Traffic Signboard Detection Dataset for Object Detection Models

This dataset contains 1,475 labeled images of traffic signboards, curated for developing and evaluating object detection and recognition models in the field of computer vision.The dataset includes annotations for 7 classes of traffic signs, commonly found in real-world road environments. Each image contains one or more signboards labeled with bounding boxes suitable for training models such as YOLO, SSD, or Faster R-CNN.

Classes and Distribution

Class Name
Count

30
243

40
48

Converge
6

Left Turn
272

Right Turn
242

School
362

T Shape
364

The dataset exhibits class imbalance, reflecting real-world conditions where certain signs are more frequent than others. Users are encouraged to apply data augmentation or class weighting strategies for balanced model performance.

Applications

Traffic sign detection and classification

Autonomous driving systems

Road safety monitoring and analytics

Computer vision research and benchmarking

Format

Image type: JPEG/PNG

Annotation format: YOLO-compatible text files (or specify if COCO/XML etc.)

Total images: 1,475

Usage

This dataset is intended for research and educational purposes. It can be used to train and evaluate deep learning models for small object detection and classification in real-world scenarios.