Ready-made dataset

High-Quality Full Locomotion Pose Dataset — Walking + Jogging + Running

Professional multi-action locomotion bundle for pose estimation, gait analysis, and motion ML — one export covering walking, jogging, and running from diverse Pexels sources.

Job bulk_190_videos_4e1065b2: 190 source videos processed with the same quality-first pipeline as our other ready-made bundles.

One-time: $400Source videos: 190Total rows: 44,691 (incl. augments)Accepted (pre-aug): 14,897

Dataset overview

  • Actions: walking, jogging, and running (combined export)
  • Quantity: 44,691 total rows (including augmentations)
  • Original HQ frames: 14,897 accepted before augmentation
  • Quality: Mean frame quality score 0.8571 / 1.0 (accepted)
  • Acceptance rate: 55.68% (pre-aug / sampled; strict HQ filtering)
  • Mean FPS (export): ~27.7 · Stride: 3

Multi-layer technical stack

  • Layer 1.0 (pose): 33-landmark MediaPipe/BlazePose set (x, y, z, visibility)
  • Layer 1.1: RGB mean/std, brightness, contrast, edge density, and 12-bin hue histograms
  • Temporal smoothing: Gaussian filter (window=5) on x/y/z
  • Body normalization: Hip-centered + torso-scale fields included
  • Augmentations: Horizontal flip + keypoint noise (duplicates rows in merged data.jsonl)

Structured data layout

  • data.jsonl — global merged dataset
  • per_video/*.jsonl — per-source splits
  • features.json — sequence-level metrics (motion consistency, velocity proxies)
  • low_quality_frames.jsonl — rejected frames for hard-example mining
  • manifest.json, global_stats.json, export_quality_report.json — metadata and QA

Free samples

Validate the format on GitHub: QualityVision-Motion-Dataset-Samples.

Purchase & licensing

  • Checkout: Gumroad (one-time; see current listing on product page)
  • Deliverable: JSONL + metadata exports — not raw source videos
  • License: Proprietary / commercial terms as listed on Gumroad
  • Contact: info@qvision.space
Buy on Gumroad Contact for custom datasetsinfo@qvision.spaceBack to pricing