Hi Everyone,
\nI understand the problem statement requires us to create a model to detect vasculatures in Human Tissue.
\nNeeded some help to understand what do we need to submit as a result, It is my understanding the output should contain some data in some specific format and I need someone's help to understand that.
PS: Relatively new to ML, AI stuff.
\nThanks.
", "messages": [ { "id": 2273766, "postDate": "2023-05-25T11:52:53.843Z", "content": "Basically you need to construct a model that is able to segment (highlight or produce a binary mask where 1 is the microvascular instance in the tissue and 0 is the rest).
\nThen you need to transform this mask to polygons (since each region can be en-globed into some sort of polygon)
\nThe coordinates of these polygons for each mask must be encoded in a certain way defined here: Evaluation
\nTo explain it further, the submission file should be of the format:
,height,width,prediction_string\n72e40acccadf,,, eNoLTDAwyrM3yI/PMwcAE94DZA==\nwhere prediction_string has the format 0 {confidence} {EncodedPolygon}
In case your mask has multiple polygons i.e multiple instances are detected in one image then the prediction_string would be:
0 {confidence_1} {EncodedPolygon_1} 0 {confidence_2} {EncodedPolygon_2} 0 {confidence_3} {EncodedPolygon_3}
each one starts with 0 and they are seperated by a space ' '
\nThe EncodedPolygon is obtained using the following steps:
The binary segmentation masks are run-length encoded (RLE), zlib compressed, and base64 encoded to be used in text format as EncodedMask. Specifically, we use the Coco masks RLE encoding/decoding (see the encode method of COCO’s mask API), the zlib compression/decompression (RFC1950), and vanilla base64 encoding.
\nThis is an example function that they gave also here : Evaluation
\n base64\n numpy np\n pycocotools _mask coco_mask\n typing t\n zlib\n\n\n () -> t.Text:\n \n\n \n mask.dtype != np.:\n ValueError(\n %\n mask.dtype)\n\n mask = np.squeeze(mask)\n (mask.shape) != :\n ValueError(\n %\n mask.shape)\n\n \n mask_to_encode = mask.reshape(mask.shape[], mask.shape[], )\n mask_to_encode = mask_to_encode.astype(np.uint8)\n mask_to_encode = np.asfortranarray(mask_to_encode)\n\n \n encoded_mask = coco_mask.encode(mask_to_encode)[][]\n\n \n binary_str = zlib.compress(encoded_mask, zlib.Z_BEST_COMPRESSION)\n base64_str = base64.b64encode(binary_str)\n base64_str\nThis explanation is great. Appreciate you taking the time to post this thanks :)
", "rawMarkdown": "This explanation is great. Appreciate you taking the time to post this thanks :)\n", "votes": 1 }, { "id": 2285994, "postDate": "2023-06-03T06:37:06.977Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 2273751, "postDate": "2023-05-25T11:38:58.833Z", "content": "Hi Everyone,
\nI understand the problem statement requires us to create a model to detect vasculatures in Human Tissue.
\nNeeded some help to understand what do we need to submit as a result, It is my understanding the output should contain some data in some specific format and I need someone's help to understand that.
PS: Relatively new to ML, AI stuff.
\nThanks.
", "rawMarkdown": "Hi Everyone, \nI understand the problem statement requires us to create a model to detect vasculatures in Human Tissue. \nNeeded some help to understand what do we need to submit as a result, It is my understanding the output should contain some data in some specific format and I need someone's help to understand that.\n\nPS: Relatively new to ML, AI stuff. \n\nThanks. \n", "votes": 8 } ], "comments": [ { "id": 2273766, "author_name": "Mohamed Amine DHIAB", "author_url": "", "post_date": "2023-05-25T11:52:53.843000", "content": "Basically you need to construct a model that is able to segment (highlight or produce a binary mask where 1 is the microvascular instance in the tissue and 0 is the rest).
\nThen you need to transform this mask to polygons (since each region can be en-globed into some sort of polygon)
\nThe coordinates of these polygons for each mask must be encoded in a certain way defined here: Evaluation
\nTo explain it further, the submission file should be of the format:
,height,width,prediction_string\n72e40acccadf,,, eNoLTDAwyrM3yI/PMwcAE94DZA==\nwhere prediction_string has the format 0 {confidence} {EncodedPolygon}
In case your mask has multiple polygons i.e multiple instances are detected in one image then the prediction_string would be:
0 {confidence_1} {EncodedPolygon_1} 0 {confidence_2} {EncodedPolygon_2} 0 {confidence_3} {EncodedPolygon_3}
each one starts with 0 and they are seperated by a space ' '
\nThe EncodedPolygon is obtained using the following steps:
The binary segmentation masks are run-length encoded (RLE), zlib compressed, and base64 encoded to be used in text format as EncodedMask. Specifically, we use the Coco masks RLE encoding/decoding (see the encode method of COCO’s mask API), the zlib compression/decompression (RFC1950), and vanilla base64 encoding.
\nThis is an example function that they gave also here : Evaluation
\n base64\n numpy np\n pycocotools _mask coco_mask\n typing t\n zlib\n\n\n () -> t.Text:\n \n\n \n mask.dtype != np.:\n ValueError(\n %\n mask.dtype)\n\n mask = np.squeeze(mask)\n (mask.shape) != :\n ValueError(\n %\n mask.shape)\n\n \n mask_to_encode = mask.reshape(mask.shape[], mask.shape[], )\n mask_to_encode = mask_to_encode.astype(np.uint8)\n mask_to_encode = np.asfortranarray(mask_to_encode)\n\n \n encoded_mask = coco_mask.encode(mask_to_encode)[][]\n\n \n binary_str = zlib.compress(encoded_mask, zlib.Z_BEST_COMPRESSION)\n base64_str = base64.b64encode(binary_str)\n base64_str\nThis explanation is great. Appreciate you taking the time to post this thanks :)
", "votes": 1, "replies": [] }, { "id": 2285994, "author_name": "", "author_url": "", "post_date": "2023-06-03T06:37:06.977000", "content": "", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "2273766": "Basically you need to construct a model that is able to segment (highlight or produce a binary mask where 1 is the microvascular instance in the tissue and 0 is the rest).\nThen you need to transform this mask to polygons (since each region can be en-globed into some sort of polygon)\nThe coordinates of these polygons for each mask must be encoded in a certain way defined here: [Evaluation](https://www.kaggle.com/competitions/hubmap-hacking-the-human-vasculature/overview/evaluation)\nTo explain it further, the submission file should be of the format:\n```python\nid,height,width,prediction_string\n72e40acccadf,512,512,0 1.0 eNoLTDAwyrM3yI/PMwcAE94DZA==\n```\nwhere `prediction_string` has the format `0 {confidence} {EncodedPolygon}`\n\nIn case your mask has multiple polygons i.e multiple instances are detected in one image then the `prediction_string` would be:\n\n`0 {confidence_1} {EncodedPolygon_1} 0 {confidence_2} {EncodedPolygon_2} 0 {confidence_3} {EncodedPolygon_3}`\n\neach one starts with 0 and they are seperated by a space ' '\n\nThe `EncodedPolygon` is obtained using the following steps:\n\nThe binary segmentation masks are run-length encoded (RLE), zlib compressed, and base64 encoded to be used in text format as EncodedMask. Specifically, we use the Coco masks RLE encoding/decoding (see the encode method of COCO’s mask API), the zlib compression/decompression (RFC1950), and vanilla base64 encoding.\n\nThis is an example function that they gave also here : [Evaluation](https://www.kaggle.com/competitions/hubmap-hacking-the-human-vasculature/overview/evaluation)\n\n```python\nimport base64\nimport numpy as np\nfrom pycocotools import _mask as coco_mask\nimport typing as t\nimport zlib\n\n\ndef encode_binary_mask(mask: np.ndarray) -> t.Text:\n \"\"\"Converts a binary mask into OID challenge encoding ascii text.\"\"\"\n\n # check input mask --\n if mask.dtype != np.bool:\n raise ValueError(\n \"encode_binary_mask expects a binary mask, received dtype == %s\" %\n mask.dtype)\n\n mask = np.squeeze(mask)\n if len(mask.shape) != 2:\n raise ValueError(\n \"encode_binary_mask expects a 2d mask, received shape == %s\" %\n mask.shape)\n\n # convert input mask to expected COCO API input --\n mask_to_encode = mask.reshape(mask.shape[0], mask.shape[1], 1)\n mask_to_encode = mask_to_encode.astype(np.uint8)\n mask_to_encode = np.asfortranarray(mask_to_encode)\n\n # RLE encode mask --\n encoded_mask = coco_mask.encode(mask_to_encode)[0][\"counts\"]\n\n # compress and base64 encoding --\n binary_str = zlib.compress(encoded_mask, zlib.Z_BEST_COMPRESSION)\n base64_str = base64.b64encode(binary_str)\n return base64_str\n\n```", "2273751": "Hi Everyone, \nI understand the problem statement requires us to create a model to detect vasculatures in Human Tissue. \nNeeded some help to understand what do we need to submit as a result, It is my understanding the output should contain some data in some specific format and I need someone's help to understand that.\n\nPS: Relatively new to ML, AI stuff. \n\nThanks. \n" } }