Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,116 @@
+---
+license: cc-by-nc-4.0
+task_categories:
+- robotics
+- image-to-3d
+language:
+- en
+tags:
+- vitra
+- mano
+- hand-object-interaction
+---
+
+# egotouch-annotations-v1
+
+**Annotations only — no images, no video.**
+
+VITRA-style hand episodes for **EgoTouch**, with per-hand instructions and paraphrases.
+
+| | |
+|---|---|
+| episodes | **111,159** |
+| training samples (`index_frame_pair` rows) | **3,687,389** |
+| 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.
+
+### Tactile
+
+**This is the only domain here that carries tactile data.** Each episode has an extra
+top-level `tactile` dict (not part of VITRA's original format):
+
+```
+d["tactile"]["left"]   ndarray (T, 21, 21) float32     # NaN outside the sensor area
+d["tactile"]["right"]  ndarray (T, 21, 21) float32
+```
+
+Row `t` of `tactile[hand]` is the pressure map for the **same frame** as row `t` of
+`joints_worldspace` and of `video_decode_frame` — i.e. tactile and pose are aligned
+frame-for-frame, no resampling.
+
+We verified that mapping end to end rather than assuming it:
+
+- the source release ships one pressure row per video frame (`frame_index` runs
+  `0..n-1`, strictly increasing, at exactly 30.00 fps) — checked on sampled scenes;
+- pressure length equals video frame count for **all 1,566 scenes x 2 hands**, with zero
+  exceptions, so nothing was silently truncated;
+- our stored values reproduce the official `pressure_grids.npz` at every frame; the
+  per-frame pressure totals cross-correlate at **lag 0 with r = 1.0000**, and a spatial
+  transpose does *not* match, ruling out both a time shift and an axis swap;
+- within an episode, the tactile slice is the same `[a:b]` window as the pose slice
+  (byte-identical on every sampled episode).
+
+Values are the official ones passed through **float16** at one intermediate stage, so they
+differ from the source by at most one float16 half-ULP (max observed 2.41e-4 on a
+0-1 normalised map). Positions of `NaN` are preserved exactly.
+
+### Files
+
+```
+egotouch.tar                 ->  Annotation/egotouch/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 egotouch.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/egotouch/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 EgoTouch — https://huggingface.co/datasets/zhouzhoujy/EgoTouch, 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.