if 'validate_object_labels' in exec_list:
print("validating object labels")
valid_label_flags = tools.validate_object_labels(
frame.laser_labels, lidar_pcl, configs_det,
0 if configs_det.use_labels_as_objects == True else 10)
else:
print('loading object labels and validation from result file')
valid_label_flags = load_object_from_file(results_fullpath,
data_filename,
'valid_labels',
cnt_frame)
## Performance evaluation for object detection
if 'measure_detection_performance' in exec_list:
print('measuring detection performance')
det_performance = eval.measure_detection_performance(
detections, frame.laser_labels, valid_label_flags)
else:
print('loading detection performance measures from file')
det_performance = load_object_from_file(
results_fullpath, data_filename, 'det_performance_' +
configs_det.arch + '_' + str(configs_det.conf_thresh),
cnt_frame)
det_performance_all.append(
det_performance
) # store all evaluation results in a list for performance assessment at the end
## Visualization for object detection
if 'show_range_image' in exec_list:
img_range = pcl.show_range_image(frame) #lider_name S3
img_range = img_range.astype(np.uint8)
print("validating object labels")
valid_label_flags = tools.validate_object_labels(
frame.laser_labels, lidar_pcl, configs_det,
0 if configs_det.use_labels_as_objects == True else 10)
else:
print('loading object labels and validation from result file')
valid_label_flags = load_object_from_file(results_fullpath,
data_filename,
'valid_labels',
cnt_frame)
## Performance evaluation for object detection
if 'measure_detection_performance' in exec_list:
print('measuring detection performance')
det_performance = eval.measure_detection_performance(
detections, frame.laser_labels, valid_label_flags,
configs_det.min_iou)
else:
print('loading detection performance measures from file')
# det_performance = load_object_from_file(results_fullpath, data_filename, 'det_performance_' + configs_det.arch + '_' + str(configs_det.conf_thresh), cnt_frame)
det_performance = load_object_from_file(results_fullpath,
data_filename,
'det_performance',
cnt_frame)
det_performance_all.append(
det_performance
) # store all evaluation results in a list for performance assessment at the end
## Visualization for object detection
if 'show_range_image' in exec_list:
img_range = pcl.show_range_image(frame, lidar_name)