EgoPressure: A Dataset for Hand Pressure and Pose Estimation in Egocentric Vision
Paper • 2409.02224 • Published
You need to agree to share your contact information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
Access to the EgoPressure dataset is provided for non-commercial academic research under the CC BY-NC-SA 4.0 license. By requesting access you agree to: (1) use the data for non-commercial research purposes only; (2) not redistribute the raw data; (3) cite the CVPR 2025 paper in any publication using this dataset; (4) acknowledge that MANO-derived annotations are additionally subject to the MANO license (https://mano.is.tue.mpg.de/license.html).
Log in or Sign Up to review the conditions and access this dataset content.
EgoPressure
A dataset for hand pressure and pose estimation in egocentric vision (CVPR 2025). 21 participants interact with a pressure-sensing touchpad (Sensel Morph), captured by 1 head-mounted egocentric camera and 7 static Azure Kinect cameras at 30 Hz, with MANO hand meshes and fine-grained pressure annotations for every contact.
- 📄 Paper: https://arxiv.org/abs/2409.02224
- 🌐 Project page: https://yiming-zhao.github.io/EgoPressure/
Layout
configs/<participant>/<sequence>.json # camera calibration + metadata
configs/<participant>/<sequence>_k4a/ # per-camera factory calibrations
data/<participant>/<sequence>/
cam-d.color.parquet cam-1.color.parquet ... cam-7.color.parquet
cam-d.depth.parquet cam-1.depth.parquet ... cam-7.depth.parquet
cam-d.mask.parquet cam-1.mask.parquet ... cam-7.mask.parquet
pressure.parquet # Sensel grid + UV pressure
annotation.parquet # MANO pose/mesh + ego pose
Each Parquet shard holds one row per frame (frame column). Data is sharded
by camera and modality so any subset — a participant, a camera, a modality —
can be downloaded selectively (e.g. huggingface_hub.snapshot_download with
allow_patterns).
Modalities
| Shard | Content |
|---|---|
cam-*.color |
RGB frames as original JPEG bytes (ego d: 1920×1080; static 1..7: 2560×1440; undistorted) |
cam-*.depth |
512×512 uint16 depth in millimetres (PNG bytes), all 8 cameras |
cam-*.mask |
hand segmentation masks (PNG bytes) |
pressure |
raw Sensel Morph grid (105×185 float32, flattened) + normalised 224×224 UV pressure map with [min, max] range |
annotation |
MANO vertices (778×3), joint_position (21×3), betas (10), full_pose (48), transl, normals, displacement, per-static-camera visible_vertices (7×778), per-frame egocentric camera pose (ego_R 3×3, ego_T 3) |
Array columns are flattened float32 lists; reshape to the dimensions above. Frames a camera dropped at capture time are absent from that camera's shard.
Cameras & geometry
- Static cameras (
1..7) carry a fixedModelViewMatrix(world→camera, millimetres) in the sequence config; the egocentric camera's pose is per-frame inannotation(metres). - Images are undistorted → pinhole projection with the config intrinsics.
- The world origin lies on the touchpad surface (240 × 137.5 mm active area,
x–y plane at z = 0); raw sensor counts convert to Newtons via
counts / 1736(cell pitch 1.25 mm). configs/<p>/<seq>_k4a/holds each camera's factory calibration (depth sensor intrinsics + depth→color transform) for depth registration.
License & citation
Released under CC BY-NC-SA 4.0 (non-commercial, academic). MANO-derived annotations are additionally subject to the MANO license.
@InProceedings{Zhao_2025_CVPR,
author = {Zhao, Yiming and Kwon, Taein and Streli, Paul and Pollefeys, Marc and Holz, Christian},
title = {EgoPressure: A Dataset for Hand Pressure and Pose Estimation in Egocentric Vision},
booktitle = {Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR)},
month = {June},
year = {2025},
pages = {27727--27738}
}
- Downloads last month
- 7