def show_toy_distributions(toy_params):
import vtool as vt
import wbia.plottool as pt
pt.ensureqt()
xdata = np.linspace(0, 8, 1000)
tp_pdf = vt.gauss_func1d(xdata, **toy_params[True])
fp_pdf = vt.gauss_func1d(xdata, **toy_params[False])
pt.plot_probabilities(
[tp_pdf, fp_pdf],
['TP', 'TF'],
prob_colors=[pt.TRUE_BLUE, pt.FALSE_RED],
xdata=xdata,
figtitle='Toy Distributions',
)
def show_top_featmatches(qreq_, cm_list):
"""
Args:
qreq_ (wbia.QueryRequest): query request object with hyper-parameters
cm_list (list):
SeeAlso:
python -m wbia --tf TestResult.draw_feat_scoresep --show --db PZ_MTEST -t best:lnbnn_on=True,lnbnn_normalizer=normlnbnn-test -a default --sephack
python -m wbia --tf TestResult.draw_feat_scoresep --show --db PZ_Master1 -t best:lnbnn_on=True -a timectrl --sephack
python -m wbia --tf TestResult.draw_feat_scoresep --show --db PZ_MTEST -t best:lnbnn_on=True -a default:size=30 --sephack
python -m wbia --tf TestResult.draw_feat_scoresep --show --db PZ_MTEST -t best:K=1,Knorm=5,lnbnn_on=True -a default:size=30 --sephack
python -m wbia --tf TestResult.draw_feat_scoresep --show --db PZ_MTEST -t best:K=1,Knorm=3,lnbnn_on=True -a default --sephack
CommandLine:
python -m wbia.viz.viz_nearest_descriptors --exec-show_top_featmatches --show
Example:
>>> # DISABLE_DOCTEST
>>> from wbia.viz.viz_nearest_descriptors import * # NOQA
>>> import wbia
>>> cm_list, qreq_ = wbia.testdata_cmlist(defaultdb='PZ_MTEST',
>>> a=['default:has_none=mother,size=30'])
>>> show_top_featmatches(qreq_, cm_list)
>>> ut.quit_if_noshow()
>>> import wbia.plottool as pt
>>> ut.show_if_requested()
"""
# for cm in cm_list:
# cm.score_annot_csum(qreq_)
import numpy as np
import vtool as vt
from functools import partial
# Stack chipmatches
ibs = qreq_.ibs
infos = [cm.get_flat_fm_info() for cm in cm_list]
flat_metadata = dict([(k, np.concatenate(v))
for k, v in ut.dict_stack2(infos).items()])
fsv_flat = flat_metadata['fsv']
flat_metadata['fs'] = fsv_flat.prod(axis=1)
aids1 = flat_metadata['aid1'][:, None]
aids2 = flat_metadata['aid2'][:, None]
flat_metadata['aid_pairs'] = np.concatenate([aids1, aids2], axis=1)
# Take sample of metadata
sortx = flat_metadata['fs'].argsort()[::-1]
num = len(cm_list) * 3
# num = 10
taker = partial(np.take, indices=sortx[:num], axis=0)
flat_metadata_top = ut.map_dict_vals(taker, flat_metadata)
aid1s, aid2s, fms = ut.dict_take(flat_metadata_top, ['aid1', 'aid2', 'fm'])
annots = {}
aids = np.unique(np.hstack((aid1s, aid2s)))
annots = {
aid: ibs.get_annot_lazy_dict(aid, config2_=qreq_.qparams)
for aid in aids
}
label_lists = (ibs.get_match_truths(
aid1s, aid2s) == ibs.const.EVIDENCE_DECISION.POSITIVE)
patch_size = 64
def extract_patches(annots, aid, fxs):
""" custom_func(lazydict, key, subkeys) for multigroup_lookup """
annot = annots[aid]
kpts = annot['kpts']
rchip = annot['rchip']
kpts_m = kpts.take(fxs, axis=0)
warped_patches, warped_subkpts = vt.get_warped_patches(
rchip, kpts_m, patch_size=patch_size)
return warped_patches
data_lists = vt.multigroup_lookup(annots, [aid1s, aid2s], fms.T,
extract_patches)
import wbia.plottool as pt # NOQA
pt.ensureqt()
import wbia_cnn
inter = wbia_cnn.draw_results.interact_patches(
label_lists,
data_lists,
flat_metadata_top,
chunck_sizes=(2, 4),
ibs=ibs,
hack_one_per_aid=False,
sortby='fs',
qreq_=qreq_,
)
inter.show()
def crftest():
"""
pip install pyqpbo
pip install pystruct
http://taku910.github.io/crfpp/#install
cd ~/tmp
#wget https://drive.google.com/folderview?id=0B4y35FiV1wh7fngteFhHQUN2Y1B5eUJBNHZUemJYQV9VWlBUb3JlX0xBdWVZTWtSbVBneU0&usp=drive_web#list
7z x CRF++-0.58.tar.gz
7z x CRF++-0.58.tar
cd CRF++-0.58
chmod +x configure
./configure
make
"""
import pystruct
import pystruct.models
inference_method_options = ['lp', 'max-product']
inference_method = inference_method_options[1]
# graph = pystruct.models.GraphCRF(
# n_states=None,
# n_features=None,
# inference_method=inference_method,
# class_weight=None,
# directed=False,
# )
num_annots = 5
num_names = num_annots
aids = np.arange(5)
rng = np.random.RandomState(0)
hidden_nids = rng.randint(0, num_names, num_annots)
unique_nids, groupxs = ut.group_indices(hidden_nids)
# Indicator vector indicating the name
node_features = np.zeros((num_annots, num_names))
node_features[(aids, hidden_nids)] = 1
toy_params = {True: {'mu': 1.0, 'sigma': 2.2}, False: {'mu': 7.0, 'sigma': 0.9}}
if False:
import vtool as vt
import wbia.plottool as pt
pt.ensureqt()
xdata = np.linspace(0, 100, 1000)
tp_pdf = vt.gauss_func1d(xdata, **toy_params[True])
fp_pdf = vt.gauss_func1d(xdata, **toy_params[False])
pt.plot_probabilities([tp_pdf, fp_pdf], ['TP', 'TF'], xdata=xdata)
def metric(aidx1, aidx2, hidden_nids=hidden_nids, toy_params=toy_params):
if aidx1 == aidx2:
return 0
rng = np.random.RandomState(int(aidx1 + aidx2))
same = hidden_nids[int(aidx1)] == hidden_nids[int(aidx2)]
mu, sigma = ut.dict_take(toy_params[same], ['mu', 'sigma'])
return np.clip(rng.normal(mu, sigma), 0, np.inf)
pairwise_aidxs = list(ut.iprod(range(num_annots), range(num_annots)))
pairwise_labels = np.array( # NOQA
[hidden_nids[a1] == hidden_nids[a2] for a1, a2 in pairwise_aidxs]
)
pairwise_scores = np.array([metric(*zz) for zz in pairwise_aidxs])
pairwise_scores_mat = pairwise_scores.reshape(num_annots, num_annots) # NOQA
graph = pystruct.models.EdgeFeatureGraphCRF( # NOQA
n_states=num_annots,
n_features=num_names,
n_edge_features=1,
inference_method=inference_method,
)
import opengm
numVar = 10
unaries = np.ones([numVar, 3], dtype=opengm.value_type)
gm = opengm.gm(np.ones(numVar, dtype=opengm.label_type) * 3)
unary_fids = gm.addFunctions(unaries)
gm.addFactors(unary_fids, np.arange(numVar))
infParam = opengm.InfParam(workflow=ut.ensure_ascii('(IC)(TTC-I,CC-I)'))
inf = opengm.inference.Multicut(gm, parameter=infParam)
visitor = inf.verboseVisitor(printNth=1, multiline=False)
inf.infer(visitor)
arg = inf.arg()
# gridVariableIndices = opengm.secondOrderGridVis(img.shape[0], img.shape[1])
# fid = gm.addFunction(regularizer)
# gm.addFactors(fid, gridVariableIndices)
# regularizer = opengm.pottsFunction([3, 3], 0.0, beta)
# gridVariableIndices = opengm.secondOrderGridVis(img.shape[0], img.shape[1])
# fid = gm.addFunction(regularizer)
# gm.addFactors(fid, gridVariableIndices)
unaries = np.random.rand(10, 10, 2)
potts = opengm.PottsFunction([2, 2], 0.0, 0.4)
gm = opengm.grid2d2Order(unaries=unaries, regularizer=potts)
inf = opengm.inference.GraphCut(gm)
inf.infer()
arg = inf.arg() # NOQA
def learn_prob_score(num_scores=5, pad=55, ret_enc=False, use_cache=None):
r"""
Args:
num_scores (int): (default = 5)
Returns:
tuple: (discr_domain, discr_p_same)
CommandLine:
python -m wbia.unstable.demobayes --exec-learn_prob_score --show
Example:
>>> # ENABLE_DOCTEST
>>> from wbia.unstable.demobayes import * # NOQA
>>> num_scores = 2
>>> (discr_domain, discr_p_same, encoder) = learn_prob_score(num_scores, ret_enc=True, use_cache=False)
>>> print('discr_p_same = %r' % (discr_p_same,))
>>> ut.quit_if_noshow()
>>> import wbia.plottool as pt
>>> encoder.visualize()
>>> ut.show_if_requested()
"""
num_annots_train = 200
num_names_train = 5
toy_data = get_toy_data_1v1(num_annots_train, num_names_train)
# pairwise_aidxs, pairwise_scores, pairwise_matches = ut.dict_take(
# toy_data, 'pairwise_aidxs, pairwise_scores, pairwise_matches'.split(', '))
diag_scores, diag_labels = ut.dict_take(
toy_data, 'diag_scores, diag_labels'.split(', '))
# is_diag = [r < c for r, c, in pairwise_aidxs]
# diag_scores = pairwise_scores.compress(is_diag)
# diag_labels = pairwise_matches.compress(is_diag)
# Learn P(S_{ij} | M_{ij})
import vtool as vt
encoder = vt.ScoreNormalizer(
reverse=True,
monotonize=True,
adjust=4,
)
encoder.fit(X=diag_scores, y=diag_labels, verbose=True)
if False:
import wbia.plottool as pt
pt.ensureqt()
encoder.visualize()
# show_toy_distributions()
def discretize_probs(encoder):
p_tp_given_score = encoder.p_tp_given_score / encoder.p_tp_given_score.sum(
)
bins = len(p_tp_given_score) - (pad * 2)
stride = int(np.ceil(bins / num_scores))
idxs = np.arange(0, bins, stride) + pad
discr_p_same = p_tp_given_score.take(idxs)
discr_p_same = discr_p_same / discr_p_same.sum()
discr_domain = encoder.score_domain.take(idxs)
return discr_domain, discr_p_same
discr_domain, discr_p_same = discretize_probs(encoder)
if ret_enc:
return discr_domain, discr_p_same, encoder
return discr_domain, discr_p_same