Child Detection Dataset (CDD) - Pediatric Object Detection
The largest public real-world dataset for child detection (1,928 images), focusing on anatomical differences, occlusions, and clinical applicability.📄 Context & Inspiration
Standard object detection models (like COCO-pretrained YOLO) often fail to detect children accurately. This is due to distinct anatomical proportions: children have a head-to-body ratio of approximately 1:4, compared to 1:7 in adults, along with proportionally… See the full description on the dataset page:
https://huggingface.co/datasets/Slownite/ChildDetectionDataset.
