# 2nd AVA Challenge@IEEE MIPR 2024 — Data

来源：https://www.kaggle.com/competitions/2nd-ava-challenge-ieee-mipr-2024/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

Your challenge in this competition is to forecast the car accident risk. In this dataset, we provide the videos of roads and freeways to let you train the model.

Please use training dataset to create your own model, and make classification on testing dataset. The name of each film is generated randomly. For your convenience, we named the road data after road_0000 to road_0299, while freeway is freeway_0000 to freeway_0301. Each film has a folder containing the corresponding jpg files.

**5/24 update : The dataset has been updated. Please refer to the overview for detailed information.**

## Files

*   **2nd_AVA_DATASET_2ed**
         |-------  **freeway**
         |          |-------  **train** - the training videos of freeway
         |          |-------  **test** - the testing videos of freeway
         |-------  **road**
         |          |-------  **train** - the training videos of road
         |          |-------  **test** - the testing videos of road
         |-------  **freeway_train.csv** - risk labels of videos in the freeway folder
         |-------  **road_train.csv** - risk labels of videos in the road folder
         |-------  **sample_submission.csv** - the example of submission format

## Columns

*   `film name` - name of videos (randomly generated)
*   `risk` - the answer, represents the risk level of the car accident's impact on the recording vehicle. 1 indicates high risk, while 0 indicates low risk.