-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathfit_dynamic_landmarks.py
209 lines (167 loc) · 8.28 KB
/
fit_dynamic_landmarks.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
197
198
199
200
201
202
203
204
205
206
207
208
209
import os
from torch.autograd import Variable
from config import parser, get_config
from fitting.landmarks_fitting import *
from fitting.silhouette_fitting import *
from fitting.correspondences_from_edges import *
from utils.perspective_camera import get_init_translation_from_lmks
from pytorch3d.renderer import OpenGLPerspectiveCameras, look_at_view_transform, OpenGLOrthographicCameras
from utils.model_ploting import plot_landmarks, plot_silhouette
from Yam_research.utils.utils import CoordTransformer, zero_pad_img
####################33
from pytorch3d.io import load_obj
# 3D transformations functions
from pytorch3d.transforms import Rotate, Translate
# rendering components
from pytorch3d.renderer import (
OpenGLPerspectiveCameras, look_at_view_transform, look_at_rotation,
RasterizationSettings, MeshRenderer, MeshRasterizer, BlendParams,
SoftSilhouetteShader, HardPhongShader, PointLights
)
#########################
def image2model_pipline(texture_mapping, target_img_path, out_path):
if not os.path.exists(target_img_path):
print('Target image not found - s' % target_img_path)
return
if not os.path.exists(out_path):
os.makedirs(out_path)
# get and transform target 2d lmks
target_img = cv2.imread(target_img_path)
target_img = zero_pad_img(target_img)
target_img = cv2.resize(target_img, (1024, 1024))
target_2d_lmks_oj = get_2d_lmks(target_img)
# target_2d_lmks_oj[:, 0] = -target_2d_lmks_oj[:, 0]
# target_2d_lmks_oj[:, 1] = target_img.shape[0] - target_2d_lmks_oj[:, 1]
# target_2d_lmks = target_2d_lmks_oj
coord_transformer = CoordTransformer(target_img.shape)
target_2d_lmks = coord_transformer.screen2cam(target_2d_lmks_oj)
flamelayer = FlameLandmarks(config, use_face_contour = True)
flamelayer.cuda()
device = torch.device("cuda:0")
distance = 0.3 # distance from camera to the object
elevation = 0.0 # angle of elevation in degrees
azimuth = 0.0 # No rotation so the camera is positioned on the +Z axis.
# Get the position of the camera based on the spherical angles
R, init_translation = look_at_view_transform(distance, elevation, azimuth, device=device)
# Guess initial camera parameters (perspective = R,T)
# init_translation = get_init_translation_from_lmks()
T = Variable(init_translation.cuda(), requires_grad=True)
cam = OpenGLPerspectiveCameras(T=T, R=R, device=device)
renderer = Renderer(cam, resulution=1024)
# Initial guess: fit by optimizing only rigid motion
vars = [flamelayer.transl, flamelayer.global_rot] # Optimize for global scale, translation and rotation
rigid_scale_optimizer = torch.optim.LBFGS(vars, tolerance_change=5e-6, max_iter=500)
vertices = fit_flame_to_2D_landmarks_perspectiv(flamelayer, cam, target_2d_lmks,
rigid_scale_optimizer)
#plotting
#plot_landmarks(renderer, target_img, target_2d_lmks, flamelayer, cam, device
# , lmk_dist=0.0, shape_reg=0.0, exp_reg=0.0, neck_pose_reg=0.0, jaw_pose_reg=0.0,
# eyeballs_pose_reg=0.0)
# Fit with all Flame parameters parameters
vars = [flamelayer.transl, flamelayer.global_rot, flamelayer.shape_params, flamelayer.expression_params,
flamelayer.jaw_pose, flamelayer.neck_pose]
all_flame_params_optimizer = torch.optim.LBFGS(vars, tolerance_change=1e-7, max_iter=1500,
line_search_fn='strong_wolfe')
fit_flame_to_2D_landmarks_perspectiv(flamelayer, cam, target_2d_lmks, all_flame_params_optimizer)
# plotting
#plot_landmarks(renderer, target_img, target_2d_lmks, flamelayer, cam, device
# , lmk_dist=0.0, shape_reg=0.0, exp_reg=0.0, neck_pose_reg=0.0, jaw_pose_reg=0.0,
# eyeballs_pose_reg=0.0)
cv2.destroyAllWindows()
mesh = make_mesh(flamelayer, device)
# Render the mesh
rendered_mesh = renderer.render_phong(mesh)
rendered_mesh = rendered_mesh.detach().cpu().numpy().squeeze()
rendered_mesh = cv2.resize(rendered_mesh, (target_img.shape[0], target_img.shape[1]))
rendered_mesh = cv2.normalize(src=rendered_mesh, dst=None, alpha=0, beta=255, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_8U)
# show edges
edges_src = cv2.Canny(target_img,150,200)
edges_mesh = cv2.Canny(rendered_mesh,150,200)
edges_overlay = np.zeros(target_img.shape, dtype = np.uint8)
edges_overlay[:,:,0] = edges_src
edges_overlay[:,:,2] = edges_mesh
cv2.imshow('edges_overlay', edges_overlay)
#cv2.waitKey(0)
#cv2.destroyAllWindows()
#cv2.waitKey(1)
print ('edges_mesh =' , edges_mesh)
print ('edges_mesh.type = ', edges_mesh.dtype)
#print ('np.array_equal(edges_mesh, edges_mesh.astype(bool)) = ', np.array_equal(edges_mesh, edges_mesh.astype(bool)))
mesh_image_coords, target_img_coords = find_edge_images_correspondences(edges_mesh, edges_src)
print ('mesh_image_coords at ', mesh_image_coords)
print ('target_img_coords at ', target_img_coords)
cv2.circle(rendered_mesh, (mesh_image_coords[0],mesh_image_coords[1]), 4, (0, 0, 255), -1)
cv2.circle(target_img, (target_img_coords[0],target_img_coords[1]), 4, (0, 0, 255), -1)
edges = np.hstack((edges_src, edges_mesh))
cv2.imshow('edges', edges)
cv2.imshow('mesh and image', np.hstack((target_img,rendered_mesh[:, :, :3])))
cv2.waitKey(0)
cv2.destroyAllWindows()
cv2.waitKey(1)
#display_debug_output(renderer, target_img, flamelayer, device)
"""
all_flame_params_optimizer = torch.optim.LBFGS(vars, tolerance_change=1e-7, max_iter=1500,
line_search_fn='strong_wolfe')
target_silh = segment_img(target_img, 10)
vars = [flamelayer.transl, flamelayer.global_rot, flamelayer.shape_params, flamelayer.expression_params,
flamelayer.jaw_pose, flamelayer.neck_pose]
fit_flame_silhouette_perspectiv(flamelayer, renderer, target_silh, all_flame_params_optimizer, device)
plot_silhouette(flamelayer, renderer, target_silh,device)
"""
def display_debug_output(renderer, target_img, flamelayer, device):
cv2.destroyAllWindows()
mesh = make_mesh(flamelayer, device)
# Render the mesh
rendered_mesh = renderer.render_phong(mesh)
rendered_mesh = rendered_mesh.detach().cpu().numpy().squeeze()
rendered_mesh = cv2.resize(rendered_mesh, (target_img.shape[0], target_img.shape[1]))
rendered_mesh = cv2.normalize(src=rendered_mesh, dst=None, alpha=0, beta=255, norm_type=cv2.NORM_MINMAX, dtype=cv2.CV_8U)
cv2.imshow('rendered_mesh', rendered_mesh)
# show edges
edges_src = cv2.Canny(target_img,75,200)
edges_mesh = cv2.Canny(rendered_mesh,75,200)
edges = np.hstack((edges_src, edges_mesh))
cv2.imshow('edges', edges)
edges_overlay = np.zeros(target_img.shape, dtype = np.uint8)
edges_overlay[:,:,0] = edges_src
edges_overlay[:,:,2] = edges_mesh
cv2.imshow('edges_overlay', edges_overlay)
cv2.waitKey(0)
cv2.destroyAllWindows()
cv2.waitKey(1)
if __name__ == '__main__':
parser.add_argument(
'--target_img_path',
type=str,
default='./data/bareteeth.000001.26_C.jpg',
help='Target image path')
parser.add_argument(
'--out_path',
type=str,
default='./Results',
help='Results folder path')
parser.add_argument(
'--texture_mapping',
type=str,
default='./data/texture_data.npy',
help='Texture data')
config = get_config()
config.batch_size = 1
config.flame_model_path = './model/male_model.pkl'
print('Running 2D landmark fitting')
image2model_pipline(config.texture_mapping, config.target_img_path, config.out_path)
# ####################################
# plt_target_lmks = target_2d_lmks_oj.copy()
# for (x, y) in plt_target_lmks:
# print('x,y -> ', x, y)
# cv2.circle(target_img, (int(x), int(y)), 4, (0, 0, 255), -1)
#
# plt_target_lmks2 = coord_transformer.cam2screen(target_2d_lmks)
# cv2.imshow('baby', target_img)
# cv2.waitKey()
# for (x, y) in plt_target_lmks2:
# print('x,y -> ', x, y)
# cv2.circle(target_img, (int(x), int(y)), 10, (0, 255, 0), -1)
# cv2.imshow('baby', target_img)
# cv2.waitKey()
# #####################################