def test_render_at_all():
W = 320
H = 240
eo = likelihood.EvaluateObj(H, W)
eo.set_params(14, 5, 3)
i1 = eo.render_source(100, 100, np.pi /2., np.pi/2)
pylab.imshow(i1, interpolation='nearest', origin='lower',
vmin=0, vmax=1.0, cmap=pylab.cm.gray)
def test_render_source_coords():
W = 320
H = 240
eo = likelihood.EvaluateObj(H, W)
eo.set_params(14, 5, 3)
for x in range(0, W, 20):
for y in range(0, H, 20):
for phi in np.arange(0, 4*np.pi, 4):
for theta in np.arange(0, np.pi, 4):
i1 = eo.render_source(x, y, phi, theta)
def pf_run(infile, outfile, posnoise, velnoise, PARTICLEN, FRAMEN, NOISE, log):
np.random.seed(0)
print "Loading data..."
d = pickle.load(open(infile))
print "done!"
env = util.Environmentz((1.5, 2), (240, 320))
eo = likelihood.EvaluateObj(240, 320)
eo.set_params(10, 4, 2)
le = likelihood.LikelihoodEvaluator(env, eo, log)
model_inst = model.LinearModel(env,
le,
POS_NOISE_STD=posnoise,
VELOCITY_NOISE_STD=velnoise)
y = d['video'][:FRAMEN]
weights, particles = pf.particle_filter(y, model_inst, len(y), PARTICLEN)
np.savez_compressed(outfile, weights=weights, particles=particles)
def render(env, state, DIODE_SEP = 10,
FRONT_PIX = 4, BACK_PIX =2):
gc = env.gc
eo = likelihood.EvaluateObj(*env.image_dim)
eo.set_params(DIODE_SEP, FRONT_PIX, BACK_PIX)
images = np.zeros((len(state), env.image_dim[0],
env.image_dim[1]),
dtype=np.uint8)
for si, s in enumerate(state):
i_x, i_y = gc.real_to_image(s['x'], s['y'])
img = eo.render_source(i_x, i_y, s['phi'], s['theta'])
x = img*255
x.flat[x.flat > 255] = 255
images[si] = x
return images
0import numpy as np
import cPickle as pickle
from matplotlib import pylab
import likelihood
import util
import model
np.random.seed(0)
d = pickle.load(open('simulate.pickle'))
env = util.Environment((1.5, 2), (240, 320))
eo = likelihood.EvaluateObj(240, 320)
eo.set_params(10, 4, 2)
le = likelihood.LikelihoodEvaluator(env, eo)
model_inst = model.LinearModel(env, le)
PARTICLEN = 4000
for frameno in range(1, 300, 30):
print "frame", frameno
prior_states = model_inst.sample_latent_from_prior(PARTICLEN)
img = d['video'][frameno]
scores = np.zeros(PARTICLEN)
for si, state in enumerate(prior_states):
#pylab.imshow(x_normed[10], interpolation='nearest',
# cmap=pylab.cm.gray)
#pylab.show
np.savez_compressed('frames.npz', x_reg=x_reg, x_normed = x_normed)
# now
#plotting.plot_state_timeseries(state, DIODE_SEP)
#pylab.show()
IMG_WIDTH = 320
IMG_HEIGHT = 240
PPM = 50
gc = util.GeomConverter((IMG_WIDTH, IMG_HEIGHT),
(PPM, PPM), (0.0, 0.0))
eo = likelihood.EvaluateObj(IMG_WIDTH, IMG_HEIGHT)
eo.set_params(14, 4, 2)
frames = np.load('frames.npz')
phi_range = np.linspace(0.0, 2*np.pi, 8)
x_range = np.linspace(0.0, IMG_WIDTH/PPM*0.98, IMG_WIDTH/4)
y_range = np.linspace(0.0, IMG_HEIGHT/PPM*0.98, IMG_HEIGHT/4)
for frame_no, frame in enumerate(frames['x_normed']):
print frame_no
deltas = np.zeros((len(phi_range), len(x_range), len(y_range)), dtype=np.float)
for phi_i, phi in enumerate(phi_range):
for x_i, x in enumerate(x_range):
for y_i, y in enumerate(y_range):
i_x, i_y = gc.real_to_image(x, y)