{
  "id": 59982,
  "title": "2nd place solution (based on Detectron)",
  "url": "/competitions/cvpr-2018-autonomous-driving/writeups/smart-vision-sg-2nd-place-solution-based-on-detect",
  "author_name": "",
  "post_date": "2018-06-29T04:20:41.216908400Z",
  "votes": 11,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Higher resolution is always a better option. \nMy solution is based on Mask R-CNN.  Source code and model are available from below link,</p>\n\n<p><a href=\"https://github.com/Computational-Camera/Kaggle-CVPR-2018-WAD-Video-Segmentation-Challenge-Solution\">https://github.com/Computational-Camera/Kaggle-CVPR-2018-WAD-Video-Segmentation-Challenge-Solution</a></p>",
  "messages": [
    {
      "id": "350048",
      "postDate": "06/29/2018 04:20:41",
      "content": "<p>Higher resolution is always a better option. \nMy solution is based on Mask R-CNN.  Source code and model are available from below link,</p>\n\n<p><a href=\"https://github.com/Computational-Camera/Kaggle-CVPR-2018-WAD-Video-Segmentation-Challenge-Solution\">https://github.com/Computational-Camera/Kaggle-CVPR-2018-WAD-Video-Segmentation-Challenge-Solution</a></p>",
      "rawMarkdown": "Higher resolution is always a better option. \nMy solution is based on Mask R-CNN.  Source code and model are available from below link,\n\nhttps://github.com/Computational-Camera/Kaggle-CVPR-2018-WAD-Video-Segmentation-Challenge-Solution",
      "votes": null
    },
    {
      "id": "350981",
      "postDate": "07/01/2018 00:26:26",
      "content": "<p>Great work, and congrats! Did you explore time based prediction, since the data were in video format? </p>",
      "rawMarkdown": "Great work, and congrats! Did you explore time based prediction, since the data were in video format?",
      "votes": null
    },
    {
      "id": "350997",
      "postDate": "07/01/2018 01:22:35",
      "content": "<p>I didn't. But i think it is definitely a promising way to improve the prediction result by using temporal information.</p>",
      "rawMarkdown": "I didn't. But i think it is definitely a promising way to improve the prediction result by using temporal information.",
      "votes": null
    },
    {
      "id": "351389",
      "postDate": "07/02/2018 04:33:57",
      "content": "<p>This is dissapointing, was hoping for some black magic</p>",
      "rawMarkdown": "This is dissapointing, was hoping for some black magic",
      "votes": null
    },
    {
      "id": "355829",
      "postDate": "07/12/2018 12:26:16",
      "content": "<p>Without investigating relevant technologies and determine best balance of performance and feasibility, the \"innovative\" black magic that you mentioned could be useless piece of junk in real world.</p>",
      "rawMarkdown": "Without investigating relevant technologies and determine best balance of performance and feasibility, the \"innovative\" black magic that you mentioned could be useless piece of junk in real world.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 350981,
      "author_name": "gobears",
      "author_url": "",
      "post_date": "07/01/2018 00:26:26",
      "content": "<p>Great work, and congrats! Did you explore time based prediction, since the data were in video format? </p>",
      "votes": null,
      "replies": [
        {
          "id": 350997,
          "author_name": "shanxu",
          "author_url": "",
          "post_date": "07/01/2018 01:22:35",
          "content": "<p>I didn't. But i think it is definitely a promising way to improve the prediction result by using temporal information.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 351389,
      "author_name": "dicksonchin93",
      "author_url": "",
      "post_date": "07/02/2018 04:33:57",
      "content": "<p>This is dissapointing, was hoping for some black magic</p>",
      "votes": null,
      "replies": [
        {
          "id": 355829,
          "author_name": "overfittinghcl",
          "author_url": "",
          "post_date": "07/12/2018 12:26:16",
          "content": "<p>Without investigating relevant technologies and determine best balance of performance and feasibility, the \"innovative\" black magic that you mentioned could be useless piece of junk in real world.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "350048": "Higher resolution is always a better option. \nMy solution is based on Mask R-CNN.  Source code and model are available from below link,\n\nhttps://github.com/Computational-Camera/Kaggle-CVPR-2018-WAD-Video-Segmentation-Challenge-Solution",
    "350981": "Great work, and congrats! Did you explore time based prediction, since the data were in video format?",
    "350997": "I didn't. But i think it is definitely a promising way to improve the prediction result by using temporal information.",
    "351389": "This is dissapointing, was hoping for some black magic",
    "355829": "Without investigating relevant technologies and determine best balance of performance and feasibility, the \"innovative\" black magic that you mentioned could be useless piece of junk in real world."
  },
  "source": "meta"
}