-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathSkeleton.py
282 lines (245 loc) · 9.81 KB
/
Skeleton.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
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
import os
import numpy as np
import cv2
import mediapipe as mp
#import gc
from Vec3 import *
#mp_pose = mp.solutions.pose
#mp_pose = mp.solutions.pose
#mp_pose_detector = mp.solutions.pose.Pose() # en global pour éviter de le recréer à chaque fois dans la classe, sinon mettre en static
mp_pose_detector = mp.solutions.pose.Pose(
static_image_mode=False, # False is for video sequence
model_complexity=2,
smooth_landmarks=True,
enable_segmentation=False,
smooth_segmentation=False,
min_detection_confidence=0.5,
min_tracking_confidence=0.5)
class Skeleton:
""" class with a skeleton
tab de Vec3
# Full skeleton
https://developers.google.com/mediapipe/solutions/vision/pose_landmarker
0 - nose
1 - left eye (inner)
2 - left eye
3 - left eye (outer)
4 - right eye (inner)
5 - right eye
6 - right eye (outer)
7 - left ear
8 - right ear
9 - mouth (left)
10 - mouth (right)
11 - left shoulder
12 - right shoulder
13 - left elbow
14 - right elbow
15 - left wrist
16 - right wrist
17 - left pinky
18 - right pinky
19 - left index
20 - right index
21 - left thumb
22 - right thumb
23 - left hip
24 - right hip
25 - left knee
26 - right knee
27 - left ankle
28 - right ankle
29 - left heel
30 - right heel
31 - left foot index
32 - right foot index
# Reduced skeleton
==> reduce 0, 11, 12, 13, 14, 15, 16, 23, 24, 25, 26, 27, 28
0 head
1 left shoulder
2 right shoulder
3 left elbow
4 right elbow
5 left wrist
6 right wrist
7 left hip
8 right hip
9 left knee
10 right knee
11 left ankle
12 right ankle
"""
reduce_indice = [0, 11, 12, 13, 14, 15, 16, 23, 24, 25, 26, 27, 28]
dim = 33
full_dim = 3*dim
reduced_dim = len(reduce_indice)*2
colors_rgb = np.array([
[255, 0, 0], # Rouge
[0, 255, 0], # Vert
[0, 0, 255], # Bleu
[255, 255, 0], # Jaune
[255, 0, 255], # Magenta
[0, 255, 255], # Cyan
[128, 0, 0], # Marron
[0, 128, 0], # Vert foncé
[0, 0, 128], # Bleu foncé
[255, 128, 0], # Orange
[128, 0, 128], # Pourpre
[128, 128, 0], # Olive
[0, 128, 128], # Sarcelle
[128, 128, 128], # Gris
[192, 192, 192], # Gris clair
[255, 165, 0], # Or
[165, 42, 42], # Brun
[0, 128, 192], # Bleu acier
[128, 0, 128], # Indigo
[128, 0, 0], # Acajou
[128, 128, 0], # Olive
[0, 128, 0], # Vert foncé
[0, 128, 128], # Sarcelle
[0, 0, 128], # Bleu foncé
[0, 165, 255], # Bleu royal
[165, 42, 42], # Brun
[255, 140, 0], # Orange rouge
[0, 250, 154], # Vert clair
[75, 0, 130], # Indigo foncé
[0, 255, 255], # Cyan clair
[218, 112, 214], # Orchidée
[210, 105, 30], # Chocolat
[240, 230, 140], # Kaki
[255, 20, 147], # Rose profond
], dtype=np.uint8)
def __init__(self, ske=None):
if ske is not None:
self.ske = ske
else:
self.ske = np.empty( Skeleton.dim, dtype=Vec3) # 33 is the size of mediapipe skeleton
for i in range(Skeleton.dim):
self.ske[i] = Vec3(0,0,0)
def __str__(self):
return str(self.ske)
def __array__(self, dtype=None, reduced=False):
""" return skeleton as a numpy array of float, if reduced is True, keep only 13 minimals joints """
if reduced:
return np.vstack( self.ske[self.reduce_indice] ).astype(float)[:, :2]
else:
return np.vstack( self.ske ).astype(float)
def reduce(self):
return self.__array__(reduced=True)
def fromImage(self, image):
""" get skeleton from image """
#results = self.pose.process(image)
results = mp_pose_detector.process(image)
if results.pose_landmarks is None:
return False
if results.pose_landmarks:
for index, landmark in enumerate(results.pose_landmarks.landmark):
self.ske[index] = Vec3(landmark.x, landmark.y, landmark.z)
ok = len(results.pose_landmarks.landmark)==Skeleton.dim
results.pose_landmarks.Clear() # free memory of mo
# del results
# gc.collect()
return ok
def crop(self, x,y,w,h):
""" crop skeleton """
for i in range(Skeleton.dim):
self.ske[i].x = (self.ske[i].x - x) / w
self.ske[i].y = (self.ske[i].y - y) / h
def boundingBox(self):
""" get bounding box of skeleton """
minx, maxx = 1, 0
miny, maxy = 1, 0
for i in range(Skeleton.dim):
minx = min(minx, self.ske[i].x)
maxx = max(maxx, self.ske[i].x)
miny = min(miny, self.ske[i].y)
maxy = max(maxy, self.ske[i].y)
return minx, miny, maxx, maxy
def distance(self, ske): # TP-TODO
""" distance between two skeletons """
d = 0.0
for i in range(Skeleton.dim):
d += norm( self.ske[i]-ske.ske[i])
return d
def draw(self, image):
""" draw skeleton on image """
image.flags.writeable = True
#image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
height,width,_ = image.shape
for i in range(Skeleton.dim):
x, y = int(self.ske[i].x * width), int(self.ske[i].y * height)
cv2.circle(image, (x,y), 3, Skeleton.colors_rgb[i].tolist() , -1)
#cv2.circle(image, (x,y), 3, (0, 0, 255), -1)
#cv2.line(image, (100, 100), (500, 500), (0, 255, 0), 2)
Skeleton.draw_reduced(self.reduce(), image)
def CoM(self,w=1,h=1):
moyenne = self.ske.mean()
moyenne = np.array( [moyenne[0] * w, moyenne[1] * h ] )
return moyenne.astype(int)
@staticmethod
def neck(ske,w,h):
ls = np.array( [int(ske[1][0] * w), int(ske[1][1] * h) ], dtype=int )
rs = np.array( [int(ske[2][0] * w), int(ske[2][1] * h) ], dtype=int )
return (0.5*(ls+rs)).astype(int)
@staticmethod
def pelvis(ske,w,h):
lh = np.array( [ int(ske[7][0] * w), int(ske[7][1] * h) ], dtype=int )
rh = np.array( [ int(ske[8][0] * w), int(ske[8][1] * h) ], dtype=int )
return (0.5*(lh+rh)).astype(int)
@staticmethod
def joint(ske,w,h,idx):
return np.array( [ ske[idx][0] * w, ske[idx][1] * h ], dtype=int )
@staticmethod
def color(idx):
return ( int(Skeleton.colors_rgb[idx][0]), int(Skeleton.colors_rgb[idx][1]), int(Skeleton.colors_rgb[idx][2]) )
@staticmethod
def draw_reduced(skr, image):
""" draw reduced skeleton skr (numpy) on image
# 0 head
# 1 left shoulder
# 2 right shoulder
# 3 left elbow
# 4 right elbow
# 5 left wrist
# 6 right wrist
# 7 left hip
# 8 right hip
# 9 left knee
# 10 right knee
# 11 left ankle
# 12 right ankle
"""
image.flags.writeable = True
h,w,_ = image.shape
pelvis = tuple(Skeleton.pelvis(skr,w,h))
neck = tuple(Skeleton.neck(skr,w,h))
cv2.line(image, pelvis, neck, Skeleton.color(0), 4)
cv2.line(image, Skeleton.joint(skr,w,h,3), Skeleton.joint(skr,w,h,1), Skeleton.color(1), 4)
cv2.line(image, Skeleton.joint(skr,w,h,5), Skeleton.joint(skr,w,h,3), Skeleton.color(2), 4)
cv2.line(image, Skeleton.joint(skr,w,h,2), Skeleton.joint(skr,w,h,4), Skeleton.color(3), 4)
cv2.line(image, Skeleton.joint(skr,w,h,4), Skeleton.joint(skr,w,h,6), Skeleton.color(4), 4)
cv2.line(image, neck, Skeleton.joint(skr,w,h,1), Skeleton.color(5), 4)
cv2.line(image, neck, Skeleton.joint(skr,w,h,2), Skeleton.color(6), 4)
cv2.line(image, neck, Skeleton.joint(skr,w,h,0), Skeleton.color(7), 4)
cv2.line(image, Skeleton.joint(skr,w,h,7), pelvis, Skeleton.color(8), 4)
cv2.line(image, Skeleton.joint(skr,w,h,8), pelvis, Skeleton.color(9), 4)
cv2.line(image, Skeleton.joint(skr,w,h,8), Skeleton.joint(skr,w,h,10), Skeleton.color(10), 4)
cv2.line(image, Skeleton.joint(skr,w,h,10), Skeleton.joint(skr,w,h,12), Skeleton.color(11), 4)
cv2.line(image, Skeleton.joint(skr,w,h,7), Skeleton.joint(skr,w,h,9), Skeleton.color(12), 4)
cv2.line(image, Skeleton.joint(skr,w,h,9), Skeleton.joint(skr,w,h,11), Skeleton.color(13), 4)
if __name__ == '__main__':
s = Skeleton()
print("Current Working Directory:", os.getcwd())
image = cv2.imread("data/test.jpg")
if image is None:
print('Lecture de l\'image a échoué.')
#image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
s.fromImage(image)
#print(s)
print( "landmarks:", s )
print( "landmarks as np:", s.__array__() )
print( "landmarks as np:", s.__array__(reduced=True) )
s.draw(image)
cv2.imshow('Image', image)
cv2.waitKey(0)
cv2.destroyAllWindows()