# Traffic Sign Localization and Classification. — Data

来源：https://www.kaggle.com/competitions/elte-traffic-signs/data

本地 Meta Kaggle 快照；不是本次实时页面。

# Dataset Description

# Data description

All files required are attached in the data section of the challenge.

## What am I predicting?  

You are predicting a row in the form: id, class,x1,y1,x2,y2 for each image where x1,y1,x2 and y2 are the corner coordinates in pixels of the bounding box of the object to localize. 
 
## Files provided

*   **train.csv** - The training set: A set of image paths and ground truth labels for class and bounding box.
*   **test.csv** - The test set: A set of image paths without ground truth predictions.
*   **sample.csv** - A sample submission file in the correct format. This shows you how your submission file should look.

## Columns
*   `id` - Identification of png image in dataset.
*   `new_path` - Path to png image in dataset.
*   `class` - Number from 0-4 that defines the class of the sign.  
*   `x1` -  X coordinate of bounding bottom box corner. 
*   `y1` -  Y coordinate of bounding bottom box corner. 
*   `x2` -  X coordinate of bounding top box corner. 
*   `y2` -  Y coordinate of bounding top box corner.