Ready-made dataset

190-Video Full Locomotion Bundle (Walking + Jogging + Running) Pose Dataset

Bulk export bundle for pose estimation, locomotion research, and robotics-style training pipelines.

A ready-to-use pose export bundle (JSONL + metadata) generated by the Quality Vision Motion Dataset Engine. Built with the Quality Vision Motion Dataset Engine (layer11_pose): 190 source clips, HQ gating, and standard augmentations where applicable.

# High-Quality Full Locomotion Bundle (Walking + Jogging + Running) Pose Dataset

Train your AI faster with clean, ready-to-use motion datasets
No preprocessing. No noise. Just results.

## Dataset Overview
- Job ID: `bulk_190_videos_613f7093`
- Action label: `Full Locomotion Bundle (Walking + Jogging + Running)`
- Videos processed: 190
- Exported frames (HQ, incl. augmentations): 40020
- Accepted frames (pre-augmentation): 13340
- Sampled candidate frames: 26755
- Rejected frames: 13415
- Acceptance rate (pre-aug / sampled): 49.86%

## Quality Metrics (Accepted Frames)
- Mean quality score: 0.8661
- Mean landmark visibility: 0.8216
- Mean lower-body visibility: 0.8112
- Mean motion_local: 0.9106

## Processing
- Gaussian temporal smoothing: window=5 (x/y/z only; visibility not smoothed)
- Body normalization: hip-centered + torso-scale
- Augmentations: horizontal flip + keypoint noise

## Deliverables (inside this ZIP)
- data.jsonl
- per_video/video_*.jsonl
- per_video/rejected_*.jsonl
- features.json
- export_quality_report.json
- manifest.json
- global_stats.json
- runtime_config.json
- SCHEMA.md
- examples/
- examples/sample_rows_accepted.jsonl
- examples/sample_rows_rejected.jsonl
- examples/sample_features.json
- examples/LOAD_EXAMPLE.py
- SHA256SUMS

## Quality Score Definition
- Per frame: frame_quality_score = 0.5 * avg_landmark_visibility + 0.0 * lower_body_visibility + 0.5 * motion_local.

https://qvision.space/

# Motion dataset export

## Overview

- **Job**: `bulk_190_videos_613f7093`
- **Source**: **190 videos from Pexels (inferred from filenames)**
- **Action label**: `Full Locomotion Bundle (Walking + Jogging + Running)`
- **HQ exported frames**: **40020** (includes augmentation duplicates)

## Key metrics (quick table)

| Metric | Value |
|---|---:|
| Videos processed | 190 |
| Frames exported (HQ, incl. augmentations) | 40020 |
| Frames sampled (pre-augmentation candidates) | 26755 |
| Frames accepted (pre-augmentation) | 13340 |
| Accepted percentage (pre-aug / sampled) | 49.86% |
| Mean frame quality score (accepted) | 0.8661 |
| Mean avg landmark visibility (accepted) | 0.8216 |
| Mean lower-body visibility (accepted) | 0.8112 |
| Mean motion_local (accepted) | 0.9106 |
| Mean fps | 27.72 |
| Stride | 3 |

See `dataset/global_stats.json` and `dataset/export_quality_report.json` for full details.

## Processing applied

- **Gaussian smoothing** (temporal): enabled (x/y/z only; visibility is not smoothed)
- **Body normalization**: enabled (hip-centered + torso-scale) (`keypoints_body_normalized`, `body_normalization`)

## Layout

- `data.jsonl` — all accepted frames from all videos (global `frame_id`).
- `per_video/video_NNN.jsonl` — same schema, only frames from source index NNN.
- `per_video/rejected_NNN.jsonl` — low-quality / no-pose frames for that source.
- `low_quality_frames.jsonl` — all rejected rows across sources.
- `bulk_combined_manifest.json` — index of combined vs per-video files.
- `features.json` — per-video sequence metrics (motion consistency, velocity proxies, etc.).
- `manifest.json` — job metadata and post-processing flags.

## Accepted percentage

- `accepted_percentage` is computed as **accepted (pre-augmentation) / sampled frames**.
- Exported frame count can exceed sampled frames when augmentations duplicate accepted rows.