{
  "id": 78699,
  "title": "1000 Hand-Annotated Humpback Whale Fluke Keypoints",
  "url": "/competitions/humpback-whale-identification/discussion/78699",
  "author_name": "",
  "post_date": "2019-01-27T04:03:03.884271600Z",
  "votes": 31,
  "comment_count": 8,
  "views": 0,
  "content": "<p>Hey all,</p>\n\n<p>I'm releasing a dataset of 1000 images that I hand annotated to track 10 specific points on whale flukes. I've provided some starter code to just see how to parse the data and plot it.</p>\n\n<p>Unfortunately I don't have time to seriously compete in this contest, but I do hope that this will be helpful to some contenders.</p>\n\n<p>If you do end up using it, I'd appreciate a footnote in your post competition write up.</p>\n\n<p>Happy Kaggling,\n-Paul</p>\n\n<p>Starter Kernel: Feel free to fork/copy/use</p>\n\n<p><a href=\"https://www.kaggle.com/oewyn000/starter-humpback-whale-fluke-keypoints\">https://www.kaggle.com/oewyn000/starter-humpback-whale-fluke-keypoints</a></p>\n\n<p>Dataset itself:</p>\n\n<p><a href=\"https://www.kaggle.com/oewyn000/humpback-whale-fluke-keypoints\">https://www.kaggle.com/oewyn000/humpback-whale-fluke-keypoints</a></p>",
  "messages": [
    {
      "id": "461817",
      "postDate": "01/27/2019 04:03:03",
      "content": "<p>Hey all,</p>\n\n<p>I'm releasing a dataset of 1000 images that I hand annotated to track 10 specific points on whale flukes. I've provided some starter code to just see how to parse the data and plot it.</p>\n\n<p>Unfortunately I don't have time to seriously compete in this contest, but I do hope that this will be helpful to some contenders.</p>\n\n<p>If you do end up using it, I'd appreciate a footnote in your post competition write up.</p>\n\n<p>Happy Kaggling,\n-Paul</p>\n\n<p>Starter Kernel: Feel free to fork/copy/use</p>\n\n<p><a href=\"https://www.kaggle.com/oewyn000/starter-humpback-whale-fluke-keypoints\">https://www.kaggle.com/oewyn000/starter-humpback-whale-fluke-keypoints</a></p>\n\n<p>Dataset itself:</p>\n\n<p><a href=\"https://www.kaggle.com/oewyn000/humpback-whale-fluke-keypoints\">https://www.kaggle.com/oewyn000/humpback-whale-fluke-keypoints</a></p>",
      "rawMarkdown": "Hey all,\n\nI'm releasing a dataset of 1000 images that I hand annotated to track 10 specific points on whale flukes. I've provided some starter code to just see how to parse the data and plot it.\n\nUnfortunately I don't have time to seriously compete in this contest, but I do hope that this will be helpful to some contenders.\n\nIf you do end up using it, I'd appreciate a footnote in your post competition write up.\n\nHappy Kaggling,\n-Paul\n\nStarter Kernel: Feel free to fork/copy/use\n\nhttps://www.kaggle.com/oewyn000/starter-humpback-whale-fluke-keypoints\n\nDataset itself:\n\nhttps://www.kaggle.com/oewyn000/humpback-whale-fluke-keypoints",
      "votes": null
    },
    {
      "id": "462029",
      "postDate": "01/27/2019 14:13:29",
      "content": "<p>Great job ! By training regression model, we can predict keypoints for other images.</p>",
      "rawMarkdown": "Great job ! By training regression model, we can predict keypoints for other images.",
      "votes": null
    },
    {
      "id": "462319",
      "postDate": "01/28/2019 05:22:03",
      "content": "<p>Awesome job.Thanks for posting this</p>",
      "rawMarkdown": "Awesome job.Thanks for posting this",
      "votes": null
    },
    {
      "id": "462384",
      "postDate": "01/28/2019 08:15:27",
      "content": "<p>Thanks for sharing</p>",
      "rawMarkdown": "Thanks for sharing",
      "votes": null
    },
    {
      "id": "463126",
      "postDate": "01/29/2019 13:24:47",
      "content": "<p>thanks!</p>\n\n<p>what annotation too did you use?</p>",
      "rawMarkdown": "thanks!\n\nwhat annotation too did you use?",
      "votes": null
    },
    {
      "id": "463207",
      "postDate": "01/29/2019 16:09:51",
      "content": "<p>I'm very much a fan of your contributions to the discussions in competitions Heng! Thanks so much for your attention!</p>\n\n<p>I used <a href=\"https://github.com/wkentaro/labelme\">https://github.com/wkentaro/labelme</a>. It seemed to be the easiest to create point annotations in an offline manner that I could find. Sometimes it has errors and you have to relaunch the program but makes it very easy to label 10 points quickly.</p>",
      "rawMarkdown": "I'm very much a fan of your contributions to the discussions in competitions Heng! Thanks so much for your attention!\n\nI used https://github.com/wkentaro/labelme. It seemed to be the easiest to create point annotations in an offline manner that I could find. Sometimes it has errors and you have to relaunch the program but makes it very easy to label 10 points quickly.",
      "votes": null
    },
    {
      "id": "463558",
      "postDate": "01/30/2019 07:19:21",
      "content": "<p>Great job !Thanks for sharing</p>",
      "rawMarkdown": "Great job !Thanks for sharing",
      "votes": null
    },
    {
      "id": "469240",
      "postDate": "02/10/2019 18:50:12",
      "content": "<p>I am wondering if someone could share how we can use those key points?</p>",
      "rawMarkdown": "I am wondering if someone could share how we can use those key points?",
      "votes": null
    },
    {
      "id": "1809118",
      "postDate": "06/02/2022 12:10:54",
      "content": "<p>thank you for sharing!</p>",
      "rawMarkdown": "thank you for sharing!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1809118,
      "author_name": "ariannaever",
      "author_url": "",
      "post_date": "06/02/2022 12:10:54",
      "content": "<p>thank you for sharing!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 462029,
      "author_name": "toshik",
      "author_url": "",
      "post_date": "01/27/2019 14:13:29",
      "content": "<p>Great job ! By training regression model, we can predict keypoints for other images.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 462319,
      "author_name": "rednivrug",
      "author_url": "",
      "post_date": "01/28/2019 05:22:03",
      "content": "<p>Awesome job.Thanks for posting this</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 462384,
      "author_name": "xf1994",
      "author_url": "",
      "post_date": "01/28/2019 08:15:27",
      "content": "<p>Thanks for sharing</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 463126,
      "author_name": "hengck23",
      "author_url": "",
      "post_date": "01/29/2019 13:24:47",
      "content": "<p>thanks!</p>\n\n<p>what annotation too did you use?</p>",
      "votes": null,
      "replies": [
        {
          "id": 463207,
          "author_name": "oewyn000",
          "author_url": "",
          "post_date": "01/29/2019 16:09:51",
          "content": "<p>I'm very much a fan of your contributions to the discussions in competitions Heng! Thanks so much for your attention!</p>\n\n<p>I used <a href=\"https://github.com/wkentaro/labelme\">https://github.com/wkentaro/labelme</a>. It seemed to be the easiest to create point annotations in an offline manner that I could find. Sometimes it has errors and you have to relaunch the program but makes it very easy to label 10 points quickly.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 463558,
      "author_name": "sanikamal",
      "author_url": "",
      "post_date": "01/30/2019 07:19:21",
      "content": "<p>Great job !Thanks for sharing</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 469240,
      "author_name": "strideradu",
      "author_url": "",
      "post_date": "02/10/2019 18:50:12",
      "content": "<p>I am wondering if someone could share how we can use those key points?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "461817": "Hey all,\n\nI'm releasing a dataset of 1000 images that I hand annotated to track 10 specific points on whale flukes. I've provided some starter code to just see how to parse the data and plot it.\n\nUnfortunately I don't have time to seriously compete in this contest, but I do hope that this will be helpful to some contenders.\n\nIf you do end up using it, I'd appreciate a footnote in your post competition write up.\n\nHappy Kaggling,\n-Paul\n\nStarter Kernel: Feel free to fork/copy/use\n\nhttps://www.kaggle.com/oewyn000/starter-humpback-whale-fluke-keypoints\n\nDataset itself:\n\nhttps://www.kaggle.com/oewyn000/humpback-whale-fluke-keypoints",
    "462029": "Great job ! By training regression model, we can predict keypoints for other images.",
    "462319": "Awesome job.Thanks for posting this",
    "462384": "Thanks for sharing",
    "463126": "thanks!\n\nwhat annotation too did you use?",
    "463207": "I'm very much a fan of your contributions to the discussions in competitions Heng! Thanks so much for your attention!\n\nI used https://github.com/wkentaro/labelme. It seemed to be the easiest to create point annotations in an offline manner that I could find. Sometimes it has errors and you have to relaunch the program but makes it very easy to label 10 points quickly.",
    "463558": "Great job !Thanks for sharing",
    "469240": "I am wondering if someone could share how we can use those key points?",
    "1809118": "thank you for sharing!"
  },
  "source": "meta"
}