{
  "id": 147171,
  "title": "iNaturalist bounding boxes",
  "url": "/competitions/iwildcam-2020-fgvc7/discussion/147171",
  "author_name": "",
  "post_date": "2020-04-29T18:19:35.844460300Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi!</p>\n\n<p>On the github page of the competition, in the <strong>Camera Trap Animal Detection Model</strong> you write that </p>\n\n<blockquote>\n  <p>We are also providing a general animal detection model which competitors are free to use as they see fit. ... We have run the detector over the three datasets, and provide the top 100 boxes and associated confidences along with the metadata for WCS.</p>\n</blockquote>\n\n<p>Do I understand correctly that by the 'three datasets' you mean WCS, iNaturalist2017 and iNaturalist2018 datasets?</p>\n\n<p>If so, I think I have not found files where the bounding boxes for iNaturalist datasets are located. I have downloaded and looked through <code>inaturalist_2018_to_iwildcam_train.json</code> file, but have not found it there, neither I could find it here <code>iwildcam2020_megadetector_results.json</code>. </p>\n\n<p>Could you, please, correct me if I missed the necessary file? </p>\n\n<p>Thank you!</p>",
  "messages": [
    {
      "id": "826600",
      "postDate": "04/29/2020 18:19:35",
      "content": "<p>Hi!</p>\n\n<p>On the github page of the competition, in the <strong>Camera Trap Animal Detection Model</strong> you write that </p>\n\n<blockquote>\n  <p>We are also providing a general animal detection model which competitors are free to use as they see fit. ... We have run the detector over the three datasets, and provide the top 100 boxes and associated confidences along with the metadata for WCS.</p>\n</blockquote>\n\n<p>Do I understand correctly that by the 'three datasets' you mean WCS, iNaturalist2017 and iNaturalist2018 datasets?</p>\n\n<p>If so, I think I have not found files where the bounding boxes for iNaturalist datasets are located. I have downloaded and looked through <code>inaturalist_2018_to_iwildcam_train.json</code> file, but have not found it there, neither I could find it here <code>iwildcam2020_megadetector_results.json</code>. </p>\n\n<p>Could you, please, correct me if I missed the necessary file? </p>\n\n<p>Thank you!</p>",
      "rawMarkdown": "Hi!\n\nOn the github page of the competition, in the **Camera Trap Animal Detection Model** you write that \n\n&gt; We are also providing a general animal detection model which competitors are free to use as they see fit. ... We have run the detector over the three datasets, and provide the top 100 boxes and associated confidences along with the metadata for WCS.\n\nDo I understand correctly that by the 'three datasets' you mean WCS, iNaturalist2017 and iNaturalist2018 datasets?\n\nIf so, I think I have not found files where the bounding boxes for iNaturalist datasets are located. I have downloaded and looked through `inaturalist_2018_to_iwildcam_train.json` file, but have not found it there, neither I could find it here `iwildcam2020_megadetector_results.json`. \n\nCould you, please, correct me if I missed the necessary file? \n\nThank you!",
      "votes": null
    },
    {
      "id": "826689",
      "postDate": "04/29/2020 19:36:50",
      "content": "<p>You found a typo!  I copied over a lot of text from iWildCam 2019, where we had three separate datasets.  iNaturalist 2017 has bounding box labels, but I don't believe iNaturalist 2018 has provided any.  We have a detector that works well on camera trap data, so we've used it to provide boxes for the WCS data.  I'll fix the mistake, good catch :)</p>",
      "rawMarkdown": "You found a typo!  I copied over a lot of text from iWildCam 2019, where we had three separate datasets.  iNaturalist 2017 has bounding box labels, but I don't believe iNaturalist 2018 has provided any.  We have a detector that works well on camera trap data, so we've used it to provide boxes for the WCS data.  I'll fix the mistake, good catch :)",
      "votes": null
    },
    {
      "id": "826726",
      "postDate": "04/29/2020 20:22:29",
      "content": "<p>=)\nthanks a lot!</p>",
      "rawMarkdown": ")\nthanks a lot!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 826689,
      "author_name": "sbeery",
      "author_url": "",
      "post_date": "04/29/2020 19:36:50",
      "content": "<p>You found a typo!  I copied over a lot of text from iWildCam 2019, where we had three separate datasets.  iNaturalist 2017 has bounding box labels, but I don't believe iNaturalist 2018 has provided any.  We have a detector that works well on camera trap data, so we've used it to provide boxes for the WCS data.  I'll fix the mistake, good catch :)</p>",
      "votes": null,
      "replies": [
        {
          "id": 826726,
          "author_name": "olegpolivin",
          "author_url": "",
          "post_date": "04/29/2020 20:22:29",
          "content": "<p>=)\nthanks a lot!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "826600": "Hi!\n\nOn the github page of the competition, in the **Camera Trap Animal Detection Model** you write that \n\n&gt; We are also providing a general animal detection model which competitors are free to use as they see fit. ... We have run the detector over the three datasets, and provide the top 100 boxes and associated confidences along with the metadata for WCS.\n\nDo I understand correctly that by the 'three datasets' you mean WCS, iNaturalist2017 and iNaturalist2018 datasets?\n\nIf so, I think I have not found files where the bounding boxes for iNaturalist datasets are located. I have downloaded and looked through `inaturalist_2018_to_iwildcam_train.json` file, but have not found it there, neither I could find it here `iwildcam2020_megadetector_results.json`. \n\nCould you, please, correct me if I missed the necessary file? \n\nThank you!",
    "826689": "You found a typo!  I copied over a lot of text from iWildCam 2019, where we had three separate datasets.  iNaturalist 2017 has bounding box labels, but I don't believe iNaturalist 2018 has provided any.  We have a detector that works well on camera trap data, so we've used it to provide boxes for the WCS data.  I'll fix the mistake, good catch :)",
    "826726": ")\nthanks a lot!"
  },
  "source": "meta"
}