Alaaharoun's picture
Walking sample: data.jsonl at root; metadata/*.json for viewer-safe layout + README
2998bd3 verified
metadata
license: cc-by-nc-4.0
language:
  - en
tags:
  - pose-estimation
  - walking
  - 2d-pose
  - mediapipe
  - motion-dataset
  - quality-vision
  - biometrics
  - human-computer-interaction
  - rehabilitation-data
  - sports-science
  - motion-capture
pretty_name: QualityVision Walking Sample (compact, 2D pose)
size_categories:
  - n<1K
configs:
  - config_name: default
    data_files:
      - split: train
        path: data.jsonl

QualityVision Walking Sample (compact)

Need production-scale pose data? Browse ready-made JSONL bundles (thousands of HQ frames, full manifests, schema docs) and Dataset Lab plans on qvision.space — Dataset pricing. This Hub repo is a small non-commercial sample so you can validate parsing and quality before you buy.

Small high-quality 2D pose clip for walking, exported from the Quality Vision Motion Dataset Engine pipeline: HQ frame filtering, optional temporal smoothing on accepted frames, and sequence diagnostics in features.json.

This sample vs full commercial exports

This Hub sample Paid exports & Lab (pricing)
Scale Compact clip (~80 rows in data.jsonl on the Hub) Thousands of HQ frames across curated bundles; higher caps via annual frame quotas
Use case Parse the format, prototype models, academic / NC research under the license Production datasets, larger actions (e.g. running, dancing bundles), repeatable exports
Support Community / docs on GitHub Checkout via Gumroad; commercial licensing and custom scope quoted separately
Deliverables JSONL + metadata/* as listed below Full dataset/ trees (manifests, diagnostics, SHA256, optional one-pager PDFs) per product page

Export format: 2D landmarks in normalized image space (see features.json). This sample does not include raw video files.

For the full 24-source bulk export (~967 HQ frames, ~2901 merged rows, mean quality ~0.82), clone the GitHub repo and use bulk_24_videos_6a173caf_dataset/, or mirror README_HF_BULK.md to a separate Hugging Face dataset via scripts/upload_bulk_to_huggingface.ps1.

Files

File Description
data.jsonl One JSON object per line: normalized image-space keypoints, timestamps, optional Layer 1.1 metadata.
metadata/features.json Sequence-level aggregates (visibility, hip motion, motion_consistency).
metadata/global_stats.json Job-level roll-up for this clip.
metadata/manifest.json Export metadata (format compatible with the Motion Dataset Engine).
metadata/export_quality_report.json Pilot-oriented quality summary.

Auxiliary JSON files live under metadata/ so the Hub dataset viewer only builds previews from data.jsonl (avoids schema clashes with report objects).

Usage

from datasets import load_dataset

# If you load as JSON lines:
ds = load_dataset("Alaaharoun/QualityVision-walking-sample", data_files="data.jsonl", split="train")

Or download raw files:

from huggingface_hub import hf_hub_download
path = hf_hub_download(
    repo_id="Alaaharoun/QualityVision-walking-sample",
    filename="data.jsonl",
    repo_type="dataset",
)

License

CC BY-NC 4.0 — see repository LICENSE on GitHub. Non-commercial use is allowed under the license terms. Commercial redistribution or product use requires a separate agreement — see Dataset pricing & contact for ready-made bundles and licensing.

Citation

@misc{qualityvision_walking_sample,
  title = {QualityVision Walking Sample (2D Pose)},
  author = {Alaaharoun},
  howpublished = {\url{https://huggingface.co/datasets/Alaaharoun/QualityVision-walking-sample}},
  year = {2026},
}