/
curt_mypydmd3.py
executable file
·216 lines (179 loc) · 6.92 KB
/
curt_mypydmd3.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
#!/usr/bin/env python3
import argparse
import cv2
import json
import matplotlib.pyplot as plt
import numpy as np
import os
import skvideo.io # pip install scikit-video
from pydmd import DMD
parser = argparse.ArgumentParser(description='virtual choir')
parser.add_argument('video', help='video file')
parser.add_argument('--scale', type=float, default=1.0, help='scale input')
parser.add_argument("--skip-frames", type=int, default=0)
parser.add_argument('--write', action='store_true', help='write out video')
args = parser.parse_args()
# pathname work
abspath = os.path.abspath(args.video)
filename, ext = os.path.splitext(abspath)
dirname = os.path.dirname(args.video)
mode_video = filename + "_modes.mp4"
metadata = skvideo.io.ffprobe(args.video)
#print(metadata.keys())
print(json.dumps(metadata["video"], indent=4))
fps_string = metadata['video']['@avg_frame_rate']
(num, den) = fps_string.split('/')
fps = float(num) / float(den)
codec = metadata['video']['@codec_long_name']
#w = int(round(int(metadata['video']['@width']) * scale))
#h = int(round(int(metadata['video']['@height']) * scale))
if "@duration" in metadata["video"]:
total_frames = int(round(float(metadata['video']['@duration']) * fps))
else:
total_frames = 1
print('fps:', fps)
print('codec:', codec)
#print('output size:', w, 'x', h)
print('total frames:', total_frames)
print("Opening ", args.video)
reader = skvideo.io.FFmpegReader(args.video, inputdict={}, outputdict={})
rows = 3
cols = 3
max_rank = ((rows*cols) * 2) - 2
print("max rank:", max_rank)
# process video at this scale factor
scale = args.scale
dmd_size = 200
window_size = 64
def draw_mode(label, mode, shape):
real = np.abs(mode.real)
equalized = 255 * (real / np.max(real))
(h, w) = shape[:2]
big = cv2.resize(np.flipud(equalized.reshape((dmd_size,dmd_size)).astype('uint8')), (w, h), interpolation=cv2.INTER_AREA)
cv2.imshow(label, big)
return big
counter = 0
X = []
dmd = DMD(svd_rank=max_rank)
inputdict = {
'-r': str(fps)
}
lossless = {
# See all options: https://trac.ffmpeg.org/wiki/Encode/H.264
'-vcodec': 'libx264', # use the h.264 codec
'-crf': '0', # set the constant rate factor to 0, (lossless)
'-preset': 'veryslow', # maximum compression
'-r': str(fps) # match input fps
}
sane = {
# See all options: https://trac.ffmpeg.org/wiki/Encode/H.264
'-vcodec': 'libx264', # use the h.264 codec
'-crf': '17', # visually lossless (or nearly so)
'-preset': 'medium', # default compression
'-r': str(fps) # match input fps
}
if args.write:
mode_writer = skvideo.io.FFmpegWriter(mode_video, inputdict=inputdict,
outputdict=sane)
print("collecting video frames")
for frame in reader.nextFrame():
counter += 1
if counter <= args.skip_frames:
continue
frame = frame[:,:,::-1] # convert from RGB to BGR (to make opencv happy)
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
scaled = cv2.resize(gray, (0,0), fx=scale, fy=scale, interpolation=cv2.INTER_AREA)
cv2.imshow("input", scaled)
small = cv2.resize(scaled, (dmd_size,dmd_size), interpolation=cv2.INTER_AREA)
X.append( np.flipud(small) )
while len(X) > window_size:
del X[0]
dmd.fit(np.array(X))
print(dmd.modes.shape)
if len(dmd.eigs):
#print(dmd.eigs)
idx = np.argsort(np.abs(dmd.eigs-1))
#idx = np.argsort(np.abs(dmd.eigs.imag))
print(idx)
print(dmd.eigs)
print(dmd.eigs[idx[0]])
print(dmd.reconstructed_data.shape)
#for i in range(len(idx)):
# draw_mode("freq index: %d" % i, dmd.modes[:,idx[i]], scaled.shape)
big = 255 * dmd.reconstructed_data[:,-1] / np.max(dmd.reconstructed_data[:,-1]) # avoid overflow
big = cv2.resize(np.flipud(big.reshape((dmd_size,dmd_size)).astype('uint8')), (scaled.shape[1], scaled.shape[0]), interpolation=cv2.INTER_AREA)
big = 255 * ( big / np.max(big) )
cv2.imshow("reconstructed", big.astype('uint8'))
def draw_text(img, label, x, y, subscale=1.0, just="center"):
font_scale = subscale * h / 700
size = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX,
font_scale, 1)
if just == "center":
locx = int(x - size[0][0]*0.5)
locy = int(y + size[0][1]*1.5)
elif just == "lower-right":
locx = int(x - size[0][0])
locy = int(y - size[0][1])
cv2.putText(img, label, (locx, locy),
cv2.FONT_HERSHEY_SIMPLEX, font_scale, (255, 255, 255),
1, cv2.LINE_AA)
(h, w) = scaled.shape[:2]
grid = np.zeros( (h*rows, w*cols) ).astype('uint8')
grid[0:h,0:w] = scaled
draw_text(grid, "Original", w*0.5, 0)
r = 0
c = 1
for i in range(0, max_rank, 2):
if i >= len(idx):
break
#print(i)
if c >= cols:
r += 1
c = 0
#print("grid:", r, c, "i:", i)
grid[r*h:(r+1)*h,c*w:(c+1)*w] = draw_mode("a", dmd.modes[:,idx[i]], scaled.shape)
#grid[r*h:(r+1)*h,c*w:(c+1)*w] = scaled
eig = dmd.eigs[idx[i]]
label = "Mode: %d (%.4f + %.4fj)" % (i, eig.real, eig.imag)
draw_text(grid, label, (c+0.5)*w, r*h)
c += 1
draw_text(grid, "www.uav.aem.umn.edu", w*(rows-0.03), h*(cols-0.03), just="lower-right")
cv2.imshow("grid", grid)
if args.write:
mode_writer.writeFrame(grid)
if 0xFF & cv2.waitKey(1) == 27:
break
if False:
(h, w) = X[0].shape
print(w,h)
print("running dmd")
dmd = DMD(svd_rank=5)
dmd.fit(np.array(X))
dmd.plot_modes_2D(figsize=(12,5))
print(X[0].shape[0])
x1 = np.array(list(range(w)))
x2 = np.array(list(range(h)))
x1grid, x2grid = np.meshgrid(x1, x2)
fig = plt.figure(figsize=(18,12))
for id_subplot, snapshot in enumerate(dmd.reconstructed_data.T[:16], start=1):
plt.subplot(4, 4, id_subplot)
plt.pcolor(x1grid, x2grid, snapshot.reshape(x1grid.shape).real, vmin=-1, vmax=1)
for eig in dmd.eigs:
print('Eigenvalue {}: distance from unit circle {}'.format(eig, np.abs(eig.imag**2+eig.real**2 - 1)))
dmd.plot_eigs(show_axes=True, show_unit_circle=True)
for mode in dmd.modes.T:
plt.plot(mode.real)
plt.title('Modes')
plt.show()
for dynamic in dmd.dynamics:
plt.plot(dynamic.real)
plt.title('Dynamics')
plt.show()
fig = plt.figure(figsize=(17,6))
for n, mode, dynamic in zip(range(131, 133), dmd.modes.T, dmd.dynamics):
plt.subplot(n)
plt.pcolor(x1grid, x2grid, (mode.reshape(-1, 1).dot(dynamic.reshape(1, -1))).real.T)
plt.subplot(133)
plt.pcolor(x1grid, x2grid, dmd.reconstructed_data.T.real)
plt.colorbar()
plt.show()