{
  "fetched_at": "2026-10-19T08:37:56Z",
  "source_url": "https://www.kaggle.com/competitions/csc532/data",
  "pages": [
    {
      "name": "data-description",
      "content": "<p><b>Phase I:</b></p>\n<p>This challenge has two labelled datasets and one unlabelled one.</p>\n<h2>File descriptions</h2>\n<ul>\n<li><strong>mini-dataset.zip</strong> - the mini-hackathon training set containing two folders of normal/non-covid and covid</li>\n<li><strong>datasets.zip</strong> - the training/test set containing two folders of training and test data with each folder containing sub-folders of negative/non-covid and positive/covid</li>\n<li><strong>sampleSubmission.csv</strong> - a sample submission file in the correct format</li>\n<li><strong>DLAI3_Validation_Set.zip</strong> - unlabelled set to be tested and submit in form of the sampleSubmission.csv which is two columns with first column as the filename and the second being \"0\" for non-covid and \"1\" for covid</li>\n</ul>\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "<p><b>Phase I:</b></p>\n<p>This is a repeat challenge from DLAI3: (<a href=\"https://www.kaggle.com/c/dlai3\">https://www.kaggle.com/c/dlai3</a>). Please watch out for another challenge from DLAI5 soon.</p>\n\n<p>The COVID-19 pandemic has been gripping much of the world recently. Although there have been much efforts gone into developing affordable testing for the masses, it has been shown that the established and widely available chest x-rays (CXR) may be used as a screening criteria. Thanks to the dedicated work by various individuals and organizations, publicly available chest x-rays of COVID-19 subjects are available for analytic usage. This challenge is a continuation from the CSC532 Machine Learning course projects from the School of Information Technology, King Mongkut's University of Technology Thonburi (KMUTT), which provided the mini-dataset and the dataset of annotated CXR images.</p>\n<h2>Goal</h2>\n<p>The following is the premise for this challenge:</p>\n<ul>\n<li>Available COVID-19 chest x-ray data from a given facility is limited. This is like the mini-dataset.</li>\n<li>Unseen images may come from various sources  that may need appropriate preprocessing.</li>\n<li><i>Besides the larger dataset provided, participants are welcome to use external sources to supplement the available data.</i></li>\n</ul>\n<h2>Acknowledgements</h2>\n<p><b>Phase I:</b></p>\n<p>The larger dataset was from the CSC532 Machine Learning project of <b>Puttipong Thammachart &amp; Vasin Virasak</b>; the mini-dataset was from the project of <b>Phubeth Rodklang</b>.</p>\n<p>A special thanks to https://github.com/ieee8023/covid-chestxray-dataset as a source for part of the datasets provided here. A follow up to this challenge will likely take place later with updated images from this site and others. Please note the following message from this website though as follows. <br>\"<i>Note: please do not claim diagnostic performance of a model without a clinical study! This is not a kaggle competition dataset. Please read this paper about evaluation issues: https://arxiv.org/abs/2004.12823 and https://arxiv.org/abs/2004.05405</i>\"</p>",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "<p>The evaluation metric is the F1 score, which combines the statistical measures of precision p and recall r. Precision is the ratio of true positives (tp) to all predicted positives (tp + fp). Recall is the ratio of true positives to all actual positives (tp + fn). The F1 score is given by:</p>\n<p>\\[ F1 = 2\\frac{p \\cdot r}{p+r}\\ \\ \\mathrm{where}\\ \\ p = \\frac{tp}{tp+fp},\\ \\ r = \\frac{tp}{tp+fn} \\]</p>\n<p>The F1 metric weights recall and precision equally, and a good retrieval algorithm will maximize both precision and recall simultaneously. Thus, moderately good performance on both will be favored over extremely good performance on one and poor performance on the other.</p>\n<h2>Submission Format</h2>\n<p><strong>For every author in the dataset</strong>, submission files should contain two columns: unlabelled image name and predicted label as \"0\" for non-covid or \"1\" for covid. </p>\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "By participating in this challenge, the participants acknowledges that the datasets are provided for the purpose of the competition only. There is no claim to ownership/authenticity/validity of the datasets and annotations. Please do not rely on the datasets provided here for diagnosis purpose without detailed validation. Any publications must cite the following: Chan, J.H., Thammachart, P., Virasak, V., and Rodklang, P. (2020, July). DLAI3 Hackathon: COVID-19 Chest x-ray challenge, Version 1.0. Retrieved [date retrieved] from https://www.kaggle.com/c/dlai3/.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}