# COVID-19 AI Detection Challenge — Data

来源：https://www.kaggle.com/competitions/covid-19-ai-detection-challenge/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

# **Data**
The competition data include Computed Tomography (CT) scan images. Each image has one label that describes a clinical state of an individual.
The competition labels consist of three types:
-**COVID-19** label indicates whether experts detect the presence of COVID-19 lesion.
-**NON-COVID-Normal** label indicates the normal clinical outcome diagnosed by the medical experts.
-**Pathological-NON-COVID** label indicates the presence of a disease other than COVID-19 detected by experts.

The competition data were collected from **Habib Bourguiba Hospital of Sfax** in 2022, thereby approved by medical experts.

The competition data are organized into three distinct sets: training, validation, and test sets. All labels are included in the different sets. We have publicly released 80% of the dataset as the training set for the competition, and we have retained the remaining 20% as test set.

**# Files**
train - Training images
train_label.csv - Tagged class for an image in train folder
test - Test Images
sample_submission.csv - a sample submission file in the correct format