forked from ericmjonas/franktrack
/
bruteforcelikelihood.py
350 lines (282 loc) · 12.1 KB
/
bruteforcelikelihood.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
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
"""
Code to brute-force compare the likelihood model
"""
import numpy as np
import scipy.stats
import scipy.ndimage
import cPickle as pickle
import likelihood
import util2 as util
import model
import time
from matplotlib import pylab
import cloud
import plotparticles
import os
import organizedata
import videotools
import glob
import measure
import template
from ruffus import *
import pf
PIX_THRESHOLD = 0
FL_DATA = "data/fl"
X_GRID_NUM = 150
Y_GRID_NUM = 150
PHI_GRID_NUM = 32
THETA_GRID_NUM = 6
USE_CLOUD = True
#cloud.start_simulator()
LIKELIHOOD_SETTING = [{'similarity' : 'dist',
'sim_params' : {'power' : 1}},
{'similarity' : 'normcc',
'sim_params' : {'scalar': 10}}]
def params():
EPOCHS = [#'bukowski_04.W1', 'bukowski_04.W2',
#'bukowski_03.W1',
'bukowski_03.W2',
#'bukowski_04.C', 'bukowski_03.C',
#'bukowski_03.linear', 'bukowski_04.linear'
]
#'bukowski_04.C', 'bukowski_04.linear']
FRAMES = np.arange(10)
for epoch in EPOCHS:
for frame in FRAMES:
for likelihood_i in range(len(LIKELIHOOD_SETTING)):
infiles = [os.path.join(FL_DATA, epoch),
os.path.join(FL_DATA, epoch, 'config.pickle'),
os.path.join(FL_DATA, epoch, 'framehist.npz'),
]
basename = '%s.likelihoodscores.%02d.%05d.%d' % (epoch, likelihood_i, frame, PIX_THRESHOLD)
outfiles = [basename + ".wait.pickle",
basename + ".wait.npz"]
yield (infiles, outfiles, epoch, frame, likelihood_i)
@files(params)
def score_frame_queue((dataset_dir, dataset_config_filename,
frame_hist_filename), (outfile_wait,
outfile_npz), dataset_name, frame, likelihood_i):
np.random.seed(0)
dataset_dir = os.path.join(FL_DATA, dataset_name)
cf = pickle.load(open(dataset_config_filename))
led_params = pickle.load(open(os.path.join(dataset_dir, "led.params.pickle")))
EO = measure.led_params_to_EO(cf, led_params)
x_range = np.linspace(0, cf['field_dim_m'][1], X_GRID_NUM)
y_range = np.linspace(0, cf['field_dim_m'][0], Y_GRID_NUM)
phi_range = np.linspace(0, 2*np.pi, PHI_GRID_NUM)
degrees_from_vertical = 30
radian_range = degrees_from_vertical/180. * np.pi
theta_range = np.linspace(np.pi/2.-radian_range,
np.pi/2. + radian_range, THETA_GRID_NUM)
sv = create_state_vect(y_range, x_range, phi_range, theta_range)
# now the input args
chunk_size = 80000
chunks = int(np.ceil(len(sv) / float(chunk_size)))
args = []
for i in range(chunks):
args += [ (i*chunk_size, (i+1)*chunk_size)]
CN = chunks
results = []
if USE_CLOUD:
print "MAPPING TO THE CLOUD"
jids = cloud.map(picloud_score_frame, [dataset_name]*CN,
[x_range]*CN, [y_range]*CN,
[phi_range]*CN, [theta_range]*CN,
args, [frame]*CN, [EO]*CN, [likelihood_i]*CN,
_type='f2', _vol="my-vol", _env="base/precise")
else:
jids = map(picloud_score_frame, [dataset_name]*CN,
[x_range]*CN, [y_range]*CN,
[phi_range]*CN, [theta_range]*CN,
args, [frame]*CN, [EO]*CN, [likelihood_i]*CN)
np.savez_compressed(outfile_npz,
x_range = x_range, y_range=y_range,
phi_range = phi_range, theta_range = theta_range)
pickle.dump({'frame' : frame,
'dataset_name' : dataset_name,
'dataset_dir' : dataset_dir,
'jids' : jids}, open(outfile_wait, 'w'))
@transform(score_frame_queue, regex(r"(.+).wait.(.+)$"), [r"\1.pickle", r"\1.npz"])
def score_frame_wait((infile_wait, infile_npz), (outfile_pickle, outfile_npz)):
dnpz = np.load(infile_npz)
p = pickle.load(open(infile_wait))
jids = p['jids']
if USE_CLOUD:
results = cloud.result(jids)
else:
results = [x for x in jids]
scores = np.concatenate(results)
np.savez_compressed(outfile_npz, scores=scores, **dnpz)
pickle.dump(p, open(outfile_pickle, 'w'))
@transform(score_frame_wait, suffix(".pickle"), [".png", ".hist.png"])
def plot_likelihood((infile_pickle, infile_npz),
(outfile, outfile_hist)):
data = np.load(infile_npz)
data_p = pickle.load(open(infile_pickle))
scores = data['scores']
sv = create_state_vect(data['y_range'], data['x_range'],
data['phi_range'], data['theta_range'])
scores = scores[:len(sv)]
pylab.figure()
scores_flat = np.array(scores.flat)
pylab.hist(scores_flat[np.isfinite(scores_flat)], bins=255)
pylab.savefig(outfile_hist, dpi=300)
scores[np.isinf(scores)] = -1e20
TOP_R, TOP_C = 3, 4
TOP_N = TOP_R * TOP_C
score_idx_sorted = np.argsort(scores)[::-1]
#get the frame
frames = organizedata.get_frames(data_p['dataset_dir'],
np.array([data_p['frame']]))
# config file
cf = pickle.load(open(os.path.join(data_p['dataset_dir'],
'config.pickle')))
env = util.Environmentz(cf['field_dim_m'],
cf['frame_dim_pix'])
img = frames[0]
f = pylab.figure()
for r in range(TOP_N):
s_i = score_idx_sorted[r]
score = scores[s_i]
ax =f.add_subplot(TOP_R, TOP_C, r+1)
ax.imshow(img, interpolation='nearest', cmap=pylab.cm.gray)
x_pix, y_pix = env.gc.real_to_image(sv[s_i]['x'], sv[s_i]['y'])
ax.axhline(y_pix, linewidth=1, c='b', alpha=0.5)
ax.axvline(x_pix, linewidth=1, c='b', alpha=0.5)
ax.set_xticks([])
ax.set_yticks([])
f.subplots_adjust(bottom=0, left=.01, right=.99, top=.90, hspace=.1, wspace=.1)
f.savefig(outfile, dpi=300)
@transform(score_frame_wait, suffix(".pickle"), [".zoom.png"])
def plot_likelihood_zoom((infile_pickle, infile_npz),
(zoom_outfile, )):
"""
zoom in on the region of interest
plot front and back diodes
"""
data = np.load(infile_npz)
data_p = pickle.load(open(infile_pickle))
scores = data['scores']
sv = create_state_vect(data['y_range'], data['x_range'],
data['phi_range'], data['theta_range'])
scores = scores[:len(sv)]
TOP_R, TOP_C = 3, 4
TOP_N = TOP_R * TOP_C
score_idx_sorted = np.argsort(scores)[::-1]
#get the frame
frames = organizedata.get_frames(data_p['dataset_dir'],
np.array([data_p['frame']]))
# config file
cf = pickle.load(open(os.path.join(data_p['dataset_dir'],
'config.pickle')))
env = util.Environmentz(cf['field_dim_m'],
cf['frame_dim_pix'])
tp = template.TemplateRenderCircleBorder()
led_params = pickle.load(open(os.path.join(data_p['dataset_dir'],
"led.params.pickle")))
EO_PARAMS = measure.led_params_to_EO(cf, led_params)
tp.set_params(*EO_PARAMS)
img = frames[0]
img_thold = img.copy()
img_thold[img < PIX_THRESHOLD] = 0
f = pylab.figure(figsize=(12, 8))
X_MARGIN = 30
Y_MARGIN = 20
for row in range(TOP_R):
for col in range(TOP_C):
r = row * TOP_C + col
s_i = score_idx_sorted[r]
score = scores[s_i]
ax = pylab.subplot2grid((TOP_R *2, TOP_C*2), (row*2, col*2))
ax.imshow(img, interpolation='nearest', cmap=pylab.cm.gray)
ax_thold = pylab.subplot2grid((TOP_R*2, TOP_C*2),
(row*2+1, col*2))
ax_thold.imshow(img_thold, interpolation='nearest',
cmap=pylab.cm.gray,
vmin=0, vmax=255)
x = sv[s_i]['x']
y = sv[s_i]['y']
phi = sv[s_i]['phi']
theta = sv[s_i]['theta']
x_pix, y_pix = env.gc.real_to_image(x, y)
# render the fake image
ax_generated = pylab.subplot2grid((TOP_R*2, TOP_C*2),
(row*2+1, col*2 + 1))
rendered_img = tp.render(phi, theta)
ax_generated.imshow(rendered_img*255, interpolation='nearest',
cmap=pylab.cm.gray,
vmin = 0, vmax=255)
# now compute position of diodes
front_pos, back_pos = util.compute_pos(tp.length, x_pix, y_pix,
phi, theta)
cir = pylab.Circle(front_pos, radius=EO_PARAMS[1], ec='g', fill=False,
linewidth=2)
ax.add_patch(cir)
cir = pylab.Circle(back_pos, radius=EO_PARAMS[2], ec='r', fill=False,
linewidth=2)
ax.add_patch(cir)
ax.set_title("%2.2f, %2.2f, %1.1f, %1.1f, %4.2f" % (x, y, phi, theta, score), size="xx-small")
for a in [ax, ax_thold]:
a.set_xticks([])
a.set_yticks([])
a.set_xlim(x_pix - X_MARGIN, x_pix+X_MARGIN)
a.set_ylim(y_pix - Y_MARGIN, y_pix+Y_MARGIN)
f.subplots_adjust(bottom=0, left=.01, right=.99, top=.90, hspace=.2, wspace=.1)
f.savefig(zoom_outfile, dpi=300)
def create_state_vect(y_range, x_range, phi_range, theta_range):
N = len(y_range) * len(x_range) * len(phi_range) * len(theta_range)
state = np.zeros(N, dtype=util.DTYPE_STATE)
i = 0
for yi, y in enumerate(y_range):
for xi, x in enumerate(x_range):
for phii, phi in enumerate(phi_range):
for thetai, theta in enumerate(theta_range):
state['x'][i] = x
state['y'][i] = y
state['phi'][i] = phi
state['theta'][i] = theta
i += 1
return state
def picloud_score_frame(dataset_name, x_range, y_range, phi_range, theta_range,
state_idx, frame, EO_PARAMS, likelihood_i):
"""
pi-cloud runner, every instance builds up full state, but
we only evaluate the states in [state_idx_to_eval[0], state_idx_to_eval[1])
and return scores
"""
print "DATSET_NAME=", dataset_name
dataset_dir = os.path.join(FL_DATA, dataset_name)
dataset_config_filename = os.path.join(dataset_dir, "config.pickle")
dataset_region_filename = os.path.join(dataset_dir, "region.pickle")
frame_hist_filename = os.path.join(dataset_dir, "framehist.npz")
np.random.seed(0)
cf = pickle.load(open(dataset_config_filename))
region = pickle.load(open(dataset_region_filename))
framehist = np.load(frame_hist_filename)
env = util.Environmentz(cf['field_dim_m'],
cf['frame_dim_pix'])
tp = template.TemplateRenderCircleBorder()
tp.set_params(*EO_PARAMS)
ls = LIKELIHOOD_SETTING[likelihood_i]
le = likelihood.LikelihoodEvaluator2(env, tp, similarity=ls['similarity'],
sim_params = ls['sim_params'])
frames = organizedata.get_frames(dataset_dir, np.array([frame]))
frame = frames[0]
frame[frame < PIX_THRESHOLD] = 0
# create the state vector
state = create_state_vect(y_range, x_range, phi_range, theta_range)
SCORE_N = state_idx[1] - state_idx[0]
scores = np.zeros(SCORE_N, dtype=np.float32)
for i, state_i in enumerate(state[state_idx[0]:state_idx[1]]):
x = state_i['x']
y = state_i['y']
if region['x_pos_min'] <= x <= region['x_pos_max'] and \
region['y_pos_min'] <= y <= region['y_pos_max']:
score = le.score_state(state_i, frame)
scores[i] = score
else:
scores[i] = -1e100
return scores
if __name__ == "__main__":
pipeline_run([score_frame_wait, plot_likelihood,
plot_likelihood_zoom], multiprocess=6)