{
  "fetched_at": "2026-10-24T02:42:53Z",
  "source_url": "https://www.kaggle.com/competitions/cs4487-course-project/data",
  "pages": [
    {
      "name": "data-description",
      "content": "#This file contains real and fake faces:\n\n## Real faces: \n- real\n\n## Fake faces: \n- fake_deepfake\n- fake_face2face",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "## Background\n- Techniques for convincing manipulations of digital facial images\nexist for decades via the use of visual effects \n- Recent technologies in deep learning, especially “DeepFakes,”\nhave led to a dramatic increase in the realism of fake content\n- While there are entertaining applications of DeepFakes, potential\nweaponization of such techniques has raised great concerns\n- DeepFake detection has become a pressing issue and a hot spot of\nresearch\n\n## Task: \nA binary classification problem\n- Input: RGB facial images of size 299 \u0002x 299 x\u0002 3\n- Output: A binary label to indicate whether an image is fake or not\n\n## Training set: \n12,000 images, within which 8,000 images contain\nfake faces, and the rest images are real\n- Download link: https://drive.google.com/drive/folders/1n18AqgUF_KqKbcxuLXrTcLq-pyfhSJD9?usp=sharing\n\n## Test set: \n3,000 images that will not be released to the students\n- Students have THREE chances to get access to the test set\n- The best result among the three will be used for ranking\n\n- Tip: It is a good practice to divide the available set further into\ntraining, validation, and test set\n\n## Group project: \nA group of at most two students is allowed\n\n## What to hand in: \na single ipython notebook (.ipynb) as the project report with source code files included\n- Source files must contain the training code for reproduction\n- An extra PDF file as project report is not recommended\n\n## When to hand in: \nDec. 6, 2021, 11:59 pm\n\n## Where to hand in: \n- Submit the clean and runable test code to the TAs\n- Submit the .ipynb file via Canvas\n\n## GPUs: \nSome GPU hours are available on Microsoft Azure\n\n## Grading (Totally 30 points)\n- 30.0% - Technical correctness (whether the\nmethodologies/algorithms are correctly used)\n- 30.0% - Experiment and analysis\n- 20.0% - Quality of the written report (organized, complete\ndescriptions, etc.)\n- 10.0% - Quality of project presentation (tentatively held in Wed,\nNov. 24)\n- 10.0% - Reserved for Top-5 teams based on the test set\nperformance",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": " - Don't cheat!\n - Apply yourself!\n - Have fun!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "## Evaluation Metric\nThe evaluation metric for this competition is Classification Accuracy, i.e. the percentage of your correct predictions over all your predictions on test set.\n\n\n## Submission Format\n\nYour submission files should be in .csv format. The data frame should contain two columns: `Id` and `Category`.  Each element of  column `Id` is the name of an image in test set, which is a string (e.g. 0.png). Each element of column `Category` should be an string from {0, 1}  where 0 for fake and 1 for real.  \n\n\nThe .csv file should contain a header and have the following format:\n```\nId,Category\n0.png,1\n1.png,1\n2.png,0\n...\n3999.png,0\n```",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Leaderboard",
      "content": "## Leaderboard of accuracy\n\nGroup 4 - 0.9786\n\nGroup 13 - 0.9779\n\nGroup 6 - 0.97755\n\nGroup 14 - 0.97755\n\nGroup 5 - 0.9514\n\nGroup 15 - 0.9384",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}