def estimation(fid, f): # @UnusedVariable
shape = [50, 50]
diffeo = diffeomorphism_from_function(shape, f)
K = 50
epsilon = 1
de = DiffeomorphismEstimator([0.2, 0.2], MATCH_CONTINUOUS)
for y0, y1 in generate_input(shape, K, diffeo, epsilon=epsilon):
de.update(y0, y1)
diff2d = de.summarize()
diffeo_learned = diff2d.d
from reprep import Report
name = f.__name__
r = Report(name)
fig = r.figure(cols=4)
diffeo_learned_rgb = diffeomorphism_to_rgb_cont(diffeo_learned)
diffeo_rgb = diffeomorphism_to_rgb_cont(diffeo)
r.data_rgb('diffeo_rgb', diffeo_rgb).add_to(fig)
r.data_rgb('diffeo_learned_rgb', diffeo_learned_rgb).add_to(fig)
L = r.data('diffeo_learned_uncertainty', diff2d.variance)
L.display('scale').add_to(fig, caption='uncertainty')
r.data('last_y0', y0).display('scale').add_to(fig, caption='last y0')
r.data('last_y1', y1).display('scale').add_to(fig, caption='last y1')
cs = [(0, 25), (10, 25), (25, 25), (25, 5)]
for c in cs:
M25 = de.get_similarity(c)
r.data('cell-%s-%s' % c,
M25).display('scale').add_to(fig,
caption='Example similarity field')
filename = 'out/diffeo_estimation_suite/%s.html' % name
print('Writing to %r.' % filename)
r.to_html(filename)
def estimation(fid, f): # @UnusedVariable
shape = [50, 50]
diffeo = diffeomorphism_from_function(shape, f)
K = 50
epsilon = 1
de = DiffeomorphismEstimator([0.2, 0.2], MATCH_CONTINUOUS)
for y0, y1 in generate_input(shape, K, diffeo, epsilon=epsilon):
de.update(y0, y1)
diff2d = de.summarize()
diffeo_learned = diff2d.d
from reprep import Report
name = f.__name__
r = Report(name)
fig = r.figure(cols=4)
diffeo_learned_rgb = diffeomorphism_to_rgb_cont(diffeo_learned)
diffeo_rgb = diffeomorphism_to_rgb_cont(diffeo)
r.data_rgb('diffeo_rgb', diffeo_rgb).add_to(fig)
r.data_rgb('diffeo_learned_rgb', diffeo_learned_rgb).add_to(fig)
L = r.data('diffeo_learned_uncertainty', diff2d.variance)
L.display('scale').add_to(fig, caption='uncertainty')
r.data('last_y0', y0).display('scale').add_to(fig, caption='last y0')
r.data('last_y1', y1).display('scale').add_to(fig, caption='last y1')
cs = [(0, 25), (10, 25), (25, 25), (25, 5)]
for c in cs:
M25 = de.get_similarity(c)
r.data('cell-%s-%s' % c, M25).display('scale').add_to(fig,
caption='Example similarity field')
filename = 'out/diffeo_estimation_suite/%s.html' % name
print('Writing to %r.' % filename)
r.to_html(filename)
def dd_get(id_diffeo):
shape = (64, 64)
func = DiffeoLibrary.diffeos[id_diffeo]
D = diffeomorphism_from_function(shape, func)
return (id_diffeo, D)
