ulid stringlengths 18 18 | task_hand stringclasses 2
values | object_phrase stringclasses 4
values | obj_px_frame0 int64 5.28k 18.6k | grasp_frame int64 110 288 | release_frame int64 156 434 | n_frames_scored int64 132 210 | box_trim_pct float64 1 1 | occlusion_pos float64 0 1 | occlusion_neg float64 0 0 |
|---|---|---|---|---|---|---|---|---|---|
01JMXF99BYFPWU5NOL | R | bagel | 7,451 | 194 | 322 | 149 | 1 | 0.287478 | 0 |
01JMXGMVZPODEGWHV5 | R | croissant | 11,093 | 116 | 250 | 134 | 1 | 0.425854 | 0 |
01JMXGPPFC2ZRTLNWH | R | croissant | 9,989 | 140 | 268 | 132 | 1 | 0.188407 | 0 |
01JMXGQEAGFX55VOQT | L | croissant | 15,964 | 122 | 258 | 151 | 1 | 0.558256 | 0 |
01JMXGR8GTQVLVREEX | L | croissant | 8,369 | 212 | 390 | 189 | 1 | 0.481659 | 0 |
01JMXGS9NVVF5XPUDQ | L | croissant | 16,542 | 148 | 322 | 157 | 1 | 0.615585 | 0 |
01JMXGT691NQAI4UGX | L | croissant | 10,002 | 144 | 304 | 154 | 1 | 0.436313 | 0 |
01JMXGW40SVV2DKJVY | R | croissant | 17,943 | 176 | 340 | 153 | 1 | 0.392521 | 0 |
01JMXGX388EP67QCGO | R | croissant | 8,710 | 162 | 296 | 138 | 1 | 0.445121 | 0 |
01JMXGXY7MKHTF37DJ | R | croissant | 10,265 | 162 | 284 | 134 | 1 | 0.524208 | 0 |
01JMXGYNQNIYVIKPAK | R | croissant | 5,284 | 160 | 310 | 145 | 1 | 0.108062 | 0 |
01JMXGZNHTYMUPD77N | R | croissant | 9,824 | 204 | 330 | 151 | 1 | 0.138029 | 0 |
01JMXH0JZ2UM6QEJFZ | L | croissant | 15,905 | 158 | 302 | 150 | 1 | 0.745677 | 0 |
01JMXH39SZD6SNVGKZ | L | croissant | 12,778 | 132 | 262 | 140 | 1 | 0.577399 | 0 |
01JMXH44Z0CAPNWQGL | L | croissant | 16,955 | 124 | 262 | 136 | 1 | 0.429372 | 0 |
01JMXH50JR4AAJWETN | L | croissant | 17,826 | 140 | 268 | 134 | 1 | 0.496578 | 0 |
01JMXH6MCHZJLL46NL | L | croissant | 10,505 | 196 | 328 | 157 | 1 | 0.245597 | 0 |
01JMXH7M1DAFQ5ZAWF | R | croissant | 5,548 | 184 | 330 | 152 | 1 | 0.381038 | 0 |
01JMXH8F3YCAGKVL6D | R | croissant | 16,804 | 154 | 316 | 149 | 1 | 0.46239 | 0 |
01JMXH9DK2YOAL5T5B | L | croissant | 11,822 | 130 | 266 | 134 | 1 | 0.925309 | 0 |
01JMXHB4XXYG2ZBSR4 | L | croissant | 9,261 | 188 | 344 | 155 | 1 | 0.677573 | 0 |
01JMXHC55J2L4AQXR4 | L | croissant | 16,941 | 130 | 280 | 139 | 1 | 0.529957 | 0 |
01JMXHD0RABKJ5RDEL | R | croissant | 7,568 | 158 | 280 | 132 | 1 | 0 | 0 |
01JMXHDWD4JHEF3RMU | R | croissant | 16,059 | 114 | 270 | 137 | 1 | 0.457999 | 0 |
01JMXHFJNGEK7ODZIQ | R | croissant | 15,343 | 122 | 270 | 143 | 1 | 0.463729 | 0 |
01JMXHGFJ8KRVXBW6O | L | croissant | 12,685 | 148 | 276 | 139 | 1 | 0.402759 | 0 |
01JMXHJAJFQFQODPYS | L | croissant | 8,615 | 178 | 342 | 168 | 1 | 0.570517 | 0 |
01JMXHK95EE35IFEXL | L | croissant | 17,515 | 154 | 308 | 172 | 1 | 0.631858 | 0 |
01JMXHM95FBDOCPX2L | L | croissant | 17,735 | 138 | 264 | 138 | 1 | 0.466761 | 0 |
01JMXHP1M4VC5AVYM6 | L | croissant | 10,582 | 168 | 296 | 145 | 1 | 0.553109 | 0 |
01JMXHPXJZG46CDQM5 | R | croissant | 15,119 | 126 | 246 | 134 | 1 | 0.13784 | 0 |
01JMXHRMJMMEN2ZKO5 | R | croissant | 8,501 | 186 | 312 | 145 | 1 | 0.199388 | 0 |
01JMXHSH5GGRR4ZWBM | R | croissant | 17,725 | 138 | 276 | 134 | 1 | 0.433343 | 0 |
01JMXHT9J86FA7GEJB | L | croissant | 13,765 | 126 | 284 | 150 | 1 | 0.570795 | 0 |
01JMXHV4WQ45DIDTZG | L | croissant | 9,365 | 194 | 332 | 168 | 1 | 0.233422 | 0 |
01JMXHW9V3XC6SBX7R | L | croissant | 18,464 | 134 | 262 | 151 | 1 | 0.511319 | 0 |
01JMXHX7PMRRVTZI73 | L | croissant | 11,654 | 186 | 326 | 155 | 1 | 0.331989 | 0 |
01JMXHY52YOGZKZEJB | L | croissant | 10,484 | 190 | 340 | 158 | 1 | 0.166349 | 0 |
01JMXHZ5XDADBGASM5 | R | croissant | 9,082 | 200 | 338 | 153 | 1 | 0 | 0 |
01JMXJ04M1ELEJ73K5 | R | croissant | 16,950 | 136 | 274 | 135 | 1 | 0.484897 | 0 |
01JMXJ0W2GY7FRZJ5Z | L | croissant | 12,517 | 138 | 312 | 156 | 1 | 0.489574 | 0 |
01JMXJ1QE1JENE3VYV | L | croissant | 15,037 | 132 | 274 | 158 | 1 | 0.265744 | 0 |
01JN0A85HBTIASV5LM | R | croissant | 10,053 | 150 | 356 | 195 | 1 | 0.451706 | 0 |
01JN0A91NCBEM2TXLL | R | croissant | 10,484 | 122 | 374 | 200 | 1 | 0.394315 | 0 |
01JN0A9WPETY4J25V5 | R | croissant | 10,208 | 178 | 434 | 198 | 1 | 0.330525 | 0 |
01JN0AAQZZJ4W2RMJS | R | croissant | 9,836 | 172 | 422 | 210 | 1 | 0.324319 | 0 |
01JN0ABKHEIDL5PFHJ | R | croissant | 9,997 | 140 | 386 | 192 | 1 | 0.151746 | 0 |
01JN0ACD62XX5SQMDB | R | croissant | 10,332 | 132 | 402 | 187 | 1 | 0.335172 | 0 |
01JN0B1N7DIF2OCR2Z | R | croissant | 12,245 | 178 | 326 | 159 | 1 | 0.512781 | 0 |
01JN0B2BNZFXLIMPNY | R | croissant | 12,081 | 162 | 328 | 165 | 1 | 0.479348 | 0 |
01JN0B35F9DZ6T4NDH | R | croissant | 13,771 | 140 | 310 | 165 | 1 | 0.486021 | 0 |
01JN0B3VETIMF4BRK7 | R | croissant | 11,050 | 136 | 284 | 153 | 1 | 0.510317 | 0 |
01JN0B4HRRYY52AGBE | R | croissant | 11,591 | 142 | 310 | 159 | 1 | 0.510569 | 0 |
01JN0B59XC5PAHKPR2 | R | croissant | 12,241 | 156 | 328 | 162 | 1 | 0.501348 | 0 |
01JN0B6144I3JRR245 | R | croissant | 13,052 | 154 | 326 | 172 | 1 | 0.472724 | 0 |
01JN0B6SM7ETEVDWLM | R | croissant | 10,663 | 152 | 320 | 161 | 1 | 0.473038 | 0 |
01JN0B7MWN7VS3RR7Q | R | croissant | 12,153 | 134 | 294 | 153 | 1 | 0.506459 | 0 |
01JN0B8G8VT6WWBBEV | R | croissant | 12,142 | 158 | 334 | 171 | 1 | 0.408005 | 0 |
01JN0B981DR4HI32AT | R | croissant | 13,091 | 142 | 292 | 141 | 1 | 0.488198 | 0 |
01JN0B9XZASBEFU6PQ | R | croissant | 12,118 | 144 | 296 | 151 | 1 | 0.523436 | 0 |
01JN0BAMK9C6VGFKV2 | R | croissant | 12,765 | 156 | 298 | 160 | 1 | 0.559812 | 0 |
01JN0BBBYQSYO4JUYW | R | croissant | 11,866 | 140 | 304 | 160 | 1 | 0.502107 | 0 |
01JN0BC2Z6ZAXRWIPI | R | croissant | 11,620 | 140 | 316 | 160 | 1 | 0.486833 | 0 |
01JN0BCTN2BA76U4D3 | R | croissant | 12,227 | 148 | 310 | 160 | 1 | 0.546495 | 0 |
01JN0BDHGV7EESL4NM | R | croissant | 12,196 | 126 | 284 | 148 | 1 | 0.404969 | 0 |
01JN0BE8GCDGAXDSCI | R | croissant | 11,733 | 146 | 290 | 154 | 1 | 0.418648 | 0 |
01JN0BEYVX3GS4FX5B | R | croissant | 11,064 | 146 | 274 | 140 | 1 | 0.499367 | 0 |
01JN0BFMVDSCEZRCOC | R | croissant | 12,609 | 116 | 264 | 137 | 1 | 0.35086 | 0 |
01JN0BG9G1WQIU236S | R | croissant | 12,216 | 110 | 270 | 150 | 1 | 0.26981 | 0 |
01JN0KMMEED4B6XL2R | L | donut | 7,864 | 182 | 306 | 150 | 1 | 0.999746 | 0 |
01JN0KNC6ENBIIANCP | R | bagel | 8,764 | 142 | 270 | 149 | 1 | 0.306481 | 0 |
01JN0XAGCRDDJ33ML4 | R | croissant | 9,820 | 208 | 366 | 161 | 1 | 0.21334 | 0 |
01JN0XB99NMRGSN34B | L | croissant | 10,165 | 224 | 390 | 171 | 1 | 0.451648 | 0 |
01JN0XC8EEQBMRX73C | R | croissant | 9,764 | 202 | 332 | 156 | 1 | 0.225113 | 0 |
01JN0XDNHPW3MSHRCZ | R | bread | 13,179 | 150 | 242 | 133 | 1 | 0.272934 | 0 |
01JN0XEAG7A6NDMS7R | R | croissant | 12,010 | 144 | 252 | 138 | 1 | 0.318235 | 0 |
01JN0XF11CKS6YENQJ | L | croissant | 13,196 | 148 | 156 | 153 | 1 | 0.271143 | 0 |
01JN0XFRTCSU3HUYBK | R | croissant | 9,781 | 166 | 336 | 160 | 1 | 0.265924 | 0 |
01JN0XGJ9NIQQDTF3O | R | croissant | 11,207 | 236 | 388 | 176 | 1 | 0.236727 | 0 |
01JN0XHCY4E6E6HE5Z | L | croissant | 13,919 | 188 | 354 | 168 | 1 | 0.312307 | 0 |
01JN0XJ7Q8CH7OG5XN | L | bread | 12,130 | 182 | 350 | 168 | 1 | 0.178318 | 0 |
01JN0XK2P5SFFTTIJU | R | croissant | 11,282 | 164 | 312 | 165 | 1 | 0.380429 | 0 |
01JN0XKY46ORUF2SQ7 | R | croissant | 11,300 | 118 | 272 | 149 | 1 | 0.435398 | 0 |
01JN0XMMZRPUNRCL4Q | R | croissant | 10,993 | 204 | 362 | 168 | 1 | 0.288911 | 0 |
01JN0XNG2226HCR27P | R | bread | 12,082 | 170 | 344 | 164 | 1 | 0.306406 | 0 |
01JN0XP8ABNHWX4ZRP | L | croissant | 15,058 | 202 | 368 | 179 | 1 | 0.234028 | 0 |
01JN0XQ2MFKJ3F6WCW | R | croissant | 11,371 | 168 | 354 | 171 | 1 | 0.373934 | 0 |
01JN0XQYZ75FPQ4GIQ | L | croissant | 17,496 | 288 | 428 | 187 | 1 | 0.64249 | 0 |
01JN0XSNAK5A6GJF76 | L | croissant | 18,606 | 204 | 344 | 154 | 1 | 0.499194 | 0 |
01JN0XTD94FJISJQBQ | R | croissant | 15,572 | 146 | 306 | 152 | 1 | 0.125225 | 0 |
01JN0XV4GT5TNULAUJ | R | croissant | 14,303 | 178 | 334 | 174 | 1 | 0.198769 | 0 |
01JN0XVX6376EU4HK6 | L | croissant | 16,274 | 188 | 194 | 201 | 1 | 0.468047 | 0 |
01JN0XWS9V2Z56BCFO | L | croissant | 16,609 | 168 | 298 | 159 | 1 | 0.282136 | 0 |
01JN0XXHP8WDWQIHB3 | L | croissant | 10,045 | 180 | 352 | 159 | 1 | 0.540866 | 0 |
01JN0XYBN8F6CWOU7K | R | croissant | 11,010 | 172 | 316 | 158 | 1 | 0.27257 | 0 |
01JN0XZ4VWVGAZWEOQ | R | croissant | 11,242 | 202 | 378 | 173 | 1 | 0.371909 | 0 |
01JN0XZYE15R6DF3VW | L | croissant | 13,653 | 158 | 338 | 161 | 1 | 0.467004 | 0 |
01JN0Y0R46ZBRGEFQI | L | croissant | 9,904 | 204 | 404 | 170 | 1 | 0.446587 | 0 |
01JN0Y1K3SR7N5LYF5 | R | croissant | 11,010 | 190 | 328 | 163 | 1 | 0.234605 | 0 |
01JN0Y2AWRMLCMJWDP | R | croissant | 11,556 | 154 | 314 | 152 | 1 | 0.244116 | 0 |
- Attribution and license
- Why this subset exists
- What is corrected here (2026-08-20)
- F0. What the filter decides
- F1. What is in here
- F2. The scene, and why it matters
- F3. How to apply Pre-Contact filtering
- F4. What came out on this set
- F5. Reproducing the verdicts
- F6. Known limitation: the boundary sphere absorbs depth leak, and here you cannot see it
- ์ ์ 2026-08-20 (๋น์ผ ๊ฐฑ์ ) โ
derived/states๋ฅผ ๊ต์ฒดํ์ต๋๋ค - 1. ๊ฒฝ๋ก
- 2. tar ์์ ๋ฌด์์ด ์๋๊ฐ
- 3. ์ด 100๊ฐ๊ฐ ๋ฌด์์ธ๊ฐ โ ํฌ๋ฃจ์์ ๊ณผ์ ๋ก ์๋์ ์ผ๋ก ๊ณ ์ ํ์ต๋๋ค
- 4. depth๋ฅผ ์ธ ๋ ์์์ผ ํ๋ ๊ฒ (์ธก์ ์ผ๋ก ํ์ธํ ์ฌ์ค)
- 5.
derived/โ ์ฐ๋ฆฌ๊ฐ ๋ง๋ ํ์ ์ฐ์ถ๋ฌผ (์งํ ์ค) - 6. ์ฐ๋ ๋ฒ
- 7. ์ฌํยท๊ฒ์ฆ
- ์ถ๊ฐ 2026-08-21 โ
derived/states_tau/์ ์ผํ ๋ณด์ ๊ฐ
ActionNet subset100 with ground-truth depth
This is a 100-episode subset redistribution of a third-party dataset, plus our derived artifacts. Read the attribution before using it.
Attribution and license
| Upstream dataset | FourierIntelligence/ActionNet (Fourier Intelligence) |
| What is redistributed | 100 episodes out of 30,121, byte-identical to the upstream tars: rgb.mp4 (1280x800 fisheye), depth.mkv (lossless 16-bit), timestamps.json, <ULID>.hdf5 |
| Upstream terms | Whatever the upstream repository states. The upstream terms govern the episode payload; if they disallow redistribution, tell us and this repo comes down. |
| Ours (CC-BY-4.0) | Everything under derived/, the mapping tables, build_subset100.py, and this document |
| Robot | Fourier GR1-T1, 6-DoF Fourier DexHands, single head-mounted camera (top) |
Why this subset exists
The widely used LeRobot conversion (lerobot/action_net) drops depth and downscales RGB to
192x288, and its observation.state is indexed at the robot's 60 Hz while its rows are
labelled at the camera's 30 Hz โ so image and state drift apart by up to half an episode.
Both problems are documented, with measurements, below. This subset keeps the original
1280x800 RGB, the lossless ground-truth depth, and a corrected per-video-frame state
built from the original hdf5.
What is corrected here (2026-08-20)
| Measured | |
|---|---|
observation.robot_joints[k] vs hdf5 /state/robot[k] |
within 1e-5 on 97/100 episodes (median max abs diff 5.95e-08); the same tar-independent check passes 21/21 on a second tar |
the documented [2k] rule |
0/100, median max abs diff 0.740 rad; and in 18/21 episodes of the second tar 2k runs past the end of the hdf5 |
FK of our corrected states vs the dataset's own /state/pose |
left EEF median 0.29 cm, right 0.27 cm, head 0.59 cm, over 3 cm on 0/100 |
| the same check with states indexed by video frame | the moving arm sits median 7.59 cm off (range 2.96-18.66), 99/100 over 3 cm; the idle arm stays at 0.33 cm, which is why this hid for so long |
Pre-Contact Level Filtering โ the inputs, the verdicts, and how to apply it
Added 2026-08-24. Everything above is about the capture: what was redistributed, what was corrected, and what the depth is. This part is about what we then did with it -- the filter that decides whether a video is consistent with the action stream that goes with it -- and it ships the inputs and verdicts so the filter can be re-run rather than taken on trust.
The top-level directories are new and are the authoritative ones. derived/ is the earlier
partial staging and is kept for the artifacts that did not move: derived/points/,
derived/hand_points2/, derived/hand_tracks3/, and derived/states/ (the un-tau-corrected
states). Two of its folders were removed rather than left to be picked by accident --
derived/masks/ held only 43 of the 100, and derived/states_tau/ was byte-identical to the new
states/ (all 100 checked by md5); that folder's evidence note moved to states/README.md.
F0. What the filter decides
Given a video and the action stream that goes with it, is the video consistent with those actions? The question exists because generated video is used as training data: a clip whose pixels show an arm going somewhere its actions never went teaches a policy the wrong thing.
The filter answers in two levels, and both must pass:
| level | question | space | what it can see |
|---|---|---|---|
| Pre-Contact | did the acting fingertip ever reach the target object's boundary? | 1280x800 pinhole | one state, the action stream, the frame-0 object. Never the pixels. |
| Occlusion | at that instant, does the arm the video draws actually cover the object? | the video's own resolution | the pixels |
The split is deliberate. Pre-Contact is a geometric screen computed from actions alone, so it cannot be fooled by a rendering. Occlusion is the pixel check, and it is only asked at the instant Pre-Contact names. Neither is a contact test; together they are a screen.
This set is the reference, not the target. Every episode here is a real teleoperated
demonstration, so the arm does reach the object. Running the filter here answers does it destroy
good data? -- the answer is 98/100 -- and calibrates the constants. The set it exists to judge is
generated video: see
glory-hyeok/robocurate-synth100.
F1. What is in here
manifest.csv one row per episode: ulid, task_hand, object_phrase,
obj_px_frame0, grasp/release frame, occlusion_pos, occlusion_neg
camera.json K_pinhole, K_fisheye, D, and the space definitions
gtdepth/depth_<U>.npz GROUND-TRUTH depth, already undistorted + downsampled:
depth (T, 400, 640) float16, metres
frame_indices (T,) which video frames these are
K_ds (3,3) the intrinsics for THIS array
ds () the downsample factor (2)
K_full (3,3) the 1280x800 pinhole K
zero_frac, sat_frac per-frame sensor quality
masks/object_masks_<U>.npz target-object masks, (T, 800, 1280) bool + frame_indices,
n_px, object_phrase, method, n_obj
states/states_<U>.npz states (T,44) + actions (T,44) + frame_indices, tau-corrected
handmasks/<U>_{L,R}.npz tracked arm masks (24 episodes -- the sampled subset)
world3d/world3d_<U>.json the Pre-Contact verdict, per episode + per frame
occlusion/occlusion_full.json the Occlusion verdict, all 100 in one table
occlusion/occlusion_series.json the per-frame score inside the window
Also in the repo, from an earlier partial staging: derived/points/ (VLM object points),
derived/hand_points2/ and derived/hand_tracks3/ (10 episodes each), derived/states/ (the
un-tau-corrected states), and subset100_gtdepth.tar. The top-level directories above are the
authoritative ones -- derived/masks/ held only 43 of the 100 and derived/states_tau/ was a
byte-identical copy of states/, so both were removed rather than left to be picked by accident.
gtdepth/ is multi-frame, but the filter only uses frame 0 -- the object is not moved before it
is touched, so its boundary is built once and held (static_box: true, box_frame: 0 in every
verdict). The rest of the frames ship because they are what a depth model gets scored against.
F2. The scene, and why it matters
Every episode has n_obj = 1: a single pastry on a plate, on a striped tablecloth. The masks
are ours, from a SAM 3 text prompt (croissant x94, bread x3, bagel x2, donut x1), keeping
the largest connected component inside an area band.
That single-object scene is doing quiet work, and it is the main reason numbers measured here are looser than what a cluttered scene needs. When a mask's border bleeds a pixel past the object, here it lands on the tablecloth immediately beside it -- at nearly the pastry's own depth. In a multi-object scene the same one-pixel error lands on a neighbouring object tens of centimetres away, and the lifted point cloud gains a second cluster that stretches the boundary. See ยงF6.
F3. How to apply Pre-Contact filtering
F3.1 What you need beyond this repo
| robot model | GR1T1_fourier_hand_6dof.urdf, from FFTAI/teleoperation @main:assets/urdf/, md5 78ec53140a125ef50c3a8701249fb036 (not redistributed here) |
| forward kinematics | any URDF FK that returns the world pose of every link |
The 44-channel state layout is ActionNet's own. Channels 0-6 are the left arm, 22-28 the right;
the 6 hand slots per side are declared in configs/hand/fourier.yaml order
(pinky, ring, middle, index, thumb_pitch, thumb_yaw) and are DexHand raw units in [0, 10.3],
not radians, on this set.
F3.2 The trajectory
Roll the state forward from the episode's first state using the recorded actions, subsampled to
15 Hz (actions[::2]), with a first-order tracking model:
q[t+1] = q[t] + beta * (q[t] - q[t-1]) + alpha * (a[t] - q[t])
beta = 0.0 alpha = 0.45 (body) alpha = 0.60 (hand)
Then add two offset layers, in this order:
rollout_offsets-- both arms and the waist/head chain, largest term 4.29 deg. These correct the controller's standing bias, which the rollout inherits.left_arm_offset-- the left arm only: shoulder pitch 0.40, roll -0.27, yaw 4.04, elbow pitch 7.21 deg. This is a left-arm encoder zero bias, not a symmetric fudge: under it, 9 of 10 fit episodes reach 0.26-1.12 px reprojection error, which is the level the right arm already reaches with no offset at all. Do not mirror it to the right arm. It moves the left fingertip by 6.11 cm, so getting this wrong on a left-handed episode is larger than the object.
Both files are in the repo root alongside this card.
F3.3 The object boundary
At frame 0 only:
z = np.load(f"gtdepth/depth_{u}.npz")
D = z["depth"][list(z["frame_indices"]).index(0)].astype(float) # (400, 640) metres
Kd, ds = z["K_ds"], int(z["ds"])
m = np.load(f"masks/object_masks_{u}.npz")
M = m["masks"][list(m["frame_indices"]).index(0)][::ds, ::ds] # match depth's grid
ok = M & (D > 0.05) & (D < 2.6) # sensor range gate
v, u_ = np.nonzero(ok); zc = D[v, u_]
lo, hi = np.percentile(zc, [5, 95]) # drop depth outliers
k = (zc >= lo) & (zc <= hi); v, u_, zc = v[k], u_[k], zc[k]
P_cam = np.stack([(u_ - Kd[0,2]) / Kd[0,0] * zc,
(v - Kd[1,2]) / Kd[1,1] * zc, zc], 1)
P_base = (base_T_cam[:3,:3] @ P_cam.T).T + base_T_cam[:3,3]
lo3 = np.percentile(P_base, 1, axis=0) # 1-99 per axis, NOT min/max
hi3 = np.percentile(P_base, 99, axis=0)
centre = (lo3 + hi3) / 2
radius = np.linalg.norm(hi3 - lo3) / 2 # circumscribed sphere of the AABB
base_T_cam = FK(q0)["head_pitch_link"] @ head_pitch_T_cam, with the extrinsic from
align_camera_hullfit_20260820.npz in the repo root.
The boundary is the circumscribed sphere of the trimmed box, not the box. The sphere is 4.2x the volume, which is the point: Pre-Contact is a screen, and a screen should be generous. On this set the median radius is 8.15 cm, from an AABB of 13.1 x 7.0 x 7.1 cm -- genuinely a croissant.
F3.4 The distance, and the event
Per frame, over the five *_tip_link fingertips of the task hand:
d = max(0.0, min(norm(FK(q_t)[tip][:3,3] - centre) for tip in tips) - radius)
d == 0 means inside. Pre-Contact is an EVENT, not a state: the entry frame is the first frame
at d == 0 having been outside on the frame before.
if entry is None and prev is not None and d <= 0 < prev:
entry = t
Frame 0 therefore can never qualify -- there is no frame before it to have been outside on. That sounds like it should cost something, because 24 of 100 episodes already start inside the sphere: the demonstration often begins with the hand near the pastry. Measured, it costs nothing. All 24 also produce a crossing later -- the hand leaves and comes back, which is what picking a thing up and putting it down looks like -- so 100 of 100 episodes here have an entry event.
The same holds on the generated set: 21 of 300 clips start inside (7 per method, and exactly 7 because all three share the same 100 conditioning frames, so frame 0 does not depend on the generator), and all 21 get an entry event anyway. Nothing is lost to the frame-0 rule in either set. Clips with no entry are clips whose fingertip never reached the boundary at all, which is the signal the filter exists to find.
F3.5 Occlusion, at that instant only
Window: +-1 s around the entry frame, i.e. +-30 frames at this set's 30 fps.
obj = object_mask_at_frame_0 # static: nothing moves it before contact
occ = max(|arm_mask[f] & obj| / |obj| for f in window)
keep = occ >= tol
The denominator is the object, not the union. Measured on the 24 episodes with tracked arm masks, the arm mask is 3.2x the object at the median (range 0.5-10.1x), so on a clean grasp that fully covers the object, IoU reads about 0.31 where coverage reads 1.0 -- it understates by roughly the size ratio, and understates most in exactly the frames where the arm is closest. Coverage has no such bias, and it is also the quantity the question asks for: what fraction of the object is covered.
F3.6 The threshold, and why this set cannot set it
| tol | task hand passes | idle hand falsely passes |
|---|---|---|
| 0.1 % | 98 / 100 | 0 / 100 |
| 1 % | 98 / 100 | 0 / 100 |
| 5 % | 98 / 100 | 0 / 100 |
The task hand's occlusion is 43.1 % at the median (10th percentile 19.8 %); the idle hand --
the negative control, the arm that never moves -- reaches 0.00 % at its maximum. Separation is
total, so the threshold is unidentifiable here: three values two orders of magnitude apart give
the same answer. It has to be set on generated video, where the scores are lower and the control is
much harder (an arm that genuinely moves and grasps, just not the instructed object). Treat any
tol in this range as provisional.
F4. What came out on this set
| episodes | 100 |
| Pre-Contact entry found | 76 (24 start inside the sphere at frame 0 -- no observable event) |
| Occlusion, task hand | median 43.1 %, p10 19.8 % |
| Occlusion, idle hand (control) | median 0.00 %, max 0.00 % |
| both levels pass | 98 of the 100 that were scored |
| boundary sphere | median 8.15 cm |
98/100 is the answer to does the filter destroy good data -- and that is all a set of real demonstrations can tell you. It cannot tell you whether the filter catches bad data, because there is no bad data here. That requires generated video.
F5. Reproducing the verdicts
world3d/world3d_<U>.json records the settings each verdict was produced with, so a
reimplementation can be checked against them rather than trusted:
{"ulid": "...", "task_hand": "R", "static_box": true, "box_frame": 0,
"box_trim_pct": 1.0, "boundary": "sphere", "grasp_frame": 194,
"grasp_frame_recorded": 191, "grasp_source": "rollout",
"tip_to_box_at_grasp_cm": 0.0, "curve_min_cm": 0.0, "n_frames_scored": 149}
grasp_source: "rollout" is worth noting: the verdict is computed from the rolled-out
trajectory, not the recorded states, so that this set and the generated set are judged by the same
machinery. grasp_frame_recorded is what the recorded states say, kept for comparison -- the two
agree to a few frames.
F6. Known limitation: the boundary sphere absorbs depth leak, and here you cannot see it
The mask's border bleeds a little past the object; the lifted points then include a few at the wrong depth. On this set that is nearly harmless, and the reason is worth stating because it does not transfer:
- The croissant's AABB is 13.1 x 7.0 x 7.1 cm. Its longest axis is real, and it dominates the half-diagonal -- depth is only 43 % of it. Adding 5 cm of leak to the depth axis grows the sphere 24 %.
- On the generated set the object is a block, 5.8 x 5.0 x 3.2 cm -- nearly isotropic, and depth is its shortest true axis, so it has the most room to be corrupted and takes over the diagonal the moment it is. The same 5 cm grows that sphere 42 %.
- Add the scene: here a border leak lands on the tablecloth at the pastry's own depth; there it lands on a neighbouring object, which is a large jump.
Measured on the generated set, clips whose object-region depth spans under 5 cm give a 4.02 cm sphere while those spanning over 15 cm give 11.46 cm -- same object class, so what varies is the measurement, not the object. The fix is to select depth as a band around the object-region median rather than by percentile; the percentile trim cannot remove a leak larger than 5 % of the mask, and on a small object the leak is 10-20 %.
Nothing was changed here. The verdicts in world3d/ use the 5-95 percentile trim, and on this
set the difference is small enough not to matter. Anyone porting these constants to a cluttered
scene or a smaller object should re-derive them first.
ActionNet subset100 โ GT depth ํฌํจ 100 ์ํผ์๋
ํ ์ค ์์ฝ: LeRobot ๋ณํ๋ณธ์๋ depth๊ฐ ์๊ณ RGB๋ 192ร288๋ก ์ค์ด ์์ด์, ์๋ณธ tar์์ 1280ร800 RGB + ๋ฌด์์ค GT depth๋ฅผ ๊ฐ์ง 100 ์ํผ์๋๋ง ๊ณจ๋ผ ์๋ฆฝํ์ผ๋ก ๋ฌถ์ ๊ฒ์ ๋๋ค.
์์ฑ 2026-08-20. ์์ ๋๋ ํ ๋ฆฌ์ README_MAPPING.md(๋ด๋ถ ๋ฌธ์, ๊ณต๊ฐ๋์ง ์์)๊ฐ ์๋ณธ โ LeRobot
๋์์ ๊ทผ๊ฑฐ ๋ฌธ์์ด๊ณ , ์ด ๋ฌธ์๋ ๊ทธ์ค 100๊ฐ ๋ถ๋ถ์งํฉ๊ณผ ํ์ ์ฐ์ถ๋ฌผ๋ง ๋ค๋ฃน๋๋ค.
์ ์ 2026-08-20 (๋น์ผ ๊ฐฑ์ ) โ derived/states๋ฅผ ๊ต์ฒดํ์ต๋๋ค
| ๋ฌด์ | ์ด์ | ์ง๊ธ |
|---|---|---|
derived/states/ |
LeRobot parquet์์ ๋ฝ์ stride 4๋ก ์ ๋ ฌ | ์๋ณธ hdf5๋ฅผ ์์์ timestamps.json ์๊ณ์ ์ต๊ทผ์ ๋ฆฌ์ํ |
| ์ด์ ํ์ ์์น | โ | derived/states_lerobot_WRONG_do_not_use/ (์ง์ฐ์ง ์๊ณ ๋จ๊น) |
| LeRobot ๋ณํ๋ณธ fps (ยง4 ํ) | "15 fps" | 30 fps, ํ๋ ์ ์๊ฐ ์๋ณธ๊ณผ ๋์ผ (447/432/491 ์ค์ธก, meta/info.json fps: 30) |
์์ธ์ ์์ README_MAPPING.md ยง4(๋ด๋ถ ๋ฌธ์)์ ์ ๋ ฌ ์ค๋ฅ์
๋๋ค โ LeRobot์ state ํ์ 30 Hz๊ฐ ์๋๋ผ
60 Hz ์ธ๋ฑ์ค์
๋๋ค. ๊ทธ๋์ ์์ ํ๋ ์ ๋ฒํธ๋ก state๋ฅผ ์ธ๋ฑ์ฑํ ์ฐ์ถ๋ฌผ์ ์ ๋ถ ๋ค์
๋ง๋ค์ด์ผ ํฉ๋๋ค(์์ง์ด๋ ํ FK๊ฐ /state/pose์์ 100 ep ์ค์ 7.59 cm ๋ฒ์ด๋จ,
์ต๋ 18.66 cm, 99/100 ep๊ฐ 3 cm ์ด๊ณผ. ์ ์งํ ํ์ 0.33 cm๋ผ ํ์ชฝ๋ง ๋ณด๋ฉด ์ ๋ณด์).
๋ง์คํฌยทpointsยทtracks๋ ์ํฅ ์์ต๋๋ค โ ์์ ํ๋ ์์์ ์ง์ ๋ง๋ ๊ฒ์ด๊ณ , ์ derived/states์
frame_indices๋ ๊ฐ์ ์๋ณธ ๋น๋์ค ํ๋ ์ ์ถ์ด๋ผ ๊ทธ๋๋ก ๋์ํฉ๋๋ค.
ยง5์ ์ถ๊ฐ๋ ์ฌ์ค ํ๋ ๋: ๊ณต์ ๋ง์คํฌ๋ ์๋ณธ ์ด์ ์์ ์ขํ๊ณ๊ฐ ์๋๋๋ค โ ์ธ๋์คํ ํธ๋
pinhole ๋ ๋๋ง ์์์ ๊ณ์ฐ๋์ต๋๋ค. raw rgb.mp4์ ๊ทธ๋๋ก ์น์ผ๋ฉด ์ฃผ๋ณ๋ถ์์ 100 px ์ด์
ํ์ด์ง๋๋ค. ยง5 ์ฐธ์กฐ.
1. ๊ฒฝ๋ก
๊ฒฝ๋ก ํ๊ธฐ ์ฃผ์: ์๋ ํ๋ ์ด ํจํค์ง๊ฐ ์ฐ๋ฆฌ ํด๋ฌ์คํฐ์ ๋์ธ ์์น์ด๊ณ , ์ธ๋ถ ๋ ์์๊ฒ๋ ์๋ฏธ๊ฐ ์์ต๋๋ค. HF ์ ์ฅ์์์๋ ๊ฐ์ ๊ฒ๋ค์ด ์ ์ฅ์ ๋ฃจํธ์ ์๊ณ , ํ์ ์ฐ์ถ๋ฌผ์
derived/๊ฐ ์๋๋ผ ์ต์์gtdepth/ masks/ states/ handmasks/ world3d/ occlusion/์ ๋๋ค.
| ๋ฌด์ | ๊ฒฝ๋ก (๋ด๋ถ) |
|---|---|
| ์ด ๋๋ ํ ๋ฆฌ | <๋ด๋ถ ์ํฐํฉํธ ๋ง์ดํธ>/action_net_og/subset100_gtdepth/ |
| โ ํ์ด๋ก๋ tar (100 ep, ~4.4 GB) | subset100_gtdepth.tar |
| โ ๋ช ๋จ (ํ ์ค์ ULID ํ๋) | ulids.txt |
| โ ๋์ํ | mapping_subset100.csv / .json |
| โ ์ฌํ ์คํฌ๋ฆฝํธ | build_subset100.py |
| โ ํ์ ์ฐ์ถ๋ฌผ (์ฐ๋ฆฌ Stage-5 ์์ ๋ฌผ) | derived/ |
| ์์: ์๋ณธ tar 4๊ฐ + ์ ์ฒด ๋์ํ | .. |
| HF ์๋ณธ | https://huggingface.co/datasets/FourierIntelligence/ActionNet |
2. tar ์์ ๋ฌด์์ด ์๋๊ฐ
์ํผ์๋๋น 4๊ฐ ํ์ผ, ํ๊ท 44 MB (rgb 31 MB + depth 11.7 MB + hdf5 1 MB):
<ULID>/top/rgb.mp4 1280x800, 30 fps, h264
<ULID>/top/depth.mkv 1280x800, 16-bit, ffv1 / gray16le (๋ฌด์์ค)
<ULID>/top/timestamps.json ํ๋ ์๋ณ ํ์์คํฌํ ๋ฌธ์์ด
<ULID>.hdf5 /state/{robot(32),hand(12),pose(27)}, /action/{...}, /timestamp
์๋ณธ tar๊ณผ ๋์ผํ ๋ ์ด์์์ด๋ผ, ๊ธฐ์กด์ tar์ ๋ค๋ฃจ๋ ์ฝ๋๊ฐ ๊ทธ๋๋ก ๋๋๋ค.
3. ์ด 100๊ฐ๊ฐ ๋ฌด์์ธ๊ฐ โ ํฌ๋ฃจ์์ ๊ณผ์ ๋ก ์๋์ ์ผ๋ก ๊ณ ์ ํ์ต๋๋ค
| ํญ๋ชฉ | ๊ฐ |
|---|---|
| ์ถ์ฒ tar | 48๊ฐ โ 01JMEX00XF-01JN0B0VXS.tar, 52๊ฐ โ 01JN0B1N7D-01JN0Y2AWR.tar |
| distinct prompt | 3์ข ๋ฟ |
| โ Pick up the croissant and put it on a plate | 55 |
| โ Put the croissant in the container | 42 |
| โ Place the croissant into the container | 3 |
์ ๋ถ ํฌ๋ฃจ์์ ๊ณผ์ ์ด๊ณ , ์ด๋ ์๋ํ ์ ํ์ ๋๋ค. Stage-5 ํํฐ๋ง์ "์์ฑ ์์์ด ์ก๋ ๊ฒ์ฒ๋ผ ๋ณด์ผ ๋ IDM action์ด 3D์์ ์ค์ ๋ก ๋ฌผ์ฒด์ ๋ฟ์๋๊ฐ"๋ฅผ ๊ฒ์ฆํ๋ ์ผ์ด๊ณ , ๊ทธ๊ฑธ ์ฒ์ ์ธ์ธ ๋๋ ๊ณผ์ ๋ฅผ ํ๋๋ก ๊ณ ์ ํ๋ ํธ์ด ํจ์ฌ ์ฝ์ต๋๋ค โ ๋ฌผ์ฒด๊ฐ ํ๋๋ฟ์ด๋ฉด ๋ฌผ์ฒด ๋ง์คํฌยท3D ๋ฐ์คยท์ ์ด ํ์ ์ด ๊ณผ์ ๋ณ ๋ณ์ ์์ด ๋น๊ต๋ฉ๋๋ค. ์ธ ๋ฌธ์ฅ์ ํํ๋ง ๋ค๋ฅด๊ณ ๊ฐ์ ๊ณผ์ ์ ๋๋ค.
๊ณผ์ ๋ค์์ฑ์ด ํ์ํ ์คํ์๋ ์์ ๋๋ ํ ๋ฆฌ์ ์ ์ฒด ๋์ํ์์ ๋ค์ ๊ณ ๋ฅด๋ฉด ๋ฉ๋๋ค
(mapping_*.csv์ prompt / task_group ์ด).
4. depth๋ฅผ ์ธ ๋ ์์์ผ ํ๋ ๊ฒ (์ธก์ ์ผ๋ก ํ์ธํ ์ฌ์ค)
| ์ฌ์ค | ๊ทผ๊ฑฐ |
|---|---|
| ์ ๋ metric์ด๋ค (์๋ยทaffine ๋ชจํธ์ฑ ์์) | ํ์ผ์ uint16 ๋ฐ๋ฆฌ๋ฏธํฐ์ด๊ณ ๋ก๋๊ฐ /1000.0์ผ๋ก ๋ฏธํฐ๋ก ๋ฐ๊พผ๋ค. ์์ ๋ถํฌ min 251~260, ์ค์ 771~799, max 2698 mm. ๊ธฐํ๋ก ๊ต์ฐจ๊ฒ์ฆ: ์นด๋ฉ๋ผ 65 cm ๋์ดยทํ
์ด๋ธ head_pitch_link(64.43 cm)์ 0.7 cm ์ฐจ |
| ํฌํ ์ํ 2698 mm | ๋ชจ๋ ํ๋ ์์์ p99 = max = 2698. 2.7 m ๋ฐ์ ์๋ ค ์์ผ๋ ๋ฐฐ๊ฒฝ ๊ฑฐ๋ฆฌ๋ก ์ฐ๋ฉด ์ ๋๋ค |
| ๋ฌดํจ ํฝ์ 17~18% | ๊ฐ 0. ๋ง์คํฌยทํ๋ฉด ์ ํฉ ์ ์ depth > 0์ผ๋ก ๊ฑธ๋ฌ์ผ ํ๋ค |
| z-depth์ด๋ค (๊ด์ ๋ฐฉํฅ ๊ฑฐ๋ฆฌ๊ฐ ์๋๋ค) | ํ ์ด๋ธ๋ณด ํ๋ฉด ์์ฐจ๋ฅผ ์ ํ๋ฉด ๋ฐ๊ฒฝ์์ ์ฌ๋ฉด z ํด์์ 0.83 cm ์ด๋ด๋ก ํํํ๊ณ , range ํด์์ โ2.85 cm๋ก ํ๋ค. ์ค์ ROI๋ง ๋ณด๋ฉด ๋ ํด์์ด ๊ตฌ๋ถ๋์ง ์๋๋ค (cos ฮธ โ 1) |
| ํ๋ฉด์ฑ | ์ ์ฒด ํ ์ด๋ธ๋ณด๋ก ๊ฐ๊ฑด ์ ํฉ ์ ์์ฐจ๊ฐ ์ ๋ฐ๊ฒฝ 0.63 cm ์ด๋ด |
| ์นด๋ฉ๋ผ ์บ๋ฆฌ๋ธ๋ ์ด์ ์ ๋ฐ์ดํฐ์ ์ ์๋ค | tar ์ํธ๋ฆฌ ์ ์ยทhdf5 attrsยทmeta/info.json์ ๋ค ํ์ธํ๋ค. hdf5 root์๋ camera_names=['top']๋ฟ์ด๊ณ intrinsicยทdistortionยทextrinsic์ ์ด๋์๋ ์๋ค |
์ด์ ์ด์ ๊ฑฐ๋ฆฌ f_native = 577.47 px |
์ถ์ฒ ์๋ ๋ฐ์ดํฐ์ํธ ์ถ์ ์น. ๋ค๋ง ์ธ์ฉ๋ ์ธ FOV(H 127ยฐ, V 79.5ยฐ, D 150ยฐ)๊ฐ r = fยทฮธ ์๋ 0.16%๋ก ์ผ์นํ๊ณ pinhole๋ก๋ 83% ๋ฒ์ด์ง๋ฏ๋ก, ๋ชจ๋ธ์ ์ ๋ท๋ฐ์นจ๋๋ค. ์์ธ๋ filtering/docs/camera.md (๋ด๋ถ RoboCurate_V2, ๋น๊ณต๊ฐ) |
| ์๊ณ๊ฐ ๋ ๊ฐ๋ค (2026-08-20 ์ ์ ) | hdf5 59.9 Hz / ์๋ณธ ๋น๋์ค 30 fps / LeRobot ์์๋ 30 fps(ํ๋ ์ ์๊ฐ ์๋ณธ๊ณผ ๋์ผ). ์ด์ ํ์ "LeRobot 15 fps"๋ ํ๋ ธ๋ค โ meta/info.json fps: 30, ffprobe ์ค์ธก 447/432/491 ํ๋ ์์ด ์๋ณธ ์นด๋ฉ๋ผ ํ๋ ์ ์์ ๊ฐ๋ค. ํ ์๋ 30 fps์ธ๋ฐ LeRobot state ๊ฐ์ 60 Hz ์ธ๋ฑ์ค์ด๋ฏ๋ก ํ๋ ์ ๋์์ ๋ฐ๋์ ๊ฐ ํ์ผ์ frame_indicesยทtimestamp๋ก ๋ง์ถ ๊ฒ |
hdf5์ /state/pose๋ ๊ทธ๋ฅ ์ง๋์น๊ธฐ ์๊น์ต๋๋ค
(T, 27) = 3 ์์ธ ร (์์น 3 + ํ์ 6) ์ด๊ณ , ์์๋ ์ผ EEF / ์ค๋ฅธ EEF / ๋จธ๋ฆฌ์
๋๋ค.
GR1T1 URDF๋ก FK๋ฅผ ๋๋ ค ๋์กฐํ๋ฉด ์ด๋ ๊ฒ ๋ง์ต๋๋ค:
๋ฐ์ดํฐ์
/state/pose |
URDF FK | ์ฐจ์ด | |
|---|---|---|---|
| ์ผ EEF | [0.2143, 0.2437, 0.1735] | left_end_effector_link [0.2132, 0.2451, 0.1697] |
0.4 cm |
| ์ค๋ฅธ EEF | [0.2268, โ0.2085, 0.1733] | right_end_effector_link [0.2284, โ0.2071, 0.1752] |
0.3 cm |
| ๋จธ๋ฆฌ | [โ0.0041, โ0.001, 0.6443] | head_pitch_link [โ0.0021, 0.005, 0.6443] |
0.6 cm |
์ฆ ์ด๋ํ๊ณผ state ๋งคํ์ ๋ฐ์ดํฐ์ ์์ฒด๋ก ๊ฒ์ฆํ ์ ์์ต๋๋ค. ๋ค๋ง ์ธ ๊ฐ์ง๋ฅผ ์งํค์ธ์.
- ๋ง๋ ๋งํฌ๋
{left,right}_end_effector_link์ ๋๋ค.{L,R}_hand_base_link๋ก ๋น๊ตํ๋ฉด ์ฝ 2.1 cm๊ฐ ๋จ์ต๋๋ค(์ํผ์๋ ์ ์ฒด์์ ๊ฑฐ์ ์ผ์ = ๋งํฌ ์ ํ ์ฐจ์ด, ์ค์ฐจ๊ฐ ์๋). ์ฆ ์ด ๋ฐ์ดํฐ์ ์ด ๋งํ๋ "EEF"๋ hand_base๊ฐ ์๋๋ผ end_effector ๋งํฌ์ ๋๋ค. /state/pose๋ 60 Hz์ ๋๋ค โ ํ ์๊ฐ mp4 ํ๋ ์ ์์ ์ฝ 2๋ฐฐ์ ๋๋ค(892/447, 799/401, 1133/567 โฆ). ๊ทธ๋์P[์์ํ๋ ์]์ผ๋ก ์ธ๋ฑ์ฑํ๋ฉด ํ๋ฆฝ๋๋ค. frame 0๋ผ๋ฆฌ๋ ๊ทธ๋ฅ ๋ง์ง๋ง, ๊ทธ ๋ฐ์ ํ๋ ์์/timestamp๋ก ๋ฆฌ์ํํ์ธ์. โ ๏ธ LeRobot์ state ํ๋ ๋๊ฐ์ด 60 Hz ์ธ๋ฑ์ค์ ๋๋ค (2026-08-20 ์ค์ธก:observation.robot_joints[k] == /state/robot[k], 100 ep ์ค 97๊ฑด์ด max|diff| 1e-5 ์ด๋ด, ์ค์ 5.949e-08;[2k]๊ฐ์ค์ 0/100, ์ค์ 0.740). LeRobot์ ํ ์๋ ์์ ํ๋ ์ ์์ ๊ฐ์์ "๊ทธ๋ฌ๋ ์์ ํ๋ ์์ผ๋ก ์ธ๋ฑ์ฑํด๋ ๋๋ค"๊ณ ์ฐฉ๊ฐํ๊ธฐ ์ฌ์ด๋ฐ, ํ ์๊ฐ ๋ง๋ ๊ฒ๊ณผ ๋ด์ฉ์ ์๊ณ๊ฐ ๋ง๋ ๊ฒ์ ๋ค๋ฅธ ๋ฌธ์ ์ ๋๋ค. hdf5๋ LeRobot์ด๋ ์์ ํ๋ ์ ๋ฒํธ๋ก state๋ฅผ ์ธ๋ฑ์ฑํ์ง ๋ง์ธ์ โ ์๊ณ๋ก ๋ฆฌ์ํํ์ธ์. ๊ทผ๊ฑฐ๋ ์์README_MAPPING.mdยง4(๋ด๋ถ ๋ฌธ์)./state/robot์ ์ฑ๋ ์์๋ LeRobot 44์ฐจ์๊ณผ ๋ค๋ฆ ๋๋ค โ ๋ค๋ง ๊ทธ ์์๋ ์ด์ ์ธก์ ์ผ๋ก ํ์ ๋ผ ์์README_MAPPING.mdยง6-2์ ํ๋ก ์ ํ ์์ต๋๋ค(์ ์ฉํ๋ฉด 100/100 ์ํผ์๋์์ max|diff| = 0).derived/states์ 44์ฐจ์๋ ๊ทธ ์์์ ๋๋ค.
5. derived/ โ ์ฐ๋ฆฌ๊ฐ ๋ง๋ ํ์ ์ฐ์ถ๋ฌผ (์งํ ์ค)
| ๊ฒฝ๋ก | ๊ฐ์ | ๋ด์ฉ |
|---|---|---|
derived/states/states_<ULID>.npz |
100 | ์๋ณธ hdf5์์ ๋ง๋ per-video-frame ์ํ, ฯ ๋ณด์ ์ . ์๋ ์ฐธ์กฐ (2026-08-20 ์ฌ์์ฑ) |
derived/states_tau/states/ |
100 | ฯ ๋ณด์ ์ํ. ์์ ํ๋ ์๊ณผ ์ง์ง์ ๋๋ ์ด๊ฒ์ ์ฐ์ธ์. ๊ทผ๊ฑฐ๋ states/README.md |
derived/states_lerobot_WRONG_do_not_use/ |
100 | ์ฐ์ง ๋ง์ธ์. ์ ๋ ฌ์ด ํ๋ฆฐ ์ด์ ํ(LeRobot ์ ๋, stride 4). ํ์ ์ฉ์ผ๋ก๋ง ๋จ๊ฒจ ๋ก๋๋ค |
derived/points/vlm_points_<ULID>.json |
99 | ๋ฌผ์ฒด point prompt (positive/negative), ์ํผ์๋๋ณ ์๋ ์ง์ |
derived/masks/masks/ |
100 | SAM3 ๋ฌผ์ฒด ๋ง์คํฌ, ์ ํ๋ ์. 43๊ฐ๋ฟ์ด๋ derived/masks/๋ ์ญ์ ํ๊ณ ์ต์์ masks/๊ฐ 100ํธ ์ ๋ถ์
๋๋ค |
derived/hand_points2/points_<ULID>.json |
10 | ํ+์ point prompt: ์๋น 3์ (์๊ฐ๋ฝยท์๋ฐ๋ฅยทํ๋) + negative |
derived/hand_tracks3/tracks_<ULID>.npz |
10 | SAM3 ํ+์ ๋ง์คํฌ, ์ ํ๋ ์, ์ข/์ฐ ๋ถ๋ฆฌ (left, right, frame_indices) |
derived/states โ ์๋ณธ hdf5์์ ๋ค์ ๋ง๋ค์์ต๋๋ค (2026-08-20)
์ด์ ํ์ LeRobot parquet์ state ์ด์ stride 4๋ก ์์ ํ๋ ์์ ์น์ ๊ฒ์ด์๋๋ฐ,
์ ์ ("LeRobot 15 fps ร stride 4 = 60 Hz")๊ฐ ๋ ๋ค ํ๋ ธ์ต๋๋ค โ LeRobot์ 30 fps์ด๊ณ
state ํ์ 60 Hz ์ธ๋ฑ์ค์
๋๋ค(ยง4, ์์ README_MAPPING.md ยง4(๋ด๋ถ ๋ฌธ์)). ์ง๊ธ ํ์ LeRobot์
์์ ๊ฑฐ์น์ง ์๊ณ , ์๋ณธ hdf5๋ฅผ ์์ ์์ ์ timestamps.json ์๊ณ์ ์ต๊ทผ์ ๋ฆฌ์ํํฉ๋๋ค
(๋ณด๊ฐ ๊ธ์ง โ ์ 12์ฑ๋์ด ๊ณ๋จ์ ์์ ๋จ์๋ผ ๋ณด๊ฐํ๋ฉด ๋ก๋ด์ด ๋ณด๊ณ ํ ์ ์๋ ๊ฐ์ด ์๊น๋๋ค).
๋น๋๋ ๋ด๋ถ RoboCurate_V2์ tools/build_states_hdf5.py(๋น๊ณต๊ฐ).
| ํค | shape (์: 01JMXF99BYFPWU5NOL, T=447) |
๋ด์ฉ |
|---|---|---|
states / actions |
(447, 44) float64 | 44์ฐจ์ GR1-T1 ์์ (์์ ๋ฌธ์ ยง6-2 ํ) |
frame_indices |
(447,) int64 | ์๋ณธ ๋น๋์ค ํ๋ ์ ๋ฒํธ (0..Tโ1, ๋ง์คํฌยทtracks์ ๊ฐ์ ์ถ) |
hdf5_rows |
(447,) int64 | ๊ฐ ํ๋ ์์ด ๊ณ ๋ฅธ hdf5 ํ |
pose |
(447, 3, 9) float64 | /state/pose ๊ฐ์ ํ (์ผ EEF / ์ค๋ฅธ EEF / ๋จธ๋ฆฌ) |
video_t, lag_s, clock_offset_s |
(447,), (447,), scalar | ์์ ์๊ณ ์ด, ํ๋ ์โํ ์์ฐจ, hdf5โ์์ ์๊ณ ์คํ์ |
๊ฒ์ฆ (100 ep ์ ์, FK vs ๋ฐ์ดํฐ์
์์ ์ /state/pose):
| ์ค์ | ์ต์ | 3 cm ์ด๊ณผ | |
|---|---|---|---|
| ์ผ EEF | 0.29 cm | 0.37 cm | 0/100 |
| ์ค๋ฅธ EEF | 0.27 cm | 0.32 cm | 0/100 |
| ๋จธ๋ฆฌ | 0.59 cm | 0.60 cm | 0/100 |
ํ๋ ์โํ ์์ฐจ(lag_s)๋ 46,927 ํ๋ ์ ์ค 98.6%๊ฐ 8.34 ms(60 Hz ๋ฐ ์นธ) ์ด๋ด
(์ค์ 2.51 ms, p95 6.42 ms, 16.7 ms ์ด๊ณผ 1 ํ๋ ์, ์ต์
24.35 ms โ ํ ์ํผ์๋
01JN0XAGCRDDJ33ML4์ ๋ชฐ๋ ค ์์). ๊ฐ ์ํผ์๋์ ๋ง์ง๋ง ํ๋ ์๋ง hdf5๊ฐ ์์๋ณด๋ค ๋จผ์
๋๋์ โ12.0 ~ โ57.7 ms(์ค์ โ29.7 ms) ๋ฒ์ด์ง๋๋ค. ์๊ณ ์คํ์
์ โ21.6 ~ +11.1 ms.
๋ง์ง๋ง ํ๋ ์์ ์ธ ๊ฑฐ๋ฉด ์ด ๊ฐ์ ๊ฐ์ํ์ธ์.
โ ๏ธ ๋ง์คํฌ๋ ์๋ณธ ์ด์ ์ขํ๊ณ๊ฐ ์๋๋๋ค โ ์ธ๋์คํ ํธ๋ pinhole ๊ณต๊ฐ์ ๋๋ค
derived/hand_tracks3/์ derived/masks/๋ raw rgb.mp4 ์์์ ๊ณ์ฐ๋ ๊ฒ์ด ์๋๋๋ค.
filtering/src/depth/undistort_fisheye.py --hfov 90์ผ๋ก ๋ค์ ๋ ๋ํ rectilinear(pinhole)
์์ ์์์ SAM3๋ฅผ ๋๋ฆฐ ๊ฒฐ๊ณผ์
๋๋ค. ์ฆ ๋ npz์ ์ขํ๊ณ๋:
| ํญ๋ชฉ | ๊ฐ |
|---|---|
| ํด์๋ | 1280ร800 (์๋ณธ๊ณผ ๊ฐ์ โ ๊ทธ๋์ ์ก์์ผ๋ก ๊ตฌ๋ถ๋์ง ์์ต๋๋ค) |
| K | [[640, 0, 639.5], [0, 640, 399.5], [0, 0, 1]] (hfov 90ยฐ) |
| ํฌ์ | pinhole (r = fยทtanฮธ). ์๋ณธ์ equidistant ์ด์ (r = fยทฮธ) |
| npz ์์ ์ฆ๊ฑฐ | video ํค = /tmp/rc_undist_<ULID>.mp4 / /tmp/rc_full_<jobid>_<ULID>.mp4, ๋ ๋ค ์ ์คํฌ๋ฆฝํธ ์ถ๋ ฅ |
raw rgb.mp4์ ์ด ๋ง์คํฌ๋ฅผ ๊ทธ๋๋ก ์น์ผ๋ฉด ์ฃผ๋ณ๋ถ์์ 100 px ์ด์ ํ์ด์ง๋๋ค.
๊ฐ์ ํฝ์
์ด ๋ ๊ณต๊ฐ์์ ์ผ๋ง๋ ๋จ์ด์ ธ ์๋์ง(์ ๋ ๋ชจ๋ธ๋ก ๊ณ์ฐ):
| ์ธ๋์คํ ํธ ์์์ ํฝ์ | ๊ด์ถ์์ | ๋์ํ๋ ์๋ณธ ์ด์ ํฝ์ | ์ด๊ธ๋จ |
|---|---|---|---|
| (1279, 399.5) ์ค๋ฅธ์ชฝ ๋ ์ค์ | 44.98ยฐ | (1092.8, 399.5) | 186 px |
| (1279, 799) ์ค๋ฅธ์๋ ์ฝ๋ | 49.68ยฐ | (1064.1, 664.8) | 253 px |
| (959.5, 399.5) ์ค๋ฅธ์ชฝ ์ ๋ฐ ์ง์ | 26.57ยฐ | (907.2, 399.5) | 52 px |
| ํ๋ฉด ์ค์ | 0ยฐ | ํ๋ฉด ์ค์ | 0 px |
์ค์์์๋ 0์ด๋ผ ์ค์ ROI๋ง ๋ณด๋ฉด ๋ ๊ณต๊ฐ์ด ๊ตฌ๋ถ๋์ง ์์ต๋๋ค โ ์ด ์ค๋ฅ๊ฐ ๋์ ๋์ง ์๋ ์ด์ ์ ๋๋ค. ๋ hfov 90ยฐ๋ ์ด์์ด ๋ด์ ํ๊ฐ(127ยฐ)๋ณด๋ค ์ข์ผ๋ฏ๋ก ์ด์์ ๋ฐ๊นฅ ์ฃผ๋ณ๋ถ๋ ์์ถ๋์ง ์๊ณ ๋ฒ๋ ค์ง๋๋ค(1280ร800์์ ๊นจ๋ํ๊ฒ ๋ฝ์ ์ ์๋ ์ต๋ hfov๋ 95.9ยฐ).
๋ง๋ ์์์ ๋ค์ ๋ง๋๋ ๋ฒ:
cd <๋ด๋ถ RoboCurate_V2>/filtering/src
python depth/undistort_fisheye.py <๋ฐ์ดํฐ>/data_og/$U/top/rgb.mp4 --hfov 90 -o /tmp/rc_undist_$U.mp4
# ํ๋ ์ ์ยท์์๋ ์๋ณธ๊ณผ ๋์ผํ๋ฏ๋ก mask์ frame_indices๊ฐ ๊ทธ๋๋ก ๋ง์ต๋๋ค
์์ ํ์ ยทdepth ์กฐํ๋ ๋ฐ๋์ ๊ฐ์ ์ธ๋์คํ ํธ ๊ณต๊ฐ์์ ํ์ธ์. depth๋
depth/gt_depth_npz.py(๊ธฐ๋ณธ --hfov 90)๊ฐ rgb์ depth๋ฅผ ๊ฐ์ ์ด์โpinhole ๋งต์ผ๋ก
๋ฆฌ์ํํ๋ฏ๋ก(RGB ์ด์ค์ ํ, depth ์ต๊ทผ์ ) ๊ทธ ์ถ๋ ฅ๊ณผ ๋ง์คํฌ๊ฐ ํฝ์
๋จ์๋ก ๋์ํฉ๋๋ค.
raw depth.mkv๋ฅผ ์ง์ ์ฝ์ด ๋ง์คํฌ ํฝ์
๋ก ์ธ๋ฑ์ฑํ๋ฉด ๋ค๋ฅธ ํฝ์
์ ๊ฑฐ๋ฆฌ๊ฐ์ ์ง๊ฒ ๋ฉ๋๋ค.
๋ง์คํฌ๋ฅผ ์ธ ๋์ ์ฃผ์ (์จ๊ธฐ์ง ์๊ณ ์ ์ต๋๋ค): ํ+์ ๋ง์คํฌ๋ 10๊ฐ ์ํผ์๋ ์ค 8๊ฐ์์
ํ์ชฝ์ด 61,3xx~61,43x px์ ๋ญ์ณ ์์ต๋๋ค. ์ด๋ ์ถ์ถ ์ํ(max_obj_frac 0.06 ร 1,024,000 = 61,440)์ ๊น์ฌ ๋ถ์ ํ
์ด๋ธ๋ณด ๋์ถ์
๋๋ค. ๊นจ๋ํ ๋ง์คํฌ๋ 14k33k px(ํ๋ ์์ 1.43.3%)์ด๋
๋ฉด์ ์ํ 0.045๋ก ๊ฑธ๋ฌ์ ์ฐ์ธ์. ํ๋ ์ ์ ๋ณ๊ธฐ๋ ๋ด๋ถ RoboCurate_V2(๋น๊ณต๊ฐ)์
filtering/src/align/filter_mask_frames.py์ ์๊ณ , ํ์ ์ ์นด๋ฉ๋ผ ์์ธ๋ฅผ ์ผ์ ์ฐ์ง ์์ต๋๋ค
(ํ์ฌ extrinsic๊ณผ ์ ๋ง๋ ํ๋ ์๋ง ๊ณ ๋ฅด๋ฉด ๊ทธ extrinsic์ ํ์ฆํด ๋ฒ๋ฆฌ๊ธฐ ๋๋ฌธ).
6. ์ฐ๋ ๋ฒ
D=. # HF ์ ์ฅ์๋ฅผ ๋ฐ์ ๋๋ ํ ๋ฆฌ (๋ด๋ถ์์๋ ์ํฐํฉํธ ๋ง์ดํธ ๊ฒฝ๋ก)
# ์ ์ฒด ํ๊ธฐ (4.4 GB)
mkdir -p ~/actionnet_subset100 && tar -xf $D/subset100_gtdepth.tar -C ~/actionnet_subset100
# ํ ์ํผ์๋๋ง
U=01JMXF99BYFPWU5NOL
tar -xf $D/subset100_gtdepth.tar -C ~/actionnet_subset100 $U/ $U.hdf5
# 16-bit depth๋ฅผ numpy๋ก (ffv1/gray16le๋ OpenCV๋ก ์ด๋ฆฌ์ง ์์ต๋๋ค)
ffmpeg -i $U/top/depth.mkv -f rawvideo -pix_fmt gray16le - \
| python -c "import sys,numpy as np; \
a=np.frombuffer(sys.stdin.buffer.read(),np.uint16).reshape(-1,800,1280); \
print(a.shape, a.dtype, a[0].max())"
depth์ ๋จ์๋ mm(uint16) ์ด๊ณ , ๋ด๋ถ RoboCurate_V2(๋น๊ณต๊ฐ)์ filtering/src/depth/gt_depth_npz.py๋
์ด mkv์ rgb๋ฅผ ๊ฐ์ ์ด์โpinhole ๋งต์ผ๋ก ๋ฆฌ์ํํด(RGB ์ด์ค์ ํ, depth ์ต๊ทผ์ ) npz๋ก ๋ง๋ญ๋๋ค.
depth๋ฅผ ๋ณด๊ฐํ๋ฉด ๋ฌผ์ฒด ๊ฒฝ๊ณ์์ ์๋ ๊ฑฐ๋ฆฌ๊ฐ์ด ์๊ธฐ๋ฏ๋ก ์ต๊ทผ์ ์ด์ด์ผ ํฉ๋๋ค.
7. ์ฌํยท๊ฒ์ฆ
cd $D
python build_subset100.py --verify # ๋์ํ ์ฌ์์ฑ + 400๊ฐ ํ์ด๋ก๋ ํ์ผ ์กด์ฌ ํ์ธ
sha256sum -c subset100_gtdepth.tar.sha256
build_subset100.py๋ ULIDโLeRobot ๋์์ ๋ค์ ๊ณ์ฐํ์ง ์์ต๋๋ค โ ์์
build_mapping.py๊ฐ ๋ง๋ per-tar ๋์ํ์์ ์ฐ๋ฆฌ 100๊ฐ ํ๋ง ๊ณจ๋ผ๋ด๋ฏ๋ก, ๋ ํ๊ฐ ์ด๊ธ๋ ์
์์ต๋๋ค. ๊ฒ์ฆ ๊ฒฐ๊ณผ๋ ์์ฑ ์์ ์ 400/400์ด์์ต๋๋ค.
์ถ๊ฐ 2026-08-21 โ derived/states_tau/์ ์ผํ ๋ณด์ ๊ฐ
์์ ํ์์คํฌํ๊ฐ ๋
ธ์ถ ์๊ฐ๋ณด๋ค ์ฝ 167 ms ๋ฆ๊ฒ ๊ธฐ๋ก๋ผ ์์์ ๋ง์คํฌ ๋๋น ์ค์ธก์ผ๋ก ํ์ธํ๊ณ ,
๊ทธ ์ํํธ๋ฅผ ๋ฐ์ํ ์ํ๋ฅผ ์ต์์ states/(์์ derived/states_tau/)์ ๋์์ต๋๋ค(derived/states/๋ ๊ทธ๋๋ก).
์์ ํ๋ ์๊ณผ ์ํ๋ฅผ ์ง์ง์ด ์ฐ๋ ์ฉ๋์๋ states_tau๋ฅผ ์ฐ์ธ์. ์ผํ ๊ด์ ์์ ๋ฐ์ด์ด์ค
๋ณด์ ๊ฐ์ ์ ์ฅ์ ๋ฃจํธ์ left_arm_offset_20260821.json์ ์์ต๋๋ค. ๋ ๋ค ๊ทผ๊ฑฐ์ ์์น๋
states/README.md์ ์ ์์ต๋๋ค.
- Downloads last month
- 226