liyang
ae28d55f81
Source: RGB-D -> Dyn-HaMR -> L20 retargeting -> FoundationPose -> reference repair -> SPIDER, documented in docs/PIPELINE_LATEST.md and docs/SETUP_AND_WEIGHTS.md. Adds FoundationPose and nvdiffrast upstream snapshots, requirements/pipeline_venv.txt and the FoundationPose weight manifest/downloader. Assets (Git LFS): weights/ (WiLoR detector, HandFlow denoiser, UniDepth-L), FoundationPose checkpoints, HaMeR checkpoint, Dyn-HaMR HMP model and BMC constraints, L20 URDF/meshes, the 20260915_171525 D405 recording and the two box CADs. MANO models are not redistributed (third_party/hamer/_DATA/data/mano/README.txt). Environments, caches and run outputs excluded. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
36 lines
2.5 KiB
Python
36 lines
2.5 KiB
Python
from pathlib import Path
|
|
import json,numpy as np
|
|
from scipy.ndimage import gaussian_filter1d
|
|
from scipy.spatial.transform import Rotation
|
|
ROOT=Path(__file__).resolve().parents[1];B=ROOT/'output/20260915_171525_dynhamr';a=dict(np.load(B/'object_refit_tracked/object_poses.npz'));report={}
|
|
for n in ['upper','lower']:
|
|
center=np.array([0, .00415 if n=='upper' else -.03335, .2185 if n=='upper' else .21975])
|
|
r=Rotation.from_quat(a[n+'_quaternion_xyzw']);c=r.apply(np.tile(center,(352,1)))+a[n+'_position_world'];cs=gaussian_filter1d(c,2,axis=0)
|
|
rs=[]
|
|
for t in range(352):
|
|
ix=np.arange(max(0,t-6),min(352,t+7));rs.append(r[ix].mean(weights=np.exp(-.5*((ix-t)/2)**2)).as_quat())
|
|
rr=Rotation.from_quat(rs)
|
|
# Resolve visible-face ambiguity using the source video: flat red face and circular blue feature.
|
|
rr=rr*Rotation.from_matrix(np.diag([-1.,-1.,1.]))
|
|
a[n+'_position_world']=cs-rr.apply(np.tile(center,(352,1)));a[n+'_quaternion_xyzw']=rr.as_quat()
|
|
angles=np.rad2deg((rr[:-1].inv()*rr[1:]).magnitude());steps=np.linalg.norm(np.diff(cs,axis=0),axis=1)*1000
|
|
report[n]={'max_center_step_mm':float(steps.max()),'max_rotation_step_deg':float(angles.max()),'center_smoothing_max_change_mm':float(np.linalg.norm(cs-c,axis=1).max()*1000)}
|
|
assert np.isfinite(cs).all() and angles.max()<20 and steps.max()<30
|
|
# Video inspection: lower part becomes occluded during placement (frames 180-214),
|
|
# and the common CAD assembly is visible from frame 215. This is an explicit assumption.
|
|
from scipy.spatial.transform import Slerp
|
|
hold_p=a['lower_position_world'][176].copy();hold_q=a['lower_quaternion_xyzw'][176].copy()
|
|
a['lower_position_world'][177:190]=hold_p;a['lower_quaternion_xyzw'][177:190]=hold_q
|
|
for t in range(190,215):
|
|
alpha=(t-189)/26
|
|
a['lower_position_world'][t]=(1-alpha)*hold_p+alpha*a['upper_position_world'][t]
|
|
a['lower_quaternion_xyzw'][t]=Slerp([0,1],Rotation.from_quat([hold_q,a['upper_quaternion_xyzw'][t]]))([alpha]).as_quat()[0]
|
|
a['lower_position_world'][215:]=a['upper_position_world'][215:]
|
|
a['lower_quaternion_xyzw'][215:]=a['upper_quaternion_xyzw'][215:]
|
|
a['lower_keyframe_valid'][177:]=False
|
|
a['lower_assembly_assumed']=np.arange(352)>=215
|
|
a['lower_occluded_transition_assumed']=(np.arange(352)>=177)&(np.arange(352)<215)
|
|
report['assembly_assumption']={'from_frame':215,'transition_frames':[177,214],'basis':'Visual inspection and shared CAD assembly coordinates; not independent measured lower poses.'}
|
|
np.savez_compressed(B/'object_aligned/object_poses.npz',**a)
|
|
(B/'object_aligned/smoothing_validation.json').write_text(json.dumps(report,indent=2));print(report)
|