def postprocess(inputs, outputs, is_training, apply_nms, nms_score_threshold,
nms_iou_threshold, nms_max_num_predicted_boxes,
use_furthest_voxel_sampling, num_furthest_voxel_samples,
sampler_score_vs_distance_coef):
"""Post-processor function."""
if not is_training:
# Squeeze voxel properties.
for key in standard_fields.get_output_voxel_fields():
if key in outputs and outputs[key] is not None:
outputs[key] = tf.squeeze(outputs[key], axis=0)
for key in standard_fields.get_output_point_fields():
if key in outputs and outputs[key] is not None:
outputs[key] = tf.squeeze(outputs[key], axis=0)
for key in standard_fields.get_output_object_fields():
if key in outputs and outputs[key] is not None:
outputs[key] = tf.squeeze(outputs[key], axis=0)
# Mask the valid voxels
mask_valid_voxels(inputs=inputs, outputs=outputs)
# NMS
postprocessor.postprocess(
outputs=outputs,
score_thresh=nms_score_threshold,
iou_thresh=nms_iou_threshold,
max_output_size=nms_max_num_predicted_boxes,
use_furthest_voxel_sampling=use_furthest_voxel_sampling,
num_furthest_voxel_samples=num_furthest_voxel_samples,
sampler_score_vs_distance_coef=sampler_score_vs_distance_coef,
apply_nms=apply_nms)
def test_postprocess(self):
num_classes = 10
n = 1000
outputs = {
standard_fields.DetectionResultFields.objects_score:
tf.random.uniform((n, num_classes),
minval=-2.0,
maxval=2.0,
dtype=tf.float32),
standard_fields.DetectionResultFields.objects_rotation_matrix:
tf.random.uniform((n, 3, 3),
minval=-1.0,
maxval=1.0,
dtype=tf.float32),
standard_fields.DetectionResultFields.objects_center:
tf.random.uniform((n, 3),
minval=10.0,
maxval=20.0,
dtype=tf.float32),
standard_fields.DetectionResultFields.objects_length:
tf.random.uniform((n, 1), minval=0.1, maxval=3.0,
dtype=tf.float32),
standard_fields.DetectionResultFields.objects_height:
tf.random.uniform((n, 1), minval=0.1, maxval=3.0,
dtype=tf.float32),
standard_fields.DetectionResultFields.objects_width:
tf.random.uniform((n, 1), minval=0.1, maxval=3.0,
dtype=tf.float32),
}
postprocessor.postprocess(outputs=outputs,
score_thresh=0.1,
iou_thresh=0.5,
max_output_size=10)
for key in [
standard_fields.DetectionResultFields.objects_length,
standard_fields.DetectionResultFields.objects_height,
standard_fields.DetectionResultFields.objects_width,
standard_fields.DetectionResultFields.objects_center,
standard_fields.DetectionResultFields.objects_class,
standard_fields.DetectionResultFields.objects_score
]:
self.assertEqual(len(outputs[key].shape), 2)
self.assertEqual(
len(outputs[standard_fields.DetectionResultFields.
objects_rotation_matrix].shape), 3)