CCTV-Pedestrian-1K: High-Angle Surveillance Images for Person Attribute Recognition

CCTV-Pedestrian-1K is a dataset of high-angle surveillance pedestrian images intended for training Vision Transformers (ViT) and Vision-Language Models (VLM). The dataset is hosted on Kaggle and is tagged for applications in public safety and computer vision. Specific details on the number of images, collection time, and creator are not provided in the available metadata.

Use Cases

• Train Vision Transformer (ViT) models for pedestrian detection based on high-angle surveillance imagery.
• Develop Vision-Language Models (VLM) for describing person attributes from video frames.
• Benchmark public safety algorithms for person identification and tracking in CCTV footage.
• Fine-tune models for attribute recognition tasks like clothing, pose, and activity classification.

Strengths

• Dataset is specifically designed for modern architectures like Vision Transformers (ViT) and Vision-Language Models (VLM).
• Focus on high-angle surveillance perspective, which is a common real-world scenario for public safety applications.

Limitations

• Row count and dataset scale are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Kaggle

Collection Method

Likely collected from high-angle surveillance camera footage.