scores_adapt_bg[d] = classifier.score(E_segment,bg)
coarse_scores[d] = classifier.coarse_score_count(E_segment)
bg = np.minimum(.1,
np.maximum(np.mean(E_segment,axis=1),
.4))
# now we get the indices sorted
indices = cl.remove_overlapping_examples(np.argsort(scores)[::-1],
classifier.window[1],
int(allowed_overlap*classifier.window[1]))
indices_adapt_bg = cl.remove_overlapping_examples(np.argsort(scores_adapt_bg)[::-1],
classifier.window[1],
int(allowed_overlap*classifier.window[1]))
indices_coarse = cl.remove_overlapping_examples(np.argsort(coarse_scores)[::-1],
classifier.window[1],
int(allowed_overlap*classifier.window[1]))
positives, negatives = cl.get_pos_neg_scores(indices,pattern_times,
scores,classifier.window[1])
positives_adapt_bg, negatives_adapt_bg = cl.get_pos_neg_scores(indices_adapt_bg,
pattern_times,
scores_adapt_bg,
classifier.window[1])
positives_coarse, negatives_coarse = cl.get_pos_neg_scores(indices_coarse,
pattern_times,
coarse_scores,
classifier.window[1])
all_positives.extend(positives)
all_negatives.extend(negatives)
all_positives_adapt_bg.extend(positives_adapt_bg)
all_negatives_adapt_bg.extend(negatives_adapt_bg)
all_positives_coarse.extend(positives_coarse)
all_negatives_coarse.extend(negatives_coarse)
else:
print "working on example", datum_id
if data_iter.next(
compute_pattern_times=True,
max_template_length=classifier.window[1]):
pattern_times = data_iter.pattern_times
num_detections = data_iter.E.shape[1] - classifier.window[1]
num_frames += data_iter.E.shape[1]
scores = -np.inf * np.ones(num_detections)
for d in xrange(num_detections):
E_segment = data_iter.E[:,d:d+classifier.window[1]].copy()
esp.threshold_edgemap(E_segment,.30,edge_feature_row_breaks,report_level=False,abst_threshold=abst_threshold)
esp.spread_edgemap(E_segment,edge_feature_row_breaks,edge_orientations,spread_length=3)
if classifier_flat.score_no_bg(E_segment[:,:classifier_flat.window[1]]) > flat_threshold:
scores[d] = classifier.score_no_bg(E_segment)
# now we get the indices sorted
indices = cl.remove_overlapping_examples(np.argsort(scores)[::-1],
classifier.window[1],
int(allowed_overlap*classifier.window[1]))
positives, negatives = cl.get_pos_neg_scores(indices,pattern_times,
scores,classifier.window[1])
all_positives.extend(positives)
all_negatives.extend(negatives)
else:
break
aar_def_roc,aar_def_roc_vals = cl.get_roc(np.sort(all_positives)[::-1],
np.sort(all_negatives)[::-1],num_frames)
np.save('aar_def_roc_vals052112',aar_def_roc_vals)
esp.threshold_edgemap(E_segment_mel,.30,e_breaks_mel,report_level=False,abst_threshold=abst_threshold)
esp.spread_edgemap(E_segment_mel,e_breaks_mel,e_orientations_mel,spread_length=3)
bg = np.minimum(.4,
np.maximum(np.mean(E_segment,axis=1),
.1))
scores_adapt_bg[d] = classifier.score(E_segment,bg)
scores_mel[d] = classifier.score_no_bg(E_segment_mel)
# now we get the indices sorted
indices_adapt_bg = cl.remove_overlapping_examples(np.argsort(scores_adapt_bg)[::-1],
classifier.window[1],
int(allowed_overlap*classifier.window[1]))
indices_mel = cl.remove_overlapping_examples(np.argsort(scores_mel)[::-1],
classifier.window[1],
int(allowed_overlap*classifier.window[1]))
positives_adapt_bg, negatives_adapt_bg = cl.get_pos_neg_scores(indices_adapt_bg,
pattern_times,
scores_adapt_bg,
classifier.window[1])
positives_mel, negatives_mel = cl.get_pos_neg_scores(indices_mel,
pattern_times,
scores_mel,
classifier.window[1])
all_positives_adapt_bg.extend(positives_adapt_bg)
all_negatives_adapt_bg.extend(negatives_adapt_bg)
all_positives_mel.extend(positives_mel)
all_negatives_mel.extend(negatives_mel)
else:
break
aar_roc_adapt_bg,aar_roc_vals_adapt_bg = cl.get_roc(np.sort(all_positives_adapt_bg)[::-1],
np.sort(all_negatives_adapt_bg)[::-1],num_frames)
aar_roc_coarse,aar_roc_vals_coarse = cl.get_roc(np.sort(all_positives_coarse)[::-1],
