Football Players Detection
This dataset contains annotated images for the detection and location of soccer players on the field. It makes it possible to develop and test computer vision models adapted to sport.
106 annotated JPG images in YOLO format (.txt), 375 files in total
Attribution 4.0 International (CC BY 4.0)
Description
The Football Players Detection Dataset offers 106 annotated images in YOLO format with a unique focus on the detection of soccer players. It includes accurate annotations to locate each player in sports scenes.
What is this dataset for?
- Train object detection models to track players on the field
- Develop video analysis tools for sports and live streaming
- Learn and experiment with computer vision and object detection algorithms (e.g. YOLO, Faster R-CNN)
Can it be enriched or improved?
It is possible to increase this dataset by collecting additional images or by data augmentation (rotation, zoom, change in brightness). Multi-class annotation (positions, actions) could also improve wealth.
🔎 In summary
🧠 Recommended for
- Computer vision students
- Sport analytics researchers
- Sports AI startups
🔧 Compatible tools
- Yolov5/yolov8
- TensorFlow Object Detection API
- PyTorch
- OpenCV
💡 Tip
Apply augmentation techniques to compensate for the small volume of data.
Frequently Asked Questions
Is this dataset sufficient to train a powerful model?
The dataset is small but useful for learning and prototyping; it is advisable to increase the data for better performance.
What annotation format is used?
The annotation format is YOLO, simple, and compatible with several detection frameworks.
Can this dataset be used for other sports?
The annotations are specific to soccer, but the images can inspire models that can be transferred to other sports with adaptation.
