{
  "id": 150943,
  "title": "more image processing",
  "url": "/competitions/alaska2-image-steganalysis/discussion/150943",
  "author_name": "",
  "post_date": "2020-05-13T21:01:37.956644200Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>we all know how this competition is about to find the most small and least change in the image channel bits to clearly define which is stegno graphed image or not so to channelize the underlying images we should see and analyze these factors so i have created this topic where people can come and discuss about what image processing would soot better for the competition</p>",
  "messages": [
    {
      "id": "846560",
      "postDate": "05/13/2020 21:01:37",
      "content": "<p>we all know how this competition is about to find the most small and least change in the image channel bits to clearly define which is stegno graphed image or not so to channelize the underlying images we should see and analyze these factors so i have created this topic where people can come and discuss about what image processing would soot better for the competition</p>",
      "rawMarkdown": "we all know how this competition is about to find the most small and least change in the image channel bits to clearly define which is stegno graphed image or not so to channelize the underlying images we should see and analyze these factors so i have created this topic where people can come and discuss about what image processing would soot better for the competition",
      "votes": null
    },
    {
      "id": "846562",
      "postDate": "05/13/2020 21:01:57",
      "content": "<p>starting with me \n<a href=\"https://www.kaggle.com/pranshu29/more-image-processing-ideas-analysis\">https://www.kaggle.com/pranshu29/more-image-processing-ideas-analysis</a></p>",
      "rawMarkdown": "starting with me \nhttps://www.kaggle.com/pranshu29/more-image-processing-ideas-analysis",
      "votes": null
    },
    {
      "id": "846791",
      "postDate": "05/14/2020 02:09:39",
      "content": "<p>Hello there,</p>\n\n<p>There is a common belief in steganalysis (hidden data detection) : do as few processing as possible over the data you will analyse.\nIndeed, while one might think this can help, it usually hurt because it makes the changes more \"fuzzy\", more difficult to target. \nIn general, in detection theory, it is also belief that applying a function over both class does never help : change the decision function instead.</p>\n\n<p>However, it may help for identifying the location of changes and, hence, tell a network where to put more focus / attention.</p>",
      "rawMarkdown": "Hello there,\n\nThere is a common belief in steganalysis (hidden data detection) : do as few processing as possible over the data you will analyse.\nIndeed, while one might think this can help, it usually hurt because it makes the changes more \"fuzzy\", more difficult to target. \nIn general, in detection theory, it is also belief that applying a function over both class does never help : change the decision function instead.\n\nHowever, it may help for identifying the location of changes and, hence, tell a network where to put more focus / attention.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 846562,
      "author_name": "pranshu29",
      "author_url": "",
      "post_date": "05/13/2020 21:01:57",
      "content": "<p>starting with me \n<a href=\"https://www.kaggle.com/pranshu29/more-image-processing-ideas-analysis\">https://www.kaggle.com/pranshu29/more-image-processing-ideas-analysis</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 846791,
          "author_name": "remicogranne",
          "author_url": "",
          "post_date": "05/14/2020 02:09:39",
          "content": "<p>Hello there,</p>\n\n<p>There is a common belief in steganalysis (hidden data detection) : do as few processing as possible over the data you will analyse.\nIndeed, while one might think this can help, it usually hurt because it makes the changes more \"fuzzy\", more difficult to target. \nIn general, in detection theory, it is also belief that applying a function over both class does never help : change the decision function instead.</p>\n\n<p>However, it may help for identifying the location of changes and, hence, tell a network where to put more focus / attention.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "846560": "we all know how this competition is about to find the most small and least change in the image channel bits to clearly define which is stegno graphed image or not so to channelize the underlying images we should see and analyze these factors so i have created this topic where people can come and discuss about what image processing would soot better for the competition",
    "846562": "starting with me \nhttps://www.kaggle.com/pranshu29/more-image-processing-ideas-analysis",
    "846791": "Hello there,\n\nThere is a common belief in steganalysis (hidden data detection) : do as few processing as possible over the data you will analyse.\nIndeed, while one might think this can help, it usually hurt because it makes the changes more \"fuzzy\", more difficult to target. \nIn general, in detection theory, it is also belief that applying a function over both class does never help : change the decision function instead.\n\nHowever, it may help for identifying the location of changes and, hence, tell a network where to put more focus / attention."
  },
  "source": "meta"
}