An image dataset for shopping trolley detection in computer vision with annotated YOLO-ready labels

This dataset is designed for training and evaluating object detection models, specifically optimized for the detection of shopping trolleys (shopping carts) in retail, supermarket, and parking lot environments.

The collection consists of 2,793 standardized images and their corresponding annotations in YOLO format. To minimize false positives and improve model robustness against complex environments, the dataset includes 250 contextual background images (empty backgrounds) with zero annotations.

Key dataset statistics:
• Total images: 2,793
• Images containing shopping trolleys: 2,543
• Background/Empty images: 250
• Total annotated object instances (bounding boxes): 3,571
• Image specification: 512x512 pixels, center-cropped, format PNG.
• Number of classes: 1 (Class 0: 'shopping-trolley')