use them yes." }, "source": "meta" }