-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathfit_SMPLD.py
196 lines (161 loc) · 7.81 KB
/
fit_SMPLD.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
"""
This code optimizes the offsets on top of SMPL.
If code works:
Author: Bharat
else:
Author: Anonymous
Cite: Combining Implicit Function Learning and Parametric Models for 3D Human Reconstruction, ECCV 2020.
"""
import os
from os.path import split, join, exists
from glob import glob
import torch
from kaolin.rep import TriangleMesh as tm
from kaolin.metrics.mesh import point_to_surface, laplacian_loss
from tqdm import tqdm
import pickle as pkl
import numpy as np
import sys
sys.path.append('../lib/')
from lib.smpl_paths import SmplPaths
from lib.th_SMPL import th_batch_SMPL
from fit_SMPL import fit_SMPL, save_meshes, batch_point_to_surface, backward_step
def get_loss_weights():
"""Set loss weights"""
loss_weight = {'s2m': lambda cst, it: 10. ** 2 * cst * (1 + it),
'm2s': lambda cst, it: 10. ** 2 * cst, #/ (1 + it),
'lap': lambda cst, it: 10. ** 4 * cst / (1 + it),
'offsets': lambda cst, it: 10. ** 1 * cst / (1 + it)}
return loss_weight
def forward_step(th_scan_meshes, smpl, init_smpl_meshes):
"""
Performs a forward step, given smpl and scan meshes.
Then computes the losses.
"""
# forward
verts, _, _, _ = smpl()
th_smpl_meshes = [tm.from_tensors(vertices=v,
faces=smpl.faces) for v in verts]
# losses
loss = dict()
loss['s2m'] = batch_point_to_surface([sm.vertices for sm in th_scan_meshes], th_smpl_meshes)
loss['m2s'] = batch_point_to_surface([sm.vertices for sm in th_smpl_meshes], th_scan_meshes)
loss['lap'] = torch.stack([laplacian_loss(sc, sm) for sc, sm in zip(init_smpl_meshes, th_smpl_meshes)])
loss['offsets'] = torch.mean(torch.mean(smpl.offsets**2, axis=1), axis=1)
return loss
def optimize_offsets(th_scan_meshes, smpl, init_smpl_meshes, iterations, steps_per_iter):
# Optimizer
optimizer = torch.optim.Adam([smpl.offsets, smpl.pose, smpl.trans, smpl.betas], 0.005, betas=(0.9, 0.999))
# Get loss_weights
weight_dict = get_loss_weights()
for it in range(iterations):
loop = tqdm(range(steps_per_iter))
loop.set_description('Optimizing SMPL+D')
for i in loop:
optimizer.zero_grad()
# Get losses for a forward pass
loss_dict = forward_step(th_scan_meshes, smpl, init_smpl_meshes)
# Get total loss for backward pass
tot_loss = backward_step(loss_dict, weight_dict, it)
tot_loss.backward()
optimizer.step()
l_str = 'Lx100. Iter: {}'.format(i)
for k in loss_dict:
l_str += ', {}: {:0.4f}'.format(k, loss_dict[k].mean().item()*100)
loop.set_description(l_str)
def fit_SMPLD(scans, smpl_pkl=None, gender='male', save_path=None, display=False):
# Get SMPL faces
sp = SmplPaths(gender=gender)
smpl_faces = sp.get_faces()
th_faces = torch.tensor(smpl_faces.astype('float32'), dtype=torch.long).cuda()
# Batch size
batch_sz = len(scans)
# Init SMPL
if smpl_pkl is None or smpl_pkl[0] is None:
print('SMPL not specified, fitting SMPL now')
pose, betas, trans = fit_SMPL(scans, None, gender, save_path, display)
else:
pose, betas, trans = [], [], []
for spkl in smpl_pkl:
smpl_dict = pkl.load(open(spkl, 'rb'), encoding='latin-1')
p, b, t = smpl_dict['pose'], smpl_dict['betas'], smpl_dict['trans']
pose.append(p)
if len(b) == 10:
temp = np.zeros((300,))
temp[:10] = b
b = temp.astype('float32')
betas.append(b)
trans.append(t)
pose, betas, trans = np.array(pose), np.array(betas), np.array(trans)
betas, pose, trans = torch.tensor(betas), torch.tensor(pose), torch.tensor(trans)
smpl = th_batch_SMPL(batch_sz, betas, pose, trans, faces=th_faces).cuda()
verts, _, _, _ = smpl()
init_smpl_meshes = [tm.from_tensors(vertices=v.clone().detach(),
faces=smpl.faces) for v in verts]
# Load scans
th_scan_meshes = []
for scan in scans:
th_scan = tm.from_obj(scan)
if save_path is not None:
th_scan.save_mesh(join(save_path, split(scan)[1]))
th_scan.vertices = th_scan.vertices.cuda()
th_scan.faces = th_scan.faces.cuda()
th_scan.vertices.requires_grad = False
th_scan_meshes.append(th_scan)
# Optimize
optimize_offsets(th_scan_meshes, smpl, init_smpl_meshes, 5, 10)
print('Done')
verts, _, _, _ = smpl()
th_smpl_meshes = [tm.from_tensors(vertices=v,
faces=smpl.faces) for v in verts]
#to get the T-pose fitted SMPLD mesh
print("Making the fitted SMPLD model to T-pose")
# Init SMPL, pose with mean smpl pose, as in ch.registration
#copy the optimized smpl model and add t-pose
pose_tpose = torch.zeros((batch_sz, 72))
smplD_tpose = th_batch_SMPL(batch_sz, smpl.betas, pose_tpose, trans, offsets=smpl.offsets, faces=th_faces).cuda()
#re-pose it to T-pose
#import torch.nn as nn
#smpl_tpose.pose = nn.Parameter(torch.zeros(batch_sz, 72))
#extract vertices
verts_tpose, _, _, _ = smplD_tpose()
th_smplD_meshes_tpose = [tm.from_tensors(vertices=v, faces=smplD_tpose.faces) for v in verts_tpose]
if save_path is not None:
if not exists(save_path):
os.makedirs(save_path)
names = [split(s)[1] for s in scans]
# Save meshes
save_meshes(th_smpl_meshes, [join(save_path, n.replace('.obj', '_smpld.obj')) for n in names])
save_meshes(th_scan_meshes, [join(save_path, n) for n in names])
# Save params
for p, b, t, d, n in zip(smpl.pose.cpu().detach().numpy(), smpl.betas.cpu().detach().numpy(),
smpl.trans.cpu().detach().numpy(), smpl.offsets.cpu().detach().numpy(), names):
smpl_dict = {'pose': p, 'betas': b, 'trans': t, 'offsets': d}
pkl.dump(smpl_dict, open(join(save_path, n.replace('.obj', '_smpld.pkl')), 'wb'))
#save smplD meshes
save_meshes(th_smplD_meshes_tpose, [join(save_path, n.replace('.obj', '_smpld_tpose.obj')) for n in names])
# Save params
for p, b, t, d, n in zip(smplD_tpose.pose.cpu().detach().numpy(), smplD_tpose.betas.cpu().detach().numpy(),
smplD_tpose.trans.cpu().detach().numpy(), smplD_tpose.offsets.cpu().detach().numpy(), names):
smpl_dict = {'pose': p, 'betas': b, 'trans': t, 'offsets': d}
pkl.dump(smpl_dict, open(join(save_path, n.replace('.obj', '_smpld_tpose.pkl')), 'wb'))
print("Done saving SMPLD")
return smpl.pose.cpu().detach().numpy(), smpl.betas.cpu().detach().numpy(), \
smpl.trans.cpu().detach().numpy(), smpl.offsets.cpu().detach().numpy()
if __name__ == "__main__":
import argparse
parser = argparse.ArgumentParser(description='Run Model')
parser.add_argument('scan_path', type=str)
parser.add_argument('save_path', type=str)
parser.add_argument('-smpl_pkl', type=str, default=None) # In case SMPL fit is already available
parser.add_argument('-gender', type=str, default='male') # can be female/ male/ neutral
parser.add_argument('--display', default=False, action='store_true')
args = parser.parse_args()
# args = lambda: None
# args.scan_path = '/BS/bharat-2/static00/renderings/renderpeople/rp_alison_posed_017_30k/rp_alison_posed_017_30k.obj'
# args.smpl_pkl = '/BS/bharat-3/work/IPNet/DO_NOT_RELEASE/test_data/rp_alison_posed_017_30k_smpl.pkl'
# args.display = False
# args.save_path = '/BS/bharat-3/work/IPNet/DO_NOT_RELEASE/test_data'
# args.gender = 'female'
_, _, _, _ = fit_SMPLD([args.scan_path], smpl_pkl=[args.smpl_pkl], display=args.display, save_path=args.save_path,
gender=args.gender)