Request access to this dataset sample
We review requests within 1 business day. Access is granted for evaluation and research use. Commercial licensing and larger cuts are handled separately by Zovda.
Full dataset, custom cuts and pricing: https://www.zovda.com/lp/llm-training-data/video/?utm_source=huggingface&utm_medium=gate_form&utm_campaign=dm0010
By requesting access you agree to use this sample for evaluation only, not to redistribute it, and that Zovda may contact you about the full dataset. We use your details only to respond to this request and to discuss licensing.
Log in or Sign Up to review the conditions and access this dataset content.
DM0010 · Egocentric Physical Interaction Samples / 第一人称物理世界交互数据样例
| Item / 项目 | Value / 值 |
|---|---|
| Dataset ID / 数据集编号 | DM0010 |
| Files / 数据文件 | 12 |
| Size / 数据体积 | 185.97 MB |
| Provided by / 提供方 | Zovda |
| Contact / 联系邮箱 | lucas@zovda.com |
Overview / 简介
An end-to-end delivery sample of one egocentric capture session: master and proxy media, an annotated preview, and seven structured annotation types covering actions, arm pose, camera motion, hands, objects, hand-object interaction and key steps.
第一人称采集 Session 的端到端交付样例：主视频、代理视频、叠加标注预览与七类结构化标注，覆盖动作、手臂姿态、相机运动、手部、物体、手物交互与关键步骤。
What's inside / 数据内容
media/proxies/: 10 fps proxy video for fast review, annotation and QA.media/proxies/：10fps 代理视频，供快速浏览、标注与质检。media/annotated_previews/: review video with annotation overlays.media/annotated_previews/：叠加标注的审核预览视频。media/masters_sanitized/: compliance-processed master video (large; contact us to obtain).media/masters_sanitized/：完成合规处理的主版视频（体积较大，如需获取请联系我们）。annotations/: seven JSONL annotation types —actions,arm_pose,camera_motion,hands,interactions,keysteps,objects.annotations/：七类逐帧/逐段 JSONL 标注——actions、arm_pose、camera_motion、hands、interactions、keysteps、objects。docs/: data card, annotation schema (ANNOTATION_SCHEMA_RGB_ONLY_L1) and delivery audit report.docs/：数据卡（DATACARD）、标注 Schema（ANNOTATION_SCHEMA_RGB_ONLY_L1）与交付审计报告。
Specifications / 规格
- Organisation: by capture session; this sample is P001_S001_T001
- 组织方式：按 Session（采集场次）组织，本样例为 P001_S001_T001
- Annotation format: JSONL, per frame or per segment, timestamped
- 标注格式：JSONL，逐帧或逐段，带时间戳
- Media: master MOV, 10 fps proxy MP4, annotated preview MP4
- 媒体：主版 MOV、10fps 代理 MP4、标注预览 MP4
Suitable tasks / 适用任务
- Embodied-AI / robot-manipulation pretraining
- 具身智能/机器人操作数据预训练
- Hand and object pose estimation, hand-object interaction detection
- 手部与物体位姿估计、手物交互检测
- Egocentric action recognition and key-step segmentation
- 第一人称动作识别与任务步骤分割
- Camera-motion modelling and scene understanding
- 相机运动建模与场景理解
Files / 文件清单
| # | File | Size |
|---|---|---|
| 1 | ego-delivery/annotations/actions/P001_S001_T001_actions.jsonl |
18.53 KB |
| 2 | ego-delivery/annotations/arm_pose/P001_S001_T001_visible_arm_estimates.jsonl |
239.06 KB |
| 3 | ego-delivery/annotations/camera_motion/P001_S001_T001_camera_motion_2d.jsonl |
177.93 KB |
| 4 | ego-delivery/annotations/hands/P001_S001_T001_hands.jsonl |
7.94 MB |
| 5 | ego-delivery/annotations/interactions/P001_S001_T001_hand_object_candidates.jsonl |
124.18 KB |
| 6 | ego-delivery/annotations/keysteps/P001_S001_T001_keysteps.jsonl |
228 B |
| 7 | ego-delivery/annotations/objects/P001_S001_T001_objects_l1.jsonl |
1.45 MB |
| 8 | ego-delivery/docs/ANNOTATION_SCHEMA_RGB_ONLY_L1.md |
3.39 KB |
| 9 | ego-delivery/docs/DATACARD.md |
1.09 KB |
| 10 | ego-delivery/docs/DELIVERABLES_AND_SOW_AUDIT.md |
30.59 KB |
| 11 | ego-delivery/media/annotated_previews/S001_annotated_h264.mp4 |
71.42 MB |
| 12 | ego-delivery/media/proxies/P001_S001_T001_proxy_10fps.mp4 |
104.59 MB |
Usage / 使用方式
from huggingface_hub import snapshot_download
snapshot_download(
repo_id="Zovda/dm0010-egocentric-physical-interaction-samples",
repo_type="dataset",
local_dir="./dm0010-egocentric-physical-interaction-samples",
)
Rights and disclaimer / 授权与声明
Collected, cleaned and annotated by Zovda. Samples are shared to evaluate data structure and annotation conventions; contact us for commercial licensing at scale.
本数据集由 Zovda 采集、清洗与标注，样例可用于评估数据结构与标注口径。规模化商用授权请联系我们。
This dataset is shared to demonstrate data organisation, technical specifications and delivery structure. For formal licensing, large-scale access or custom production, contact lucas@zovda.com.
本数据集仅用于展示数据组织方式、技术规格与交付结构。正式授权、规模化取数与定制生产请联系 lucas@zovda.com。
About Zovda / 关于 Zovda
Zovda builds, cleans, annotates and packages training data across video, audio, images, e-commerce image-text pairs, exam question banks, embodied AI and coding-agent corpora.
Zovda 提供多模态与文本训练语料的数据采集、清洗、标注与结构化交付服务，覆盖视频、音频、图像、电商图文、K12 与职业题库、具身智能、代码智能体等方向。
Contact / 联系方式: lucas@zovda.com
- Downloads last month
- 133