{
  "id": 32857,
  "title": "Correct Sea Lion Coordinates and Counts",
  "url": "/competitions/noaa-fisheries-steller-sea-lion-population-count/discussion/32857",
  "author_name": "",
  "post_date": "2017-05-12T01:09:18.021082900Z",
  "votes": 32,
  "comment_count": 14,
  "views": 0,
  "content": "<p><strong>Hello everyone!</strong></p>\n\n<p>I have correct coordinates available for download <strong><a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_coordinates.csv\">here</a></strong>, and correct counts available for download <strong><a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\">here</a></strong>. If you are curious about how these are generated I have an IPython Notebook that explains everything and can be run to generate the same results <strong><a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/Correct%20Coordinates%20%26%20Sea%20Lion%20Counts.ipynb\">here</a></strong>. </p>\n\n<p><strong>Briefly Summarizing What I Did</strong>:</p>\n\n<ul>\n<li>Use <strong><a href=\"https://www.kaggle.com/radustoicescu/get-coordinates-using-blob-detection\">Radu's Method</a></strong> to extract preliminary coordinates</li>\n<li>Generate a count for each image based on number of preliminary coordinates for each class</li>\n<li>Compare that count to the <strong>train.csv</strong> counts</li>\n<li>Draw coordinates for all the images that have count discrepancies</li>\n<li>Manually look over the images with discrepancies (over 252 total)</li>\n<li>Add any missing coordinates, remove any misplaced coordinates</li>\n<li>Generate correct count based on final number of coordinates</li>\n</ul>\n\n<p><strong>Other Issues Noted, But Not Addressed</strong>:</p>\n\n<ul>\n<li>448 ~ 3 unmarked sea lions in the rocky islands on the middle</li>\n<li>513 ~ darker unmarked sea lions</li>\n<li>519 ~ 2 unmarked sea lions in the bottom right corner</li>\n<li>555 ~ many unmarked sea lions in the water</li>\n<li>571 ~ unmarked sea lion at the center of the island</li>\n<li>578 ~ unmarked sea lion towards in the bottom left quadrant</li>\n<li>629 ~ unmarked sea lion in the ocean</li>\n<li>655 ~ darker unmarked sea lions</li>\n<li>675 ~ unmarked sea lion top right quadrant</li>\n<li>679 ~ unmarked sea lion in the water near the coast</li>\n<li>686 ~ unmarked sea lion in the water near the coast</li>\n<li>724 ~ 3 unmarked sea lions in the water, 1 unmarked sea lion on tiny island</li>\n<li>732 ~ unmarked sea lion in the water in the bottom left quadrant</li>\n<li>759 ~ sea lion marked DEAD, meaning is not understood</li>\n<li>761 ~ possibly a few adult female sea lions marked as adult male</li>\n<li>764 ~ 4 unmarked sea lions</li>\n<li>776 ~ unmarked black sea lions</li>\n<li>790 ~ unmarked black sea lions</li>\n<li>802 ~ unmarked sea lion in bottom right quadrant</li>\n<li>806 ~ 2 unmarked sea lions that had just dove into the water </li>\n<li>822 ~ unmarked sea lion in the upper right quadrant, in the water</li>\n<li>823 ~ unmarked sea lion in the water</li>\n<li>837 ~ unmarked sea lions in bottom right quadrant, text in the top left quadrant (\"Bulls: 1 SAMs:0 Females:0 Juveniles:0 Pups:0 Total:1\")</li>\n<li>845 ~ possibly a few adult female sea lions marked as adult male</li>\n<li>847 ~ unmarked sea lion bottom right quadrant</li>\n<li>857 ~ no marked sea lions present</li>\n<li>858 ~ unmarked sea lion in bottom right quadrant ocean, unmarked sea lion to left of island in the ocean</li>\n<li>899 ~ unmarked sea lion top left quadrant</li>\n<li>900 ~ unmarked sea lion bottom left quadrant</li>\n<li>906 ~ unmarked sea lion in water off the coast</li>\n<li>924 ~ shiny unmarked sea lions (most likely not stellar sea lions)</li>\n<li>929 ~ unmarked sea lion in water (top right quadrant)</li>\n<li>939 ~ 2 unmarked sea lions in cluster of sea lions</li>\n</ul>\n\n<p><strong>Good Luck, Hope This Helps!</strong></p>\n\n<p><strong>If you have questions, I'll gladly answer them down below</strong></p>\n\n<h2>Change Log:</h2>\n\n<ul>\n<li><strong>5/16/2017</strong>, credit to <strong>@Lowik Chanussot</strong>, changes made to data for images 80, 112, 127, 170, 174, 186, 197, 291, 416, 418, 420, 452, 495, 566, 587, 629, 699, 715, 791, 828 (refer to indepth explanation of changes below, github commits are <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\">here</a> and <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665\">here</a>)</li>\n<li><strong>5/30/2017</strong>, credit to <strong>@Artem.Sanakoev</strong>, changes made to data for image 691 (refer to indepth explanation of changes below, github commit is <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5\">here</a>)</li>\n</ul>",
  "messages": [
    {
      "id": "182015",
      "postDate": "05/12/2017 01:09:18",
      "content": "<p><strong>Hello everyone!</strong></p>\n\n<p>I have correct coordinates available for download <strong><a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_coordinates.csv\">here</a></strong>, and correct counts available for download <strong><a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\">here</a></strong>. If you are curious about how these are generated I have an IPython Notebook that explains everything and can be run to generate the same results <strong><a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/Correct%20Coordinates%20%26%20Sea%20Lion%20Counts.ipynb\">here</a></strong>. </p>\n\n<p><strong>Briefly Summarizing What I Did</strong>:</p>\n\n<ul>\n<li>Use <strong><a href=\"https://www.kaggle.com/radustoicescu/get-coordinates-using-blob-detection\">Radu's Method</a></strong> to extract preliminary coordinates</li>\n<li>Generate a count for each image based on number of preliminary coordinates for each class</li>\n<li>Compare that count to the <strong>train.csv</strong> counts</li>\n<li>Draw coordinates for all the images that have count discrepancies</li>\n<li>Manually look over the images with discrepancies (over 252 total)</li>\n<li>Add any missing coordinates, remove any misplaced coordinates</li>\n<li>Generate correct count based on final number of coordinates</li>\n</ul>\n\n<p><strong>Other Issues Noted, But Not Addressed</strong>:</p>\n\n<ul>\n<li>448 ~ 3 unmarked sea lions in the rocky islands on the middle</li>\n<li>513 ~ darker unmarked sea lions</li>\n<li>519 ~ 2 unmarked sea lions in the bottom right corner</li>\n<li>555 ~ many unmarked sea lions in the water</li>\n<li>571 ~ unmarked sea lion at the center of the island</li>\n<li>578 ~ unmarked sea lion towards in the bottom left quadrant</li>\n<li>629 ~ unmarked sea lion in the ocean</li>\n<li>655 ~ darker unmarked sea lions</li>\n<li>675 ~ unmarked sea lion top right quadrant</li>\n<li>679 ~ unmarked sea lion in the water near the coast</li>\n<li>686 ~ unmarked sea lion in the water near the coast</li>\n<li>724 ~ 3 unmarked sea lions in the water, 1 unmarked sea lion on tiny island</li>\n<li>732 ~ unmarked sea lion in the water in the bottom left quadrant</li>\n<li>759 ~ sea lion marked DEAD, meaning is not understood</li>\n<li>761 ~ possibly a few adult female sea lions marked as adult male</li>\n<li>764 ~ 4 unmarked sea lions</li>\n<li>776 ~ unmarked black sea lions</li>\n<li>790 ~ unmarked black sea lions</li>\n<li>802 ~ unmarked sea lion in bottom right quadrant</li>\n<li>806 ~ 2 unmarked sea lions that had just dove into the water </li>\n<li>822 ~ unmarked sea lion in the upper right quadrant, in the water</li>\n<li>823 ~ unmarked sea lion in the water</li>\n<li>837 ~ unmarked sea lions in bottom right quadrant, text in the top left quadrant (\"Bulls: 1 SAMs:0 Females:0 Juveniles:0 Pups:0 Total:1\")</li>\n<li>845 ~ possibly a few adult female sea lions marked as adult male</li>\n<li>847 ~ unmarked sea lion bottom right quadrant</li>\n<li>857 ~ no marked sea lions present</li>\n<li>858 ~ unmarked sea lion in bottom right quadrant ocean, unmarked sea lion to left of island in the ocean</li>\n<li>899 ~ unmarked sea lion top left quadrant</li>\n<li>900 ~ unmarked sea lion bottom left quadrant</li>\n<li>906 ~ unmarked sea lion in water off the coast</li>\n<li>924 ~ shiny unmarked sea lions (most likely not stellar sea lions)</li>\n<li>929 ~ unmarked sea lion in water (top right quadrant)</li>\n<li>939 ~ 2 unmarked sea lions in cluster of sea lions</li>\n</ul>\n\n<p><strong>Good Luck, Hope This Helps!</strong></p>\n\n<p><strong>If you have questions, I'll gladly answer them down below</strong></p>\n\n<h2>Change Log:</h2>\n\n<ul>\n<li><strong>5/16/2017</strong>, credit to <strong>@Lowik Chanussot</strong>, changes made to data for images 80, 112, 127, 170, 174, 186, 197, 291, 416, 418, 420, 452, 495, 566, 587, 629, 699, 715, 791, 828 (refer to indepth explanation of changes below, github commits are <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\">here</a> and <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665\">here</a>)</li>\n<li><strong>5/30/2017</strong>, credit to <strong>@Artem.Sanakoev</strong>, changes made to data for image 691 (refer to indepth explanation of changes below, github commit is <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5\">here</a>)</li>\n</ul>",
      "rawMarkdown": "**Hello everyone!**\n\nI have correct coordinates available for download **[here][1]**, and correct counts available for download **[here][2]**. If you are curious about how these are generated I have an IPython Notebook that explains everything and can be run to generate the same results **[here][3]**. \n\n**Briefly Summarizing What I Did**:\n\n - Use **[Radu's Method][4]** to extract preliminary coordinates\n - Generate a count for each image based on number of preliminary coordinates for each class\n - Compare that count to the **train.csv** counts\n - Draw coordinates for all the images that have count discrepancies\n - Manually look over the images with discrepancies (over 252 total)\n - Add any missing coordinates, remove any misplaced coordinates\n - Generate correct count based on final number of coordinates\n\n**Other Issues Noted, But Not Addressed**:\n\n - 448 ~ 3 unmarked sea lions in the rocky islands on the middle\n - 513 ~ darker unmarked sea lions\n - 519 ~ 2 unmarked sea lions in the bottom right corner\n - 555 ~ many unmarked sea lions in the water\n - 571 ~ unmarked sea lion at the center of the island\n - 578 ~ unmarked sea lion towards in the bottom left quadrant\n - 629 ~ unmarked sea lion in the ocean\n - 655 ~ darker unmarked sea lions\n - 675 ~ unmarked sea lion top right quadrant\n - 679 ~ unmarked sea lion in the water near the coast\n - 686 ~ unmarked sea lion in the water near the coast\n - 724 ~ 3 unmarked sea lions in the water, 1 unmarked sea lion on tiny island\n - 732 ~ unmarked sea lion in the water in the bottom left quadrant\n - 759 ~ sea lion marked DEAD, meaning is not understood\n - 761 ~ possibly a few adult female sea lions marked as adult male\n - 764 ~ 4 unmarked sea lions\n - 776 ~ unmarked black sea lions\n - 790 ~ unmarked black sea lions\n - 802 ~ unmarked sea lion in bottom right quadrant\n - 806 ~ 2 unmarked sea lions that had just dove into the water \n - 822 ~ unmarked sea lion in the upper right quadrant, in the water\n - 823 ~ unmarked sea lion in the water\n - 837 ~ unmarked sea lions in bottom right quadrant, text in the top left quadrant (\"Bulls: 1 SAMs:0 Females:0 Juveniles:0 Pups:0 Total:1\")\n - 845 ~ possibly a few adult female sea lions marked as adult male\n - 847 ~ unmarked sea lion bottom right quadrant\n - 857 ~ no marked sea lions present\n - 858 ~ unmarked sea lion in bottom right quadrant ocean, unmarked sea lion to left of island in the ocean\n - 899 ~ unmarked sea lion top left quadrant\n - 900 ~ unmarked sea lion bottom left quadrant\n - 906 ~ unmarked sea lion in water off the coast\n - 924 ~ shiny unmarked sea lions (most likely not stellar sea lions)\n - 929 ~ unmarked sea lion in water (top right quadrant)\n - 939 ~ 2 unmarked sea lions in cluster of sea lions\n\n**Good Luck, Hope This Helps!**\n\n**If you have questions, I'll gladly answer them down below**\n\n##Change Log: ##\n* **5/16/2017**, credit to **@Lowik Chanussot**, changes made to data for images 80, 112, 127, 170, 174, 186, 197, 291, 416, 418, 420, 452, 495, 566, 587, 629, 699, 715, 791, 828 (refer to indepth explanation of changes below, github commits are [here][5] and [here][6])\n* **5/30/2017**, credit to **@Artem.Sanakoev**, changes made to data for image 691 (refer to indepth explanation of changes below, github commit is [here][7])\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_coordinates.csv\n  [2]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\n  [3]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/Correct%20Coordinates%20%26%20Sea%20Lion%20Counts.ipynb\n  [4]: https://www.kaggle.com/radustoicescu/get-coordinates-using-blob-detection\n  [5]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\n  [6]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665\n  [7]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5",
      "votes": null
    },
    {
      "id": "182027",
      "postDate": "05/12/2017 04:12:23",
      "content": "<p>Looks like you've put a lot of hard work into this -- congrats on this milestone, and moreover, thanks for sharing!</p>\n\n<p>I'm a little concerned that we haven't had any feedback from the organizers regarding the accuracy + validity of the training dataset. Over the past month, there have been many kernels (almost every public kernel so far in fact), which has attempted to address the problem of training data discrepancies. As we're closing in on just a month and a half left, I hope the organizers agree to do something such as use the same pipeline outlined above by @LivingProgram to: 1) formally accept his updated training coordinates as G.T., and 2) apply the same process to clean up the unreleased stage2 data, if need be.</p>",
      "rawMarkdown": "Looks like you've put a lot of hard work into this -- congrats on this milestone, and moreover, thanks for sharing!\n\nI'm a little concerned that we haven't had any feedback from the organizers regarding the accuracy + validity of the training dataset. Over the past month, there have been many kernels (almost every public kernel so far in fact), which has attempted to address the problem of training data discrepancies. As we're closing in on just a month and a half left, I hope the organizers agree to do something such as use the same pipeline outlined above by @LivingProgram to: 1) formally accept his updated training coordinates as G.T., and 2) apply the same process to clean up the unreleased stage2 data, if need be.",
      "votes": null
    },
    {
      "id": "182087",
      "postDate": "05/12/2017 10:31:43",
      "content": "<p>Thanks @authman , \nI strongly agree with you about verifying the validity of the training dataset, and just to add on, those unmarked sea lions I mentioned in the <strong>\"Other Issues Noted, But Not Addressed\"</strong> could cause some discrepancies both in the training stage as well as determining the final results if there is the same problem in the final Leaderboard Data. So hopefully those issues can be addressed as well by organizers. :)</p>",
      "rawMarkdown": "Thanks @authman , \nI strongly agree with you about verifying the validity of the training dataset, and just to add on, those unmarked sea lions I mentioned in the **\"Other Issues Noted, But Not Addressed\"** could cause some discrepancies both in the training stage as well as determining the final results if there is the same problem in the final Leaderboard Data. So hopefully those issues can be addressed as well by organizers. :)",
      "votes": null
    },
    {
      "id": "182415",
      "postDate": "05/13/2017 18:24:20",
      "content": "<p>Hi, I see that you did a nice job and worked a lot. Thanks,</p>\n\n<p>Is it legal to use your data?</p>",
      "rawMarkdown": "Hi, I see that you did a nice job and worked a lot. Thanks,\n\nIs it legal to use your data?",
      "votes": null
    },
    {
      "id": "182460",
      "postDate": "05/14/2017 00:12:21",
      "content": "<p>Hi @armamut,\nI <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/LICENSE.txt\">open sourced my data</a>, so I think it is legal as everyone in the competition can access it (so there should be no problem with the <strong>\"No private sharing outside teams\"</strong> rule). I am also not distributing the original data, so it should comply with the competition rules. If need be we could try to get some clarification from competition organizers.</p>",
      "rawMarkdown": "Hi @armamut,\nI [open sourced my data][1], so I think it is legal as everyone in the competition can access it (so there should be no problem with the **\"No private sharing outside teams\"** rule). I am also not distributing the original data, so it should comply with the competition rules. If need be we could try to get some clarification from competition organizers.\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/LICENSE.txt",
      "votes": null
    },
    {
      "id": "182547",
      "postDate": "05/14/2017 14:01:28",
      "content": "<p>Thank you for your awesome work! That's really helpful.</p>\n\n<p>I checked discrepancies between your coordinates and mines, I found a few errors that you may want to correct ((x, y) format):</p>\n\n<ul>\n<li>80: (1697, 414) should be subadult_males instead of adult_females (it will gives the right numbers in train.csv)</li>\n<li>112: Replace (1362, 1156) (adult_females) by (1365, 1153) (position of the dot incorrectly detected)</li>\n<li>127: add (3014, 3053) in adult_females</li>\n<li>170: (3234, 623) should be adult_females instead of adult_males (to match counts if train.csv)</li>\n<li>174: remove adult_male: (3710, 3500), this is a false detection</li>\n<li>186: add (2335, 1273) and (3177, 1408) in adult_females (missed detections)</li>\n<li>197: add (2959, 1042) in adult_females</li>\n<li>291: remove 3 wrong detections in adult_males: (1452, 49), (1482, 49), (1503, 49)</li>\n<li>416: add  (1385, 543) in juveniles</li>\n<li>418: add (1622, 1032), (1715, 1151), (1510, 1229) and (1648, 1109) in adult_females</li>\n<li>420: add (4733, 437) in adult_females</li>\n<li>452: add (4379, 2287) in pups</li>\n<li>495: remove (2823, 1661) in adult_females (two dots on the same sealion)</li>\n<li>566: add (1550, 3103) in subadult_males</li>\n<li>587: add (1708, 122), (1667, 150), (1553, 174), (1684, 210) in pups</li>\n<li>629: add (4738, 2395) in pups</li>\n<li>699: add (2046, 1885) in adult_females</li>\n<li>715: add (3287, 552) in pups</li>\n<li>791: add (2256, 585) in adult_females</li>\n<li>803: add (1541, 3342) in adult_females</li>\n<li>828: add (2682, 2349) in adult_females</li>\n</ul>",
      "rawMarkdown": "Thank you for your awesome work! That's really helpful.\n\nI checked discrepancies between your coordinates and mines, I found a few errors that you may want to correct ((x, y) format):\n\n - 80: (1697, 414) should be subadult_males instead of adult_females (it will gives the right numbers in train.csv)\n - 112: Replace (1362, 1156) (adult_females) by (1365, 1153) (position of the dot incorrectly detected)\n - 127: add (3014, 3053) in adult_females\n - 170: (3234, 623) should be adult_females instead of adult_males (to match counts if train.csv)\n - 174: remove adult_male: (3710, 3500), this is a false detection\n - 186: add (2335, 1273) and (3177, 1408) in adult_females (missed detections)\n - 197: add (2959, 1042) in adult_females\n - 291: remove 3 wrong detections in adult_males: (1452, 49), (1482, 49), (1503, 49)\n - 416: add  (1385, 543) in juveniles\n - 418: add (1622, 1032), (1715, 1151), (1510, 1229) and (1648, 1109) in adult_females\n - 420: add (4733, 437) in adult_females\n - 452: add (4379, 2287) in pups\n - 495: remove (2823, 1661) in adult_females (two dots on the same sealion)\n - 566: add (1550, 3103) in subadult_males\n - 587: add (1708, 122), (1667, 150), (1553, 174), (1684, 210) in pups\n - 629: add (4738, 2395) in pups\n - 699: add (2046, 1885) in adult_females\n - 715: add (3287, 552) in pups\n - 791: add (2256, 585) in adult_females\n - 803: add (1541, 3342) in adult_females\n - 828: add (2682, 2349) in adult_females",
      "votes": null
    },
    {
      "id": "182828",
      "postDate": "05/16/2017 02:39:34",
      "content": "<p>Hi @Lowik Chanussot,\nThank you very much for your helpful comment! I assumed the coordinates you gave were in (x,y) format, correct me if I am wrong. </p>\n\n<p><strong>Based on your errors I have a couple questions</strong>:</p>\n\n<ul>\n<li>112: you mentioned to change a coordinate but the coordinate you requested to change to was the exact same, did you perhaps mean (1365,1153)?</li>\n<li>170: (3234, 623) was labeled as an adult_males, does it still need correction?</li>\n<li>803: for the (846, 2120) in adult_males, I have the coordinate for that adult_males at (850,2121) which is slightly to the left so it should be fine.  the (1541, 3342) in adult_females has been added.</li>\n</ul>\n\n<p><strong>Here are the images that have been corrected</strong>:</p>\n\n<ul>\n<li>80</li>\n<li>127</li>\n<li>174</li>\n<li>186</li>\n<li>197</li>\n<li>291</li>\n<li>416</li>\n<li>418</li>\n<li>420</li>\n<li>452</li>\n<li>495</li>\n<li>566</li>\n<li>587</li>\n<li>629</li>\n<li>699</li>\n<li>715</li>\n<li>791</li>\n<li>828</li>\n</ul>\n\n<p><strong>I will re-upload the changed correct_coordinates.csv, correct_train.csv, and changes.csv. If there are any other issues, I'd be happy to fix those too!</strong></p>",
      "rawMarkdown": "Hi @Lowik Chanussot,\nThank you very much for your helpful comment! I assumed the coordinates you gave were in (x,y) format, correct me if I am wrong. \n\n**Based on your errors I have a couple questions**:\n\n* 112: you mentioned to change a coordinate but the coordinate you requested to change to was the exact same, did you perhaps mean (1365,1153)?\n* 170: (3234, 623) was labeled as an adult_males, does it still need correction?\n* 803: for the (846, 2120) in adult_males, I have the coordinate for that adult_males at (850,2121) which is slightly to the left so it should be fine.  the (1541, 3342) in adult_females has been added.\n\n**Here are the images that have been corrected**:\n\n* 80\n* 127\n* 174\n* 186\n* 197\n* 291\n* 416\n* 418\n* 420\n* 452\n* 495\n* 566\n* 587\n* 629\n* 699\n* 715\n* 791\n* 828\n\n**I will re-upload the changed correct_coordinates.csv, correct_train.csv, and changes.csv. If there are any other issues, I'd be happy to fix those too!**",
      "votes": null
    },
    {
      "id": "182984",
      "postDate": "05/16/2017 15:54:07",
      "content": "<ul>\n<li>Yes, it is in (x, y) format</li>\n<li>112: Right.</li>\n<li>170: I corrected my post. My suggestion is to put this point in adult_female instead of adult_males, to match the counts in train.csv (but it is hard to see the correct category otherwise)</li>\n<li>803: ok\nI edited my post to take into account your remarks. Thanks! </li>\n</ul>",
      "rawMarkdown": "Yes, it is in (x, y) format\n- 112: Right.\n- 170: I corrected my post. My suggestion is to put this point in adult_female instead of adult_males, to match the counts in train.csv (but it is hard to see the correct category otherwise)\n- 803: ok\nI edited my post to take into account your remarks. Thanks!",
      "votes": null
    },
    {
      "id": "183071",
      "postDate": "05/16/2017 21:40:30",
      "content": "<p>@Lowik Chanussot ,</p>\n\n<ul>\n<li>112: (1362, 1156) (adult_females) has been replaced by (1365, 1153)</li>\n<li>170: (3234, 623) has been changed to adult_females instead of adult_males</li>\n<li>803: no changes made</li>\n</ul>\n\n<p>If you want to review the changes made to the files, the first set of changes are located in <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\">this github commit</a>, and the changes to images 112 and 170 are located in <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665\">this github commit</a>. </p>\n\n<p><strong>Once again, thanks for the help! I'll re-edit the main post to keep an edits log with time markers and contributors.</strong></p>",
      "rawMarkdown": "Lowik Chanussot ,\n\n* 112: (1362, 1156) (adult_females) has been replaced by (1365, 1153)\n* 170: (3234, 623) has been changed to adult_females instead of adult_males\n* 803: no changes made\n\nIf you want to review the changes made to the files, the first set of changes are located in [this github commit][1], and the changes to images 112 and 170 are located in [this github commit][2]. \n\n**Once again, thanks for the help! I'll re-edit the main post to keep an edits log with time markers and contributors.**\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\n  [2]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665",
      "votes": null
    },
    {
      "id": "186633",
      "postDate": "05/28/2017 19:16:26",
      "content": "<p>image 691: coordinates of the male must be (1550, 105) </p>",
      "rawMarkdown": "image 691: coordinates of the male must be (1550, 105)",
      "votes": null
    },
    {
      "id": "187342",
      "postDate": "05/31/2017 02:02:33",
      "content": "<p>@Artem.Sanakoev , </p>\n\n<p>Thanks for catching that coordinate, I removed the coordinate in the ocean, and added your suggested coordinate, if you want to review changes made to the files, <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5\">here</a> is the github commit. </p>",
      "rawMarkdown": "Artem.Sanakoev , \n\nThanks for catching that coordinate, I removed the coordinate in the ocean, and added your suggested coordinate, if you want to review changes made to the files, [here][1] is the github commit. \n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5",
      "votes": null
    },
    {
      "id": "189546",
      "postDate": "06/06/2017 10:04:10",
      "content": "<p>I ran this on the  small training sample and am getting huge mismatches. is the actual data and the sample same?</p>",
      "rawMarkdown": "I ran this on the  small training sample and am getting huge mismatches. is the actual data and the sample same?",
      "votes": null
    },
    {
      "id": "189550",
      "postDate": "06/06/2017 10:23:23",
      "content": "<p>Sorry @saurabh bansal, I don't quite understand what you mean, did you run your model on a subset of the data I posted here and get very incorrect results? Or is it you are saying the data I posted differs greatly from the data given by the competition?</p>",
      "rawMarkdown": "Sorry @saurabh bansal, I don't quite understand what you mean, did you run your model on a subset of the data I posted here and get very incorrect results? Or is it you are saying the data I posted differs greatly from the data given by the competition?",
      "votes": null
    },
    {
      "id": "189572",
      "postDate": "06/06/2017 11:24:34",
      "content": "<p>I used the code on the sample of data (10 images) provided in the competition and the results are very different from the actuals. (Data:  <a href=\"https://www.kaggle.com/c/noaa-fisheries-steller-sea-lion-population-count/download/TrainSmall2.7z\">https://www.kaggle.com/c/noaa-fisheries-steller-sea-lion-population-count/download/TrainSmall2.7z</a> )</p>",
      "rawMarkdown": "I used the code on the sample of data (10 images) provided in the competition and the results are very different from the actuals. (Data:  https://www.kaggle.com/c/noaa-fisheries-steller-sea-lion-population-count/download/TrainSmall2.7z )",
      "votes": null
    },
    {
      "id": "189970",
      "postDate": "06/07/2017 02:07:00",
      "content": "<p>Sorry if I'm mistaken, I so you are saying the number of sea lions in my train.csv file <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\">here</a> is different from the number of sea lions reported by the original train.csv distributed by the competition hosts?</p>\n\n<p>If so, that is correct that my data is different from the competition provided data, because the small sample of data like you said is images 41 to 50, and according to my <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/changes.csv\">changes.csv file</a> the number of sea lions was wrong for image 47 so that count had to be changed, but besides that there should be no major discrepancies.</p>",
      "rawMarkdown": "Sorry if I'm mistaken, I so you are saying the number of sea lions in my train.csv file [here][1] is different from the number of sea lions reported by the original train.csv distributed by the competition hosts?\n\nIf so, that is correct that my data is different from the competition provided data, because the small sample of data like you said is images 41 to 50, and according to my [changes.csv file][2] the number of sea lions was wrong for image 47 so that count had to be changed, but besides that there should be no major discrepancies.\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\n  [2]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/changes.csv",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 182027,
      "author_name": "authman",
      "author_url": "",
      "post_date": "05/12/2017 04:12:23",
      "content": "<p>Looks like you've put a lot of hard work into this -- congrats on this milestone, and moreover, thanks for sharing!</p>\n\n<p>I'm a little concerned that we haven't had any feedback from the organizers regarding the accuracy + validity of the training dataset. Over the past month, there have been many kernels (almost every public kernel so far in fact), which has attempted to address the problem of training data discrepancies. As we're closing in on just a month and a half left, I hope the organizers agree to do something such as use the same pipeline outlined above by @LivingProgram to: 1) formally accept his updated training coordinates as G.T., and 2) apply the same process to clean up the unreleased stage2 data, if need be.</p>",
      "votes": null,
      "replies": [
        {
          "id": 182087,
          "author_name": "livingprogram",
          "author_url": "",
          "post_date": "05/12/2017 10:31:43",
          "content": "<p>Thanks @authman , \nI strongly agree with you about verifying the validity of the training dataset, and just to add on, those unmarked sea lions I mentioned in the <strong>\"Other Issues Noted, But Not Addressed\"</strong> could cause some discrepancies both in the training stage as well as determining the final results if there is the same problem in the final Leaderboard Data. So hopefully those issues can be addressed as well by organizers. :)</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 182415,
          "author_name": "armamut",
          "author_url": "",
          "post_date": "05/13/2017 18:24:20",
          "content": "<p>Hi, I see that you did a nice job and worked a lot. Thanks,</p>\n\n<p>Is it legal to use your data?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 182460,
          "author_name": "livingprogram",
          "author_url": "",
          "post_date": "05/14/2017 00:12:21",
          "content": "<p>Hi @armamut,\nI <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/LICENSE.txt\">open sourced my data</a>, so I think it is legal as everyone in the competition can access it (so there should be no problem with the <strong>\"No private sharing outside teams\"</strong> rule). I am also not distributing the original data, so it should comply with the competition rules. If need be we could try to get some clarification from competition organizers.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 182547,
      "author_name": "nzeuwik",
      "author_url": "",
      "post_date": "05/14/2017 14:01:28",
      "content": "<p>Thank you for your awesome work! That's really helpful.</p>\n\n<p>I checked discrepancies between your coordinates and mines, I found a few errors that you may want to correct ((x, y) format):</p>\n\n<ul>\n<li>80: (1697, 414) should be subadult_males instead of adult_females (it will gives the right numbers in train.csv)</li>\n<li>112: Replace (1362, 1156) (adult_females) by (1365, 1153) (position of the dot incorrectly detected)</li>\n<li>127: add (3014, 3053) in adult_females</li>\n<li>170: (3234, 623) should be adult_females instead of adult_males (to match counts if train.csv)</li>\n<li>174: remove adult_male: (3710, 3500), this is a false detection</li>\n<li>186: add (2335, 1273) and (3177, 1408) in adult_females (missed detections)</li>\n<li>197: add (2959, 1042) in adult_females</li>\n<li>291: remove 3 wrong detections in adult_males: (1452, 49), (1482, 49), (1503, 49)</li>\n<li>416: add  (1385, 543) in juveniles</li>\n<li>418: add (1622, 1032), (1715, 1151), (1510, 1229) and (1648, 1109) in adult_females</li>\n<li>420: add (4733, 437) in adult_females</li>\n<li>452: add (4379, 2287) in pups</li>\n<li>495: remove (2823, 1661) in adult_females (two dots on the same sealion)</li>\n<li>566: add (1550, 3103) in subadult_males</li>\n<li>587: add (1708, 122), (1667, 150), (1553, 174), (1684, 210) in pups</li>\n<li>629: add (4738, 2395) in pups</li>\n<li>699: add (2046, 1885) in adult_females</li>\n<li>715: add (3287, 552) in pups</li>\n<li>791: add (2256, 585) in adult_females</li>\n<li>803: add (1541, 3342) in adult_females</li>\n<li>828: add (2682, 2349) in adult_females</li>\n</ul>",
      "votes": null,
      "replies": [
        {
          "id": 182828,
          "author_name": "livingprogram",
          "author_url": "",
          "post_date": "05/16/2017 02:39:34",
          "content": "<p>Hi @Lowik Chanussot,\nThank you very much for your helpful comment! I assumed the coordinates you gave were in (x,y) format, correct me if I am wrong. </p>\n\n<p><strong>Based on your errors I have a couple questions</strong>:</p>\n\n<ul>\n<li>112: you mentioned to change a coordinate but the coordinate you requested to change to was the exact same, did you perhaps mean (1365,1153)?</li>\n<li>170: (3234, 623) was labeled as an adult_males, does it still need correction?</li>\n<li>803: for the (846, 2120) in adult_males, I have the coordinate for that adult_males at (850,2121) which is slightly to the left so it should be fine.  the (1541, 3342) in adult_females has been added.</li>\n</ul>\n\n<p><strong>Here are the images that have been corrected</strong>:</p>\n\n<ul>\n<li>80</li>\n<li>127</li>\n<li>174</li>\n<li>186</li>\n<li>197</li>\n<li>291</li>\n<li>416</li>\n<li>418</li>\n<li>420</li>\n<li>452</li>\n<li>495</li>\n<li>566</li>\n<li>587</li>\n<li>629</li>\n<li>699</li>\n<li>715</li>\n<li>791</li>\n<li>828</li>\n</ul>\n\n<p><strong>I will re-upload the changed correct_coordinates.csv, correct_train.csv, and changes.csv. If there are any other issues, I'd be happy to fix those too!</strong></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 182984,
          "author_name": "nzeuwik",
          "author_url": "",
          "post_date": "05/16/2017 15:54:07",
          "content": "<ul>\n<li>Yes, it is in (x, y) format</li>\n<li>112: Right.</li>\n<li>170: I corrected my post. My suggestion is to put this point in adult_female instead of adult_males, to match the counts in train.csv (but it is hard to see the correct category otherwise)</li>\n<li>803: ok\nI edited my post to take into account your remarks. Thanks! </li>\n</ul>",
          "votes": null,
          "replies": []
        },
        {
          "id": 183071,
          "author_name": "livingprogram",
          "author_url": "",
          "post_date": "05/16/2017 21:40:30",
          "content": "<p>@Lowik Chanussot ,</p>\n\n<ul>\n<li>112: (1362, 1156) (adult_females) has been replaced by (1365, 1153)</li>\n<li>170: (3234, 623) has been changed to adult_females instead of adult_males</li>\n<li>803: no changes made</li>\n</ul>\n\n<p>If you want to review the changes made to the files, the first set of changes are located in <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\">this github commit</a>, and the changes to images 112 and 170 are located in <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665\">this github commit</a>. </p>\n\n<p><strong>Once again, thanks for the help! I'll re-edit the main post to keep an edits log with time markers and contributors.</strong></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 186633,
      "author_name": "asanakoev",
      "author_url": "",
      "post_date": "05/28/2017 19:16:26",
      "content": "<p>image 691: coordinates of the male must be (1550, 105) </p>",
      "votes": null,
      "replies": [
        {
          "id": 187342,
          "author_name": "livingprogram",
          "author_url": "",
          "post_date": "05/31/2017 02:02:33",
          "content": "<p>@Artem.Sanakoev , </p>\n\n<p>Thanks for catching that coordinate, I removed the coordinate in the ocean, and added your suggested coordinate, if you want to review changes made to the files, <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5\">here</a> is the github commit. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 189546,
      "author_name": "saurabhbansal",
      "author_url": "",
      "post_date": "06/06/2017 10:04:10",
      "content": "<p>I ran this on the  small training sample and am getting huge mismatches. is the actual data and the sample same?</p>",
      "votes": null,
      "replies": [
        {
          "id": 189550,
          "author_name": "livingprogram",
          "author_url": "",
          "post_date": "06/06/2017 10:23:23",
          "content": "<p>Sorry @saurabh bansal, I don't quite understand what you mean, did you run your model on a subset of the data I posted here and get very incorrect results? Or is it you are saying the data I posted differs greatly from the data given by the competition?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 189572,
          "author_name": "saurabhbansal",
          "author_url": "",
          "post_date": "06/06/2017 11:24:34",
          "content": "<p>I used the code on the sample of data (10 images) provided in the competition and the results are very different from the actuals. (Data:  <a href=\"https://www.kaggle.com/c/noaa-fisheries-steller-sea-lion-population-count/download/TrainSmall2.7z\">https://www.kaggle.com/c/noaa-fisheries-steller-sea-lion-population-count/download/TrainSmall2.7z</a> )</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 189970,
          "author_name": "livingprogram",
          "author_url": "",
          "post_date": "06/07/2017 02:07:00",
          "content": "<p>Sorry if I'm mistaken, I so you are saying the number of sea lions in my train.csv file <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\">here</a> is different from the number of sea lions reported by the original train.csv distributed by the competition hosts?</p>\n\n<p>If so, that is correct that my data is different from the competition provided data, because the small sample of data like you said is images 41 to 50, and according to my <a href=\"https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/changes.csv\">changes.csv file</a> the number of sea lions was wrong for image 47 so that count had to be changed, but besides that there should be no major discrepancies.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "182015": "**Hello everyone!**\n\nI have correct coordinates available for download **[here][1]**, and correct counts available for download **[here][2]**. If you are curious about how these are generated I have an IPython Notebook that explains everything and can be run to generate the same results **[here][3]**. \n\n**Briefly Summarizing What I Did**:\n\n - Use **[Radu's Method][4]** to extract preliminary coordinates\n - Generate a count for each image based on number of preliminary coordinates for each class\n - Compare that count to the **train.csv** counts\n - Draw coordinates for all the images that have count discrepancies\n - Manually look over the images with discrepancies (over 252 total)\n - Add any missing coordinates, remove any misplaced coordinates\n - Generate correct count based on final number of coordinates\n\n**Other Issues Noted, But Not Addressed**:\n\n - 448 ~ 3 unmarked sea lions in the rocky islands on the middle\n - 513 ~ darker unmarked sea lions\n - 519 ~ 2 unmarked sea lions in the bottom right corner\n - 555 ~ many unmarked sea lions in the water\n - 571 ~ unmarked sea lion at the center of the island\n - 578 ~ unmarked sea lion towards in the bottom left quadrant\n - 629 ~ unmarked sea lion in the ocean\n - 655 ~ darker unmarked sea lions\n - 675 ~ unmarked sea lion top right quadrant\n - 679 ~ unmarked sea lion in the water near the coast\n - 686 ~ unmarked sea lion in the water near the coast\n - 724 ~ 3 unmarked sea lions in the water, 1 unmarked sea lion on tiny island\n - 732 ~ unmarked sea lion in the water in the bottom left quadrant\n - 759 ~ sea lion marked DEAD, meaning is not understood\n - 761 ~ possibly a few adult female sea lions marked as adult male\n - 764 ~ 4 unmarked sea lions\n - 776 ~ unmarked black sea lions\n - 790 ~ unmarked black sea lions\n - 802 ~ unmarked sea lion in bottom right quadrant\n - 806 ~ 2 unmarked sea lions that had just dove into the water \n - 822 ~ unmarked sea lion in the upper right quadrant, in the water\n - 823 ~ unmarked sea lion in the water\n - 837 ~ unmarked sea lions in bottom right quadrant, text in the top left quadrant (\"Bulls: 1 SAMs:0 Females:0 Juveniles:0 Pups:0 Total:1\")\n - 845 ~ possibly a few adult female sea lions marked as adult male\n - 847 ~ unmarked sea lion bottom right quadrant\n - 857 ~ no marked sea lions present\n - 858 ~ unmarked sea lion in bottom right quadrant ocean, unmarked sea lion to left of island in the ocean\n - 899 ~ unmarked sea lion top left quadrant\n - 900 ~ unmarked sea lion bottom left quadrant\n - 906 ~ unmarked sea lion in water off the coast\n - 924 ~ shiny unmarked sea lions (most likely not stellar sea lions)\n - 929 ~ unmarked sea lion in water (top right quadrant)\n - 939 ~ 2 unmarked sea lions in cluster of sea lions\n\n**Good Luck, Hope This Helps!**\n\n**If you have questions, I'll gladly answer them down below**\n\n##Change Log: ##\n* **5/16/2017**, credit to **@Lowik Chanussot**, changes made to data for images 80, 112, 127, 170, 174, 186, 197, 291, 416, 418, 420, 452, 495, 566, 587, 629, 699, 715, 791, 828 (refer to indepth explanation of changes below, github commits are [here][5] and [here][6])\n* **5/30/2017**, credit to **@Artem.Sanakoev**, changes made to data for image 691 (refer to indepth explanation of changes below, github commit is [here][7])\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_coordinates.csv\n  [2]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\n  [3]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/Correct%20Coordinates%20%26%20Sea%20Lion%20Counts.ipynb\n  [4]: https://www.kaggle.com/radustoicescu/get-coordinates-using-blob-detection\n  [5]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\n  [6]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665\n  [7]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5",
    "182027": "Looks like you've put a lot of hard work into this -- congrats on this milestone, and moreover, thanks for sharing!\n\nI'm a little concerned that we haven't had any feedback from the organizers regarding the accuracy + validity of the training dataset. Over the past month, there have been many kernels (almost every public kernel so far in fact), which has attempted to address the problem of training data discrepancies. As we're closing in on just a month and a half left, I hope the organizers agree to do something such as use the same pipeline outlined above by @LivingProgram to: 1) formally accept his updated training coordinates as G.T., and 2) apply the same process to clean up the unreleased stage2 data, if need be.",
    "182087": "Thanks @authman , \nI strongly agree with you about verifying the validity of the training dataset, and just to add on, those unmarked sea lions I mentioned in the **\"Other Issues Noted, But Not Addressed\"** could cause some discrepancies both in the training stage as well as determining the final results if there is the same problem in the final Leaderboard Data. So hopefully those issues can be addressed as well by organizers. :)",
    "182415": "Hi, I see that you did a nice job and worked a lot. Thanks,\n\nIs it legal to use your data?",
    "182460": "Hi @armamut,\nI [open sourced my data][1], so I think it is legal as everyone in the competition can access it (so there should be no problem with the **\"No private sharing outside teams\"** rule). I am also not distributing the original data, so it should comply with the competition rules. If need be we could try to get some clarification from competition organizers.\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/LICENSE.txt",
    "182547": "Thank you for your awesome work! That's really helpful.\n\nI checked discrepancies between your coordinates and mines, I found a few errors that you may want to correct ((x, y) format):\n\n - 80: (1697, 414) should be subadult_males instead of adult_females (it will gives the right numbers in train.csv)\n - 112: Replace (1362, 1156) (adult_females) by (1365, 1153) (position of the dot incorrectly detected)\n - 127: add (3014, 3053) in adult_females\n - 170: (3234, 623) should be adult_females instead of adult_males (to match counts if train.csv)\n - 174: remove adult_male: (3710, 3500), this is a false detection\n - 186: add (2335, 1273) and (3177, 1408) in adult_females (missed detections)\n - 197: add (2959, 1042) in adult_females\n - 291: remove 3 wrong detections in adult_males: (1452, 49), (1482, 49), (1503, 49)\n - 416: add  (1385, 543) in juveniles\n - 418: add (1622, 1032), (1715, 1151), (1510, 1229) and (1648, 1109) in adult_females\n - 420: add (4733, 437) in adult_females\n - 452: add (4379, 2287) in pups\n - 495: remove (2823, 1661) in adult_females (two dots on the same sealion)\n - 566: add (1550, 3103) in subadult_males\n - 587: add (1708, 122), (1667, 150), (1553, 174), (1684, 210) in pups\n - 629: add (4738, 2395) in pups\n - 699: add (2046, 1885) in adult_females\n - 715: add (3287, 552) in pups\n - 791: add (2256, 585) in adult_females\n - 803: add (1541, 3342) in adult_females\n - 828: add (2682, 2349) in adult_females",
    "182828": "Hi @Lowik Chanussot,\nThank you very much for your helpful comment! I assumed the coordinates you gave were in (x,y) format, correct me if I am wrong. \n\n**Based on your errors I have a couple questions**:\n\n* 112: you mentioned to change a coordinate but the coordinate you requested to change to was the exact same, did you perhaps mean (1365,1153)?\n* 170: (3234, 623) was labeled as an adult_males, does it still need correction?\n* 803: for the (846, 2120) in adult_males, I have the coordinate for that adult_males at (850,2121) which is slightly to the left so it should be fine.  the (1541, 3342) in adult_females has been added.\n\n**Here are the images that have been corrected**:\n\n* 80\n* 127\n* 174\n* 186\n* 197\n* 291\n* 416\n* 418\n* 420\n* 452\n* 495\n* 566\n* 587\n* 629\n* 699\n* 715\n* 791\n* 828\n\n**I will re-upload the changed correct_coordinates.csv, correct_train.csv, and changes.csv. If there are any other issues, I'd be happy to fix those too!**",
    "182984": "Yes, it is in (x, y) format\n- 112: Right.\n- 170: I corrected my post. My suggestion is to put this point in adult_female instead of adult_males, to match the counts in train.csv (but it is hard to see the correct category otherwise)\n- 803: ok\nI edited my post to take into account your remarks. Thanks!",
    "183071": "Lowik Chanussot ,\n\n* 112: (1362, 1156) (adult_females) has been replaced by (1365, 1153)\n* 170: (3234, 623) has been changed to adult_females instead of adult_males\n* 803: no changes made\n\nIf you want to review the changes made to the files, the first set of changes are located in [this github commit][1], and the changes to images 112 and 170 are located in [this github commit][2]. \n\n**Once again, thanks for the help! I'll re-edit the main post to keep an edits log with time markers and contributors.**\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/bd82a5c0746b0670c4dfb6fcf5ed5b99cd59ff58\n  [2]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/90958a3bf3b70a65f8442965d36e382792f31665",
    "186633": "image 691: coordinates of the male must be (1550, 105)",
    "187342": "Artem.Sanakoev , \n\nThanks for catching that coordinate, I removed the coordinate in the ocean, and added your suggested coordinate, if you want to review changes made to the files, [here][1] is the github commit. \n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/commit/d312278783d7ba8e0a7986fa6f21fd6876e451e5",
    "189546": "I ran this on the  small training sample and am getting huge mismatches. is the actual data and the sample same?",
    "189550": "Sorry @saurabh bansal, I don't quite understand what you mean, did you run your model on a subset of the data I posted here and get very incorrect results? Or is it you are saying the data I posted differs greatly from the data given by the competition?",
    "189572": "I used the code on the sample of data (10 images) provided in the competition and the results are very different from the actuals. (Data:  https://www.kaggle.com/c/noaa-fisheries-steller-sea-lion-population-count/download/TrainSmall2.7z )",
    "189970": "Sorry if I'm mistaken, I so you are saying the number of sea lions in my train.csv file [here][1] is different from the number of sea lions reported by the original train.csv distributed by the competition hosts?\n\nIf so, that is correct that my data is different from the competition provided data, because the small sample of data like you said is images 41 to 50, and according to my [changes.csv file][2] the number of sea lions was wrong for image 47 so that count had to be changed, but besides that there should be no major discrepancies.\n\n\n  [1]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/correct_train.csv\n  [2]: https://github.com/LivingProgram/kaggle-sea-lion-data/blob/master/changes.csv"
  },
  "source": "meta"
}