{
  "id": 161861,
  "title": "Image quality 95 most difficult to detect?",
  "url": "/competitions/alaska2-image-steganalysis/discussion/161861",
  "author_name": "",
  "post_date": "2020-06-26T12:54:36.452915200Z",
  "votes": 3,
  "comment_count": 4,
  "views": 0,
  "content": "<p>I train three separate 4 class models (predict cover, juniward, uerd, jmipod) for each image quality. I'm currently getting the following CV (weighted AUC):\n- Q 70: 0.919\n- Q 90: 0.935\n- Q 95: 0.889</p>\n\n<p>Is anyone else finding that there model performs the worst for the 95 quality images? </p>",
  "messages": [
    {
      "id": "902911",
      "postDate": "06/26/2020 12:54:36",
      "content": "<p>I train three separate 4 class models (predict cover, juniward, uerd, jmipod) for each image quality. I'm currently getting the following CV (weighted AUC):\n- Q 70: 0.919\n- Q 90: 0.935\n- Q 95: 0.889</p>\n\n<p>Is anyone else finding that there model performs the worst for the 95 quality images? </p>",
      "rawMarkdown": "I train three separate 4 class models (predict cover, juniward, uerd, jmipod) for each image quality. I'm currently getting the following CV (weighted AUC):\n- Q 70: 0.919\n- Q 90: 0.935\n- Q 95: 0.889\n\nIs anyone else finding that there model performs the worst for the 95 quality images?",
      "votes": null
    },
    {
      "id": "903025",
      "postDate": "06/26/2020 14:30:22",
      "content": "<ul>\n<li>Q70: 0.940</li>\n<li>Q90: 0.941</li>\n<li>Q95: 0.895</li>\n</ul>",
      "rawMarkdown": "Q70: 0.940\n- Q90: 0.941\n- Q95: 0.895",
      "votes": null
    },
    {
      "id": "903130",
      "postDate": "06/26/2020 15:36:27",
      "content": "<p>Thanks for sharing! Looks like I'm not the only one.</p>",
      "rawMarkdown": "Thanks for sharing! Looks like I'm not the only one.",
      "votes": null
    },
    {
      "id": "903351",
      "postDate": "06/26/2020 18:47:15",
      "content": "<p>QF70:0.934  QF90:0.9377 QF95:0.8853</p>",
      "rawMarkdown": "QF70:0.934\tQF90:0.9377\tQF95:0.8853",
      "votes": null
    },
    {
      "id": "913315",
      "postDate": "07/03/2020 06:17:17",
      "content": "<p>Q70: 0.931\nQ90: 0.940\nQ95: 0.888</p>",
      "rawMarkdown": "Q70: 0.931\nQ90: 0.940\nQ95: 0.888",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 903025,
      "author_name": "ren4yu",
      "author_url": "",
      "post_date": "06/26/2020 14:30:22",
      "content": "<ul>\n<li>Q70: 0.940</li>\n<li>Q90: 0.941</li>\n<li>Q95: 0.895</li>\n</ul>",
      "votes": null,
      "replies": [
        {
          "id": 903130,
          "author_name": "davelo12",
          "author_url": "",
          "post_date": "06/26/2020 15:36:27",
          "content": "<p>Thanks for sharing! Looks like I'm not the only one.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 903351,
      "author_name": "chenbaoying",
      "author_url": "",
      "post_date": "06/26/2020 18:47:15",
      "content": "<p>QF70:0.934  QF90:0.9377 QF95:0.8853</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 913315,
      "author_name": "sharksbeer",
      "author_url": "",
      "post_date": "07/03/2020 06:17:17",
      "content": "<p>Q70: 0.931\nQ90: 0.940\nQ95: 0.888</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "902911": "I train three separate 4 class models (predict cover, juniward, uerd, jmipod) for each image quality. I'm currently getting the following CV (weighted AUC):\n- Q 70: 0.919\n- Q 90: 0.935\n- Q 95: 0.889\n\nIs anyone else finding that there model performs the worst for the 95 quality images?",
    "903025": "Q70: 0.940\n- Q90: 0.941\n- Q95: 0.895",
    "903130": "Thanks for sharing! Looks like I'm not the only one.",
    "903351": "QF70:0.934\tQF90:0.9377\tQF95:0.8853",
    "913315": "Q70: 0.931\nQ90: 0.940\nQ95: 0.888"
  },
  "source": "meta"
}