Minh124689/test / README.md
Minh124689's picture
|
download
raw
3.12 kB
metadata
language:
  - vi
  - en
license: cc-by-4.0
task_categories:
  - object-detection
tags:
  - traffic-sign
  - vietnam
  - yolov8
  - autonomous-vehicles
size_categories:
  - 10K<n<100K
configs:
  - config_name: default
    data_files:
      - split: train
        path: classid.csv

Vietnam Traffic Sign Detection Dataset

YOLO License

This repository contains the dataset for detecting road traffic signs in Vietnam using the state-of-the-art YOLO object detection model.

Viet Nam Traffic Sign


๐Ÿ“‚ Repository Structure

The dataset is structured in the standard YOLO format, containing images and corresponding annotations divided into training, validation, and testing sets.

โ”œโ”€โ”€ classid.xlsx        # Excel file mapping class IDs to names
โ”œโ”€โ”€ dataset/
โ”‚   โ”œโ”€โ”€ train/          # Training split
โ”‚   โ”‚   โ”œโ”€โ”€ images/     # Raw images (.jpg)
โ”‚   โ”‚   โ””โ”€โ”€ labels/     # Normalized bounding box coordinates (.txt)
โ”‚   โ”œโ”€โ”€ val/            # Validation split
โ”‚   โ”‚   โ”œโ”€โ”€ images/
โ”‚   โ”‚   โ””โ”€โ”€ labels/
โ”‚   โ””โ”€โ”€ test/           # Testing split
โ”‚       โ”œโ”€โ”€ images/
โ”‚       โ””โ”€โ”€ labels/
โ””โ”€โ”€ README.md

๐Ÿ“Š Dataset Distribution

The dataset statistics below show the division of images and labels across the different splits:

Split Images Details
Train 8,125 8,124 labeled images + 36 background (negative) images
Validation 1,016 1,014 labeled images + 2 background (negative) images
Test 1,016 1,014 labeled images + 2 background (negative) images
Total 10,157

๐Ÿท๏ธ Annotation Format

Annotations follow the standard YOLO format. For every image (e.g., 0002.jpg), there is a corresponding label file (e.g., 0002.txt) in the labels/ directory.

Each row in the text file represents a single bounding box formatted as:

<class_id> <x_center> <y_center> <width> <height>
  • class_id: Integer representing the target class index (ranging from 0 to 81).
  • x_center, y_center: Center coordinates of the bounding box, normalized to [0.0, 1.0] by dividing by the image's width and height.
  • width, height: The width and height of the bounding box, normalized to [0.0, 1.0].

๐Ÿšฆ Supported Traffic Sign Classes (82 Classes)

The dataset contains 82 categories of road signs commonly found on Vietnamese streets:

Class distribution


Xet Storage Details

Size:
3.12 kB
ยท
Xet hash:
56dfef5f0e259ad840cb669bb3ed308c82001be33fa098e7e516cf9a9700a717

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.