{ "id": 197542, "title": "RLE to mask converter", "url": "/competitions/hubmap-kidney-segmentation/discussion/197542", "author_name": "", "post_date": "2020-11-17T00:34:55.100870300Z", "votes": 23, "comment_count": 7, "views": 0, "content": "

Hello Kagglers,

\n

lets start with converting the images from RLE string.

\n

Run-length encoding (RLE) is a form of lossless data compression in which runs of data (sequences in which the same data value occurs in many consecutive data elements) are stored as a single data value and count, rather than as the original run.
\n\"\"

\n

A sample method of converting the RLE to mask
\nhttps://www.kaggle.com/robertkag/rle-to-mask-converter

", "messages": [ { "id": "1081280", "postDate": "11/17/2020 00:34:55", "content": "

Hello Kagglers,

\n

lets start with converting the images from RLE string.

\n

Run-length encoding (RLE) is a form of lossless data compression in which runs of data (sequences in which the same data value occurs in many consecutive data elements) are stored as a single data value and count, rather than as the original run.
\n\"\"

\n

A sample method of converting the RLE to mask
\nhttps://www.kaggle.com/robertkag/rle-to-mask-converter

", "rawMarkdown": "Hello Kagglers,\n\nlets start with converting the images from RLE string.\n\n**Run-length encoding (RLE)** is a form of lossless data compression in which runs of data (sequences in which the same data value occurs in many consecutive data elements) are stored as a single data value and count, rather than as the original run.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F232304%2F34b5170604cdfcfc6cf5b22c45cd157c%2Frle.jpg?generation=1605573268174582&alt=media)\n\nA sample method of converting the RLE to mask \nhttps://www.kaggle.com/robertkag/rle-to-mask-converter", "votes": null }, { "id": "1083386", "postDate": "11/18/2020 23:16:50", "content": "

Thanks for posting. I just tried to encode a huge mask image (35000+, 40000+) with methods used in previous competitions but have a OOM exception! We may need to create a new procedure that is more memory efficient.

", "rawMarkdown": "Thanks for posting. I just tried to encode a huge mask image (35000+, 40000+) with methods used in previous competitions but have a OOM exception! We may need to create a new procedure that is more memory efficient.", "votes": null }, { "id": "1083854", "postDate": "11/19/2020 13:17:53", "content": "

I'll reply myself because I was able to fix the OOM by skipping one step on RLE function. It's shared here: https://www.kaggle.com/bguberfain/memory-aware-rle-encoding

", "rawMarkdown": "I'll reply myself because I was able to fix the OOM by skipping one step on RLE function. It's shared here: https://www.kaggle.com/bguberfain/memory-aware-rle-encoding", "votes": null }, { "id": "1086758", "postDate": "11/22/2020 03:05:29", "content": "

I've applied the function, and had this target.
\nit seems fine, thank you for correct code!

", "rawMarkdown": "I've applied the function, and had this target.\nit seems fine, thank you for correct code!", "votes": null }, { "id": "1087000", "postDate": "11/22/2020 09:04:43", "content": "

@subbuvolvosekar Have you made inverse function of this?
\n(I mean mask to RLE)

", "rawMarkdown": "subbuvolvosekar Have you made inverse function of this?\n(I mean mask to RLE)", "votes": null }, { "id": "1089372", "postDate": "11/24/2020 12:56:22", "content": "

No.. But going to try.. you made a curious question to start :)

", "rawMarkdown": "No.. But going to try.. you made a curious question to start :)", "votes": null }, { "id": "1155733", "postDate": "01/16/2021 16:14:49", "content": "

I'm trying the code in https://www.kaggle.com/robertkag/rle-to-mask-converter, and have a question about the results:

\n

I start with a very simple 3 x 3 \"mask\"

\n
[ [ 0, 1, 0 ],\n  [ 1, 0, 1 ],\n  [ 0, 1, 0 ]  ]\n
\n

I manually encode this into the string 1 1 3 1 5 1 8 1. However, when I pass this to the code, I get

\n
[[255   0   0]\n [  0 255 255]\n [255   0   0]]\n
\n

Doesn't seem to agree with the original mask. Is my manual encoding incorrect, or is the decoder?
\nThanks,
\n-- Mark

", "rawMarkdown": "I'm trying the code in https://www.kaggle.com/robertkag/rle-to-mask-converter, and have a question about the results:\n\nI start with a very simple 3 x 3 \"mask\"\n```\n[ [ 0, 1, 0 ],\n [ 1, 0, 1 ],\n [ 0, 1, 0 ] ]\n```\nI manually encode this into the string ```1 1 3 1 5 1 8 1```. However, when I pass this to the code, I get \n```\n[[255 0 0]\n [ 0 255 255]\n [255 0 0]]\n```\nDoesn't seem to agree with the original mask. Is my manual encoding incorrect, or is the decoder?\nThanks,\n-- Mark", "votes": null }, { "id": "1288698", "postDate": "04/30/2021 09:04:46", "content": "

Try string '2 1 4 1 6 1 8 1'

", "rawMarkdown": "Try string '2 1 4 1 6 1 8 1'", "votes": null } ], "comments": [ { "id": 1083386, "author_name": "bguberfain", "author_url": "", "post_date": "11/18/2020 23:16:50", "content": "

Thanks for posting. I just tried to encode a huge mask image (35000+, 40000+) with methods used in previous competitions but have a OOM exception! We may need to create a new procedure that is more memory efficient.

", "votes": null, "replies": [ { "id": 1083854, "author_name": "bguberfain", "author_url": "", "post_date": "11/19/2020 13:17:53", "content": "

I'll reply myself because I was able to fix the OOM by skipping one step on RLE function. It's shared here: https://www.kaggle.com/bguberfain/memory-aware-rle-encoding

", "votes": null, "replies": [] } ] }, { "id": 1086758, "author_name": "ryunosukeishizaki", "author_url": "", "post_date": "11/22/2020 03:05:29", "content": "

I've applied the function, and had this target.
\nit seems fine, thank you for correct code!

", "votes": null, "replies": [ { "id": 1087000, "author_name": "ryunosukeishizaki", "author_url": "", "post_date": "11/22/2020 09:04:43", "content": "

@subbuvolvosekar Have you made inverse function of this?
\n(I mean mask to RLE)

", "votes": null, "replies": [] }, { "id": 1089372, "author_name": "subbuvolvosekar", "author_url": "", "post_date": "11/24/2020 12:56:22", "content": "

No.. But going to try.. you made a curious question to start :)

", "votes": null, "replies": [] } ] }, { "id": 1155733, "author_name": "markalavin", "author_url": "", "post_date": "01/16/2021 16:14:49", "content": "

I'm trying the code in https://www.kaggle.com/robertkag/rle-to-mask-converter, and have a question about the results:

\n

I start with a very simple 3 x 3 \"mask\"

\n
[ [ 0, 1, 0 ],\n  [ 1, 0, 1 ],\n  [ 0, 1, 0 ]  ]\n
\n

I manually encode this into the string 1 1 3 1 5 1 8 1. However, when I pass this to the code, I get

\n
[[255   0   0]\n [  0 255 255]\n [255   0   0]]\n
\n

Doesn't seem to agree with the original mask. Is my manual encoding incorrect, or is the decoder?
\nThanks,
\n-- Mark

", "votes": null, "replies": [ { "id": 1288698, "author_name": "mgcyung", "author_url": "", "post_date": "04/30/2021 09:04:46", "content": "

Try string '2 1 4 1 6 1 8 1'

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1081280": "Hello Kagglers,\n\nlets start with converting the images from RLE string.\n\n**Run-length encoding (RLE)** is a form of lossless data compression in which runs of data (sequences in which the same data value occurs in many consecutive data elements) are stored as a single data value and count, rather than as the original run.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F232304%2F34b5170604cdfcfc6cf5b22c45cd157c%2Frle.jpg?generation=1605573268174582&alt=media)\n\nA sample method of converting the RLE to mask \nhttps://www.kaggle.com/robertkag/rle-to-mask-converter", "1083386": "Thanks for posting. I just tried to encode a huge mask image (35000+, 40000+) with methods used in previous competitions but have a OOM exception! We may need to create a new procedure that is more memory efficient.", "1083854": "I'll reply myself because I was able to fix the OOM by skipping one step on RLE function. It's shared here: https://www.kaggle.com/bguberfain/memory-aware-rle-encoding", "1086758": "I've applied the function, and had this target.\nit seems fine, thank you for correct code!", "1087000": "subbuvolvosekar Have you made inverse function of this?\n(I mean mask to RLE)", "1089372": "No.. But going to try.. you made a curious question to start :)", "1155733": "I'm trying the code in https://www.kaggle.com/robertkag/rle-to-mask-converter, and have a question about the results:\n\nI start with a very simple 3 x 3 \"mask\"\n```\n[ [ 0, 1, 0 ],\n [ 1, 0, 1 ],\n [ 0, 1, 0 ] ]\n```\nI manually encode this into the string ```1 1 3 1 5 1 8 1```. However, when I pass this to the code, I get \n```\n[[255 0 0]\n [ 0 255 255]\n [255 0 0]]\n```\nDoesn't seem to agree with the original mask. Is my manual encoding incorrect, or is the decoder?\nThanks,\n-- Mark", "1288698": "Try string '2 1 4 1 6 1 8 1'" }, "source": "meta" }