def make_test_pairwise_fetaures(case1, case2, label, rng):
import vtool as vt
mu_fm = 50 if label == 1 else 10
sigma_fm = 10 if label == 1 else 20
mu_fs = 0.2 if label == 1 else 0.4
sigma_fs = 0.1 if label == 1 else 0.1
num_top = 4
max_feats = min(case1['nfeats'], case2['nfeats'])
num_matches = (
np.clip(rng.normal(mu_fm, sigma_fm, size=1)[0], num_top + 1, max_feats).astype(
np.int
),
)
perb = np.abs(rng.normal(0.001, 0.001, size=num_matches))
sift_dists = np.clip(rng.normal(mu_fs, sigma_fs, size=num_matches), 0, 1) + perb
sortx = np.argsort(sift_dists)
local_feat_simvecs = np.vstack([sift_dists]).T[sortx]
local_simvec = local_feat_simvecs[0:num_top]
yaw1 = case1['yaw']
yaw2 = case2['yaw']
global_simvec = np.array(
[
case1['qual'],
case2['qual'],
yaw1 / vt.TAU,
yaw2 / vt.TAU,
vt.ori_distance(yaw1, yaw2) / np.pi,
np.abs(case1['qual'] - case1['qual']),
]
)
simvec = np.hstack([global_simvec, local_simvec.ravel()])
return simvec
def make_test_pairwise_labels2(cases1, cases2):
import vtool as vt
is_same = np.array(cases1['name']) == np.array(cases2['name'])
yaw1 = np.array(cases1['yaw'])
yaw2 = np.array(cases2['yaw'])
yaw_dist = vt.ori_distance(yaw1, yaw2) / np.pi
qual1 = np.array(cases1['qual'])
qual2 = np.array(cases2['qual'])
_and = np.logical_and
_or = np.logical_or
_not = np.logical_not
poorqual = _or(qual1 < 2, qual2 < 2)
bestqual = _and(_or(qual1 > 3, qual2 > 3), _not(poorqual))
goodqual = _and(_not(bestqual), _not(poorqual))
is_comp = _or.reduce(
[
_not(poorqual),
_and(bestqual, yaw_dist <= (1 / 4)),
_and(goodqual, yaw_dist <= (1 / 8)),
]
)
label = is_same.astype(np.int32)
label[_not(is_comp)] = 2
return label
def make_test_pairwise_fetaures(case1, case2, label, rng):
import vtool as vt
mu_fm = 50 if label == 1 else 10
sigma_fm = 10 if label == 1 else 20
mu_fs = .2 if label == 1 else .4
sigma_fs = .1 if label == 1 else .1
num_top = 4
max_feats = min(case1['nfeats'], case2['nfeats'])
num_matches = np.clip(rng.normal(mu_fm, sigma_fm, size=1)[0], num_top + 1, max_feats).astype(np.int),
perb = np.abs(rng.normal(.001, .001, size=num_matches))
sift_dists = np.clip(rng.normal(mu_fs, sigma_fs, size=num_matches), 0, 1) + perb
sortx = np.argsort(sift_dists)
local_feat_simvecs = np.vstack([sift_dists]).T[sortx]
local_simvec = local_feat_simvecs[0:num_top]
yaw1 = case1['yaw']
yaw2 = case2['yaw']
global_simvec = np.array([
case1['qual'],
case2['qual'],
yaw1 / vt.TAU,
yaw2 / vt.TAU,
vt.ori_distance(yaw1, yaw2) / np.pi,
np.abs(case1['qual'] - case1['qual']),
])
simvec = np.hstack([global_simvec, local_simvec.ravel()])
return simvec
def make_test_similarity(test_case):
# toy_params = {
# True: {'mu': 0.9, 'sigma': .1},
# False: {'mu': 0.1, 'sigma': .4}
# }
# tau = np.pi * 2
from wbia import constants as const
# view_to_ori = const.VIEWTEXT_TO_YAW_RADIANS
view_to_ori = ut.map_dict_keys(
lambda x: const.YAWALIAS[x], const.VIEWTEXT_TO_YAW_RADIANS
)
# view_to_ori = {
# 'F': -1 * tau / 4,
# 'L': 0 * tau / 4,
# 'B': 1 * tau / 4,
# 'R': 2 * tau / 4,
# }
import vtool as vt
nid_list = np.array(ut.dict_take_column(test_case, 'name'))
yaw_list = np.array(
ut.dict_take(view_to_ori, ut.dict_take_column(test_case, 'view'))
)
rng = np.random.RandomState(0)
pmat = []
for idx in range(len(test_case)):
nid = nid_list[idx]
yaw = yaw_list[idx]
p_same = nid == nid_list
p_comp = 1 - vt.ori_distance(yaw_list, yaw) / np.pi
# estimate noisy measurements
p_same_m = np.clip(p_same + rng.normal(0, 0.5, size=len(p_same)), 0, 0.9)
p_comp_m = np.clip(p_comp + rng.normal(0, 0.5, size=len(p_comp)), 0, 0.9)
#
p_same_and_comp = p_same_m * p_comp_m
pmat.append(p_same_and_comp)
#
P = np.array(pmat)
P[np.diag_indices(len(P))] = 0
P = P + P.T / 2
P = np.clip(P, 0.01, 0.99)
logger.info(ut.hz_str(' P = ', ut.repr2(P, precision=2, max_line_width=140)))
return P
def make_test_pairwise_labels(case1, case2):
import vtool as vt
is_same = case1['name'] == case2['name']
yaw1 = case1['yaw']
yaw2 = case2['yaw']
yaw_dist = vt.ori_distance(yaw1, yaw2) / np.pi
if case1['qual'] < 2 or case2['qual'] < 2:
# Bad quality means not comparable
is_comp = False
else:
if case1['qual'] > 3 or case2['qual'] > 3:
# Better quality, better chance of being comparable
is_comp = yaw_dist <= (1 / 4)
else:
is_comp = yaw_dist <= (1 / 8)
if is_comp:
label = int(is_same)
else:
label = 2
return label
def make_test_pairwise_labels(case1, case2):
import vtool as vt
is_same = case1['name'] == case2['name']
yaw1 = case1['yaw']
yaw2 = case2['yaw']
yaw_dist = vt.ori_distance(yaw1, yaw2) / np.pi
if case1['qual'] < 2 or case2['qual'] < 2:
# Bad quality means not comparable
is_comp = False
else:
if case1['qual'] > 3 or case2['qual'] > 3:
# Better quality, better chance of being comparable
is_comp = yaw_dist <= (1 / 4)
else:
is_comp = yaw_dist <= (1 / 8)
if is_comp:
label = int(is_same)
else:
label = 2
return label
def make_test_similarity(test_case):
#toy_params = {
# True: {'mu': 0.9, 'sigma': .1},
# False: {'mu': 0.1, 'sigma': .4}
#}
# tau = np.pi * 2
from ibeis import constants as const
# view_to_ori = const.VIEWTEXT_TO_YAW_RADIANS
view_to_ori = ut.map_dict_keys(lambda x: const.YAWALIAS[x], const.VIEWTEXT_TO_YAW_RADIANS)
# view_to_ori = {
# 'F': -1 * tau / 4,
# 'L': 0 * tau / 4,
# 'B': 1 * tau / 4,
# 'R': 2 * tau / 4,
# }
import vtool as vt
nid_list = np.array(ut.dict_take_column(test_case, 'name'))
yaw_list = np.array(ut.dict_take(view_to_ori, ut.dict_take_column(test_case, 'view')))
rng = np.random.RandomState(0)
pmat = []
for idx in range(len(test_case)):
nid = nid_list[idx]
yaw = yaw_list[idx]
p_same = nid == nid_list
p_comp = 1 - vt.ori_distance(yaw_list, yaw) / np.pi
# estimate noisy measurements
p_same_m = np.clip(p_same + rng.normal(0, .5, size=len(p_same)), 0, .9)
p_comp_m = np.clip(p_comp + rng.normal(0, .5, size=len(p_comp)), 0, .9)
#
p_same_and_comp = p_same_m * p_comp_m
pmat.append(p_same_and_comp)
#
P = np.array(pmat)
P[np.diag_indices(len(P))] = 0
P = P + P.T / 2
P = np.clip(P, .01, .99)
print(ut.hz_str(' P = ', ut.array_repr2(P, precision=2, max_line_width=140)))
return P
def make_test_pairwise_labels2(cases1, cases2):
import vtool as vt
is_same = np.array(cases1['name']) == np.array(cases2['name'])
yaw1 = np.array(cases1['yaw'])
yaw2 = np.array(cases2['yaw'])
yaw_dist = vt.ori_distance(yaw1, yaw2) / np.pi
qual1 = np.array(cases1['qual'])
qual2 = np.array(cases2['qual'])
_and = np.logical_and
_or = np.logical_or
_not = np.logical_not
poorqual = _or(qual1 < 2, qual2 < 2)
bestqual = _and(_or(qual1 > 3, qual2 > 3), _not(poorqual))
goodqual = _and(_not(bestqual), _not(poorqual))
is_comp = _or.reduce([_not(poorqual), _and(bestqual, yaw_dist <= (1 / 4)), _and(goodqual, yaw_dist <= (1 / 8))])
label = is_same.astype(np.int32)
label[_not(is_comp)] = 2
return label
