{
  "id": 154843,
  "title": "Take into account the cost functions of the embedding schemes?",
  "url": "/competitions/alaska2-image-steganalysis/discussion/154843",
  "author_name": "Panagiotis Kyriakis",
  "post_date": "2020-05-30T04:15:05.534000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hi guys,</p>\n\n<p>I was looking for a quick recommendation. Intuitiely, I would say that we need to take into account the cost functions of different embedding schemes because it seems like it would help.  For example, if the embedding scheme was something really simple like always modifying a predefined set of pixels (independently of the image), then it would obviously be much easier to build a classifier. </p>\n\n<p>To me, this seems to be the way to approach this problem but surprisingly enough I haven't found any papers of methods in the literature. </p>\n\n<p>Let me know what you think. </p>",
  "messages": [
    {
      "id": 867166,
      "postDate": "2020-05-30T04:15:05.533Z",
      "content": "<p>Hi guys,</p>\n\n<p>I was looking for a quick recommendation. Intuitiely, I would say that we need to take into account the cost functions of different embedding schemes because it seems like it would help.  For example, if the embedding scheme was something really simple like always modifying a predefined set of pixels (independently of the image), then it would obviously be much easier to build a classifier. </p>\n\n<p>To me, this seems to be the way to approach this problem but surprisingly enough I haven't found any papers of methods in the literature. </p>\n\n<p>Let me know what you think. </p>",
      "rawMarkdown": "Hi guys,\n\nI was looking for a quick recommendation. Intuitiely, I would say that we need to take into account the cost functions of different embedding schemes because it seems like it would help.  For example, if the embedding scheme was something really simple like always modifying a predefined set of pixels (independently of the image), then it would obviously be much easier to build a classifier. \n\nTo me, this seems to be the way to approach this problem but surprisingly enough I haven't found any papers of methods in the literature. \n\nLet me know what you think. "
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "867166": "Hi guys,\n\nI was looking for a quick recommendation. Intuitiely, I would say that we need to take into account the cost functions of different embedding schemes because it seems like it would help.  For example, if the embedding scheme was something really simple like always modifying a predefined set of pixels (independently of the image), then it would obviously be much easier to build a classifier. \n\nTo me, this seems to be the way to approach this problem but surprisingly enough I haven't found any papers of methods in the literature. \n\nLet me know what you think. "
  }
}