if exp.has_pattern(phns):
s = exp.get_s(path_idx)
E,edge_feature_row_breaks,\
edge_orientations= exp.get_edgemap_no_threshold(s)
phn_times = exp.get_phn_times(path_idx)
# get the backgrounds for all detection spots
P,C=exp.get_detection_scores(E,mean_template,
bg_len,
mean_background,
edge_feature_row_breaks,
edge_orientations)
scores = P+C
# select the object
pattern_times = exp.get_pattern_times(E,phns,phn_times,s)
# find the maxima
scores_maxima = ps.get_maxima(scores,5)
maxima_idx = np.arange(scores.shape[0])[scores_maxima]
# see if we pick up the patterns
for i in xrange(len(pattern_times)):
pattern_array =np.empty(scores.shape[0],dtype=bool)
pattern_array[:]=False
# consider something a detection if its within a third of the template length around the start of the pattern
pattern_array[pattern_times[i][0]-template_length/3:pattern_times[i][0]+template_length/3] = True
pattern_maxima = np.logical_and(scores_maxima,pattern_array)
if pattern_maxima.any():
true_pos_thresholds.append( np.max(scores[pattern_maxima]))
else:
# this was a false negative
pattern_thresholds.append(-np.inf)
# remove the maxima that are contained within the pattern radius
# at end of for loop only maxima left will be related to false positives
if t < bg_len:
cur_bgd = mean_background
else:
cur_bgd = np.maximum(np.minimum(np.mean(E[:,t-bg_len:t],
axis=1),
.4),
.05)
E_segment = E[:,t:t+template_length].copy()
_,edge_thresholds = esp.threshold_edgemap(E_segment,.30,edge_feature_row_breaks,abst_threshold=abst_threshold)
esp.spread_edgemap(E_segment,edge_feature_row_breaks,edge_orientations,spread_length=spread_length)
P,C = tt.score_template(template,cur_bgd,E_segment)
score_vals_edge_thresholds.append((P+C,edge_thresholds))
patterns = exp.get_patterns(E,phns,phn_times,s)
bgds = exp.get_pattern_bgds(E,phns,phn_times,s,bg_len)
# find the maxima
scores_maxima = ps.get_maxima(scores,maxima_radius)
maxima_idx = np.arange(scores.shape[0])[scores_maxima]
# see if we pick up the patterns
for i in xrange(len(pattern_times)):
esp.threshold_edgemap(patterns[i],.30,edge_feature_row_breaks,abst_threshold=abst_threshold)
esp.spread_edgemap(patterns[i],edge_feature_row_breaks,edge_orientations,spread_length=3)
esp.threshold_edgemap(bgds[i],.30,edge_feature_row_breaks,abst_threshold=abst_threshold)
esp.spread_edgemap(bgds[i],edge_feature_row_breaks,edge_orientations,spread_length=spread_length)
# compute background mean
bgds[i] = np.mean(bgds[i],axis=1)
# impose floor and ceiling constraints on values
bgds[i] = np.maximum(np.minimum(bgds[i],.4),.05)
registered_example = np.empty((template_height,template_length))
et._register_template(patterns[i],registered_example,template_height,template_length)
P_reg,C_reg = tt.score_template_background_section(mean_template,bgds[i],registered_example)
pattern_array =np.empty(scores.shape[0],dtype=bool)