Walking Pose Dataset With Over 1500 High-Quality Frames

QualityVision Walking Pose Dataset contains a sample of over 1500 high-quality image frames. The dataset is designed for human pose estimation tasks, focusing on walking poses. The author and original collection date are unknown.

Use Cases

• Train 2D keypoint detection models on high-quality frames of walking humans.
• Benchmark pose estimation algorithms using the provided walking pose annotations.
• Develop models for gait analysis by leveraging the specific walking pose data.
• Create synthetic data pipelines by learning pose distributions from the labeled frames.

Strengths

• Contains over 1500 image frames.
• Frames are described as high-quality (HQ).

Limitations

• Sample size of 1500+ frames is relatively small for deep learning training.
• Limited to walking poses, which may not generalize to other activities.
• Potential for demographic or environmental bias in the image collection is unknown.

Provenance

Source

QualityVision (implied by dataset name).