| # H2O Dataset information |
|
|
| ## News |
|
|
| **24/Sep./2022**: We are pleased to announce that our total prize will be up to 4000 CHF (4000 USD) for our two challenges ([hand](https://codalab.lisn.upsaclay.fr/competitions/4822), [action](https://codalab.lisn.upsaclay.fr/competitions/4820)). <br> |
| **15/Aug./2022**: Our challenges are open here ([hand](https://codalab.lisn.upsaclay.fr/competitions/4822), [action](https://codalab.lisn.upsaclay.fr/competitions/4820)). Our 1st phase of the challenges will be closed at October 1st, 2022. <br> |
| **9/May/2022**: The download script (downlaod_script.py) is uploaded. <br> |
| **8/May/2022**: Labels for hand pose and object pose are released. Please check https://h2odataset.ethz.ch <br> |
| **4/May/2022**: The H2O dataset will be part of the workshop, Human Body, Hands, and Activities from Egocentric and Multi-view Cameras@ECCV2022. Please check our workshop site (https://sites.google.com/view/egocentric-hand-body-activity) and our challenges. <br> |
| **12/Mar./2022**: Visualization code is published. https://github.com/taeinkwon/h2oplayer <br> |
| **11/Mar./2022**: Updated new labels to the existing files. <br> |
| **8/Dec./2021**: Please Download new labels "manolabel_v1.1.tar.gz" in https://h2odataset.ethz.ch for MANO parameters. We will also update existing files soon. <br> |
|
|
| # Visualization code (H2OPlayer) |
|  |
| Please check the visualization code in the following link: https://github.com/taeinkwon/h2oplayer. |
|
|
| # Download script |
|
|
| ## Dependencies |
| - python3 |
| - tqdm |
| - pathlib |
| - requests |
| - argparse |
|
|
|
|
| Once you receive the username and password from the download page (https://h2odataset.ethz.ch), you can either download directly from the download page (https://h2odataset.ethz.ch) or use download_script.py. |
| ``` |
| python download_script.py --username "username" --password "password" --mode "type of view" --dest "dest folder path" |
| ``` |
| We provide three different modes (views): |
| - all: download the entire data, including five views (four fixed views, one egocentric view). |
| - ego: download only egocentric-view data. |
| - pose: download only pose (hand, object, egocentric view) without RGB-D images. |
| |
| # Dataset Structure |
| |
| <pre> |
| . |
| ├── h1 |
| │ ├── 0 |
| │ │ │── cam0 |
| │ │ │ ├── rgb |
| │ │ │ ├── depth |
| │ │ │ ├── cam_pose |
| │ │ │ ├── hand_pose |
| │ │ │ ├── hand_pose_MANO |
| │ │ │ ├── obj_pose |
| │ │ │ ├── obj_pose_RT |
| │ │ │ ├── action_label (only in cam4) |
| │ │ │ ├── rgb256 (only in cam4) |
| │ │ │ ├── verb_label |
| │ │ │ └── cam_intrinsics.txt |
| │ │ ├── cam1 |
| │ │ ├── cam2 |
| │ │ ├── cam3 |
| │ │ └── cam4 |
| │ ├── 1 |
| │ ├── 2 |
| │ ├── 3 |
| │ └── ... |
| ├── h2 |
| ├── k1 |
| ├── k2 |
| └── ... |
| </pre> |
| |
| cam0 ~ cam3 are fixed cameras. cam4 is an head-mounted camera (egocentric view). <br> |
| train_sequences = ['subject1/h1', 'subject1/h2', 'subject1/k1', 'subject1/k2', 'subject1/o1', 'subject1/o2', 'subject2/h1', 'subject2/h2', 'subject2/k1', |
| 'subject2/k2', 'subject2/o1', 'subject2/o2', 'subject3/h1', 'subject3/h2', 'subject3/k1'] (subject 1,2,3) <br> |
| val_sequences = ['subject3/k2', 'subject3/o1', 'subject3/o2'] (subject 3)<br> |
| test_sequences = ['subject4/h1', 'subject4/h2', 'subject4/k1', 'subject4/k2', 'subject4/o1', 'subject4/o2'] (subject4)<br> |
| |
| |
| ### rgb |
| 1280 * 720 resolution rgb images |
| |
| ### rgb256 |
| 455 * 256 resolution resized rgb images |
| |
| ### depth |
| 1280 * 720 resolution depth images |
| |
| ### cam_pose |
| cam_to_world rotion matrix <br> |
| |
| 16 numbers, 4x4 camera matrix |
| ### hand_pose |
| cam_to_hand <br> |
| 1 (whether annotate or not, 0: not annotate 1: annotate) + 21 * 3 (x, y, z in order) + 1 + 21 * 3 (right hand) <br> |
| First 64 numbers belong to the left hand. Next 64 numbers belong to the right hand |
| ### hand_pose_MANO |
| 1 (whether annotate or not, 0: not annotate 1: annotate) + 3 translation values + 48 pose values + 10 shape values + 1 + 3 + 48 + 10 (right hand) <br> |
| First 59 numbers belong to the left hand. Next 59 numbers belong to the right hand |
| ### obj_pose |
| cam_to_obj <br> |
| 1 (object class) + 21 * 3 (x, y, z in order) <br> |
| 21 numbers : 1 center, 8 corners, 12 mid edge point. <br> |
| 0 background (no object) <br> |
| 1 book <br> |
| 2 espresso <br> |
| 3 lotion <br> |
| 4 spray <br> |
| 5 milk <br> |
| 6 cocoa <br> |
| 7 chips <br> |
| 8 capuccino <br> |
| |
| <img id="image_canv" src="obj_point_order.jpg"/> |
| |
| ### object_pose_RT |
| 1 (object class) + 16 numbers, 4x4 camera matrix |
| |
| ### verb_label |
| 0 background (no verb) <br> |
| 1 grab <br> |
| 2 place <br> |
| 3 open <br> |
| 4 close <br> |
| 5 pour <br> |
| 6 take out <br> |
| 7 put in <br> |
| 8 apply <br> |
| 9 read <br> |
| 10 spray <br> |
| 11 squeeze <br> |
| |
| ### action_label |
| Combination of a noun (object class) and a verb (verb label) |
| |
| 0 background <br> |
| 1 grab book <br> |
| 2 grab espresso <br> |
| 3 grab lotion <br> |
| 4 grab spray <br> |
| 5 grab milk <br> |
| 6 grab cocoa <br> |
| 7 grab chips <br> |
| 8 grab cappuccino <br> |
| 9 place book <br> |
| 10 place espresso <br> |
| 11 place lotion <br> |
| 12 place spray <br> |
| 13 place milk <br> |
| 14 place cocoa <br> |
| 15 place chips <br> |
| 16 place cappuccino <br> |
| 17 open lotion <br> |
| 18 open milk <br> |
| 19 open chips <br> |
| 20 close lotion <br> |
| 21 close milk <br> |
| 22 close chips <br> |
| 23 pour milk <br> |
| 24 take out espresso <br> |
| 25 take out cocoa <br> |
| 26 take out chips <br> |
| 27 take out cappuccino <br> |
| 28 put in espresso <br> |
| 29 put in cocoa <br> |
| 30 put in cappuccino <br> |
| 31 apply lotion <br> |
| 32 apply spray <br> |
| 33 read book <br> |
| 34 read espresso <br> |
| 35 spray spray <br> |
| 36 squeeze lotion <br> |
| |
| ## For Actions |
| [Train set file](action_labels/action_train.txt) <br> |
| [Validation set file](action_labels/action_val.txt) <br> |
| [Test set file](action_labels/action_test.txt) <br> |
| |
| ## For Poses |
| [Train set file](pose_lists/pose_test.txt) <br> |
| [Validation set file](pose_lists/pose_train.txt) <br> |
| [Test set file](pose_lists/pose_test.txt) <br> |
| |
| ### cam_instrinsics.txt |
| six numbers : fx, fy, cx, cy, width, height |
| |
| # Citations |
| If you find any usefulness in this dataset, please consider citing: |
| ``` |
| @InProceedings{Kwon_2021_ICCV, |
| author = {Kwon, Taein and Tekin, Bugra and St\"uhmer, Jan and Bogo, Federica and Pollefeys, Marc}, |
| title = {H2O: Two Hands Manipulating Objects for First Person Interaction Recognition}, |
| booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)}, |
| month = {October}, |
| year = {2021}, |
| pages = {10138-10148} |
| } |
| ``` |
| |
