{
  "fetched_at": "2026-10-19T07:35:27Z",
  "source_url": "https://www.kaggle.com/competitions/ads5035-01/data",
  "pages": [
    {
      "name": "data-description",
      "content": "Dataset Description\n=============\n* Dataset Copyright\n  * All data is generated from the dataset provided by [Faceforensics](https://github.com/ondyari/FaceForensics).\n  * We don't have any ownership of this data, nor do we have participants in this contest.\n* Format of Dataset\n * Inside the dataset, **the images have different sizes**. You should manage the input image by yourself.\n * Every image in this dataset has channel order as follows\n    * RGB\n\n![RGB format image](https://raw.githubusercontent.com/nkmk/python-snippets/7b135319354f28ed55244fae00a7695a2c14e372/notebook/data/dst/lena_bgr_cv.jpg)    \n    * BGR\n\n![BGR format image](https://raw.githubusercontent.com/nkmk/python-snippets/7b135319354f28ed55244fae00a7695a2c14e372/notebook/data/dst/lena_bgr_pillow.jpg)\n\n* Dataset ratio\n    * Training dataset\n         * Total 10000 (Real:5000 counts ,Fake:5000 counts )\n         * Real: Real image from Faceforensics dataset\n         * Fake: Synthetic image that generated by Neural Texture and input was Real Faceforensics dataset    \n    * Validation dataset\n        * Total 5000 ( Real: 2500 counts, Fake:2500 counts)\n        * Real: Real image from Faceforensics dataset\n        * Fake: Synthetic image that generated by Neural Texture and input was Real Faceforensics     \n    * Test dataset\n        * Total 2000 (Unknown Real, Fake ratio)\n        * Real: Real image from Faceforensics dataset\n        * Fake: Synthetic image that generated by Neural Texture and input was Real Faceforensics \n  ",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "<p>Deepfake is an algorithm that allows you to express a person's face by changing it to someone else. Due to recent advances in deepfake technology, high-level deepfake images are being produced. More and more attempts are being made to harm people by using these images with malicious intent.</p>\n\n<p>\n\n</p>\n\n<p>The problem situation given is as follows: You are a security officer at an IT company that actively utilizes face recognition technology. Recently, you have heard that there is a loophole in your company's facial recognition technology. While you were investigating technology loopholes, some malicious people used the Neural Texture model to create fake faces, and you found an attempt to bypass your company's facial recognition technology.\n\n</p>\n<p>The company has assigned you to solve this problem after hearing that you have taken a security class.</p>\n\n\n<p>3 Submission</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "<p>Submissions are scored by categorization accuracy.\nYou need to be as accurate as possible. There are no other restrictions.</p>\n<p> 3 Submission is available per day and ONLY ONE SUBMISSION WILL BE CALCULATED AS FINAL</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "1. Your submissions should be reproducible. \n2. Do not use the validation dataset in training.\n3. We highly recommend training your model in Kaggle or Colab. Because each person has a different capacity of computer and GPU, code the model to learn in a Kagle, Collaboration as much as possible.\n4. Do not share the dataset with others. Do not share it on the Internet.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}