README.md

---
license: cc-by-nc-4.0
task_categories:
- robotics
- image-to-3d
language:
- en
tags:
- vitra
- mano
- hand-object-interaction
---

# oakink2-annotations-v1

**Annotations only — no images, no video.**

VITRA-style hand episodes for **OakInk2**, with per-hand instructions and paraphrases.

| | |
|---|---|
| episodes | **29,058** |
| training samples (`index_frame_pair` rows) | **1,052,924** |
| annotation | MANO pose + world/camera joints + per-frame extrinsics |
| text | one instruction per episode + 1.93 paraphrases on average |
| **images / video** | **not included** — see *Getting the frames* below |

### What we did

**Episodes are ours.** The source release ships either raw video or differently-segmented
clips, so we re-cut it with VITRA's method — speed minima of the 3D wrist in world space:

    gaussian smooth (sigma=1.0) -> local speed minima in a fixed window (win=15, i.e. 0.5 s
    at 30 Hz) -> merge runs shorter than min_seg=16 -> pad 2 frames on each end

sigma and win are quantities in *time*, converted per source frame rate. Left and right hands
are cut independently, with the other hand's motion ignored.

**Instructions are ours.** Two rounds, both with `Qwen3.5-122B-A10B-FP8`:
round 1 captions 8 frames per episode with the palm's future trajectory drawn on them;
round 2 checks the sentence belongs to that hand, strips same-hand references
("Rinse the right hand." -> "Rinse the hand.", because training already prepends
`Left hand: ... Right hand: ...`), and writes 1-3 paraphrases.

**Episodes with no instruction are not included.** Round 1 returns `N/A` when an episode
shows no object interaction. Those episodes are excluded from both the archive and the
index, so every episode here has a usable instruction.

### Egocentric view only

OakInk2 records each sequence from four cameras — `allocentric_left`, `allocentric_right`,
`allocentric_top` and `egocentric`. **Only the egocentric view is here.** The three allocentric
views are the same sequences from a fixed third-person rig; we train on head-mounted footage,
so they are not included. Of 1,878 allocentric episodes, 1,875 have an egocentric counterpart
of the same sequence — dropping them loses the content of exactly one sequence.

Episodes are cut per hand, and a frame counts only when **the palm actually projects inside the
egocentric frame**. The source ships `kept_frames` and `annotation_valid` as all-`True`
(the pose comes from multi-view mocap, so it exists for every frame regardless of what the
head camera sees), and `visible` alone is not enough — it only guarantees that *some* joint is
in frame. Without the palm test, 5,243 of 45,281 segments would have had the hand outside the
picture for their entire length.

### Files

```
oakink2.tar                 ->  Annotation/oakink2/episodic_annotations/*.npy
episode_frame_index.npz        index_frame_pair (N,2) uint32 + index_to_episode_id (E,)
```

`index_frame_pair` **row number is the sample id**: row r = (episode ordinal, frame within
that episode). `len(index_frame_pair)` is the size of the training set.

```python
import numpy as np
# tar -xf oakink2.tar
z = np.load("episode_frame_index.npz", allow_pickle=True)
ep_slot, frame_id = z["index_frame_pair"][sample_id]
eid = str(z["index_to_episode_id"][ep_slot])
d = np.load(f"Annotation/oakink2/episodic_annotations/{eid}.npy", allow_pickle=True).item()
rgb_frame_id = int(d["video_decode_frame"][frame_id])
```

Each `.npy` is a dict with `video_name`, `video_decode_frame`, `intrinsics`,
per-frame `extrinsics` (world->camera), `anno_type` (which hand this episode is for),
`text`, `text_rephrase`, and a `left`/`right` dict holding `beta`, `hand_pose`,
`global_orient_worldspace`, `transl_worldspace`, `joints_worldspace`, `kept_frames`.
`text[hand] = [(sentence, (0, T))]` and `text_rephrase[hand] = [([paraphrases...], (0, T))]`.

### Getting the frames

`video_decode_frame` indexes the **source** video, which we do not redistribute.
Get it from OakInk2 — https://oakink.net/v2/, then decode by index (we use `decord`; a self-maintained
sequential counter drifts silently if the decoder ever skips a frame).

### Known limitations

- Paraphrase count averages 1.93, not a fixed number. Past 3 the model starts inventing;
  a sentence with no prepositional phrase honestly supports only one or two.
- Verified: the index lists exactly the episodes that have an instruction, every episode's
  stored frame count matches its index rows, and no index entry points at a missing episode.

---

## Revision — 2026-09-08 (re-cut)

**This release replaces the previous one. The previous episodes contained invalid frames
and should not be used.**

A defect in our episode-cutting step let frames with `kept_frames == False`
(invalid hand pose — all-zero or NaN wrist coordinates) stay inside published episodes.
The validity mask was only used to keep a cut point from landing on an invalid frame; it did
not constrain what a segment contained. Worse, a run of invalid frames could suppress cutting
altogether, so the gap was swallowed into one long segment instead of being excluded.

Every episode here is now built from a run of consecutive valid frames, so `kept_frames` is
all-`True` by construction — verified over the whole collection: **568,369 episodes /
16.5 M frames, zero `kept_frames == False`**. Both instruction rounds were regenerated for the
new segmentation.

Segment counts and episode ids therefore changed, and the index and splits were rebuilt:

| | previous | this release |
|---|---:|---:|
| episodes on disk | 37,691 | 38,444 |
| episodes published (with an instruction) | 28,264 | 29,058 |
| training samples | 1,371,721 | 1,052,924 |

Splits are `video_test` (test owns whole videos disjoint from train/val; train and val share
the remaining videos and are split at the episode level), balanced on **frames** at 90/5/5,
seed 1.

<!-- COLLECTION_TABLE_START -->
| dataset | episodes | training samples | our contribution | size | HF |
|---|---:|---:|---|---:|---|
| **EPIC-KITCHENS-100** | 149,570 | 4,019,534 | episodes + text | 8.70 GB | [`epic30-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/epic30-annotations-v1) |
| **EgoTouch** | 107,364 | 3,123,675 | episodes + text + **tactile** | 17.02 GB | [`egotouch-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/egotouch-annotations-v1) |
| **GigaHands** | 70,486 | 2,266,087 | episodes + text | 2.89 GB | [`gigahands-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/gigahands-annotations-v1) |
| **Ego-Exo4D** | 67,051 | 1,757,474 | text only | 4.09 GB | [`egoexo4d-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/egoexo4d-annotations-v1) |
| **OakInk2** | 29,058 | 1,052,924 | episodes + text | 1.56 GB | [`oakink2-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/oakink2-annotations-v1) |
| **TACO** | 23,757 | 736,136 | episodes + text | 1.34 GB | [`taco-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/taco-annotations-v1) |
| **HOT3D** | 18,805 | 619,680 | episodes + text | 1.51 GB | [`hot3d-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/hot3d-annotations-v1) |
| **ARCTIC** | 12,610 | 425,796 | episodes + text | 0.85 GB | [`arctic-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/arctic-annotations-v1) |
| **H2O** | 5,845 | 200,332 | episodes + text | 0.40 GB | [`h2o-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/h2o-annotations-v1) |
| **Tachin** | 2,669 | 79,584 | episodes + text + **tactile** | 3.38 GB | [`tachin-annotations-v1`](https://huggingface.co/datasets/MIT-Media-Lab/tachin-annotations-v1) |
| **total** | **487,215** | **14,281,222** | | **41.7 GB** | |

> **Something-Something V2 was dropped** from the collection (12 fps against
> 30 fps everywhere else, so a 16-step action chunk spans 1.33 s instead of 0.53 s).
> The repository still exists but should not be used.
<!-- COLLECTION_TABLE_END -->