Person Detection Images from COCO for YOLO Models

Subset of the COCO dataset filtered to contain only images with person instances, formatted for training YOLO object detection models. The dataset is tagged for Computer Vision tasks. Specific details on the number of images, annotations, or features are not provided in the input.

Use Cases

• Fine-tune a YOLO model for person detection using the provided bounding box annotations.
• Benchmark person detection model performance on a curated subset of COCO images.
• Use the pre-processed images and labels to train or evaluate YOLO-based architectures.

Strengths

• Derived from the widely-used COCO benchmark dataset, ensuring a standard annotation format.
• Specifically formatted for YOLO models, reducing preprocessing effort for users of that framework.
• Focuses on a single object class (person), which is a common and critical detection task.

Limitations

• The input provides no information on the dataset's size, number of images, or annotation counts.
• Potential class imbalance is inherent as it contains only person instances, limiting multi-class training.
• The recency, geographic coverage, and update frequency of this specific subset are unknown.

Provenance

Source

COCO (Common Objects in Context) dataset.

Collection Method

Subset created by filtering the original COCO dataset for images containing the 'person' category.