def _BuildMetric(self, feed_data, classid):
"""Construct tensors and the feed_dict for KITTI metric op.
Args:
feed_data: a NestedMap returned by _GetData()
classid: integer. Unused in this implementation.
Returns:
scalar_metrics: a dict mapping all the metric names to fetch tensors.
curves: a dict mapping all the curve names to fetch tensors.
feed_dict: a dict mapping all the tensors in feed_tensors to feed values.
"""
if feed_data is None:
dummy_scalar = tf.constant(np.nan)
dummy_curve = tf.zeros([self.metadata.NumberOfPrecisionRecallPoints(), 2],
tf.float32)
scalar_metrics = {'ap': dummy_scalar}
curve_metrics = {'pr': dummy_curve}
return scalar_metrics, curve_metrics, {}
feed_dict = {}
f_iou = tf.placeholder(tf.float32)
feed_dict[f_iou] = feed_data.iou_threshold
f_gt_bbox = tf.placeholder(tf.float32)
feed_dict[f_gt_bbox] = feed_data.gt.bbox
f_gt_imgid = tf.placeholder(tf.int32)
feed_dict[f_gt_imgid] = feed_data.gt.imgid
f_gt_ignore = tf.placeholder(tf.int32)
feed_dict[f_gt_ignore] = feed_data.gt.ignore
f_pd_bbox = tf.placeholder(tf.float32)
feed_dict[f_pd_bbox] = feed_data.pd.bbox
f_pd_imgid = tf.placeholder(tf.int32)
feed_dict[f_pd_imgid] = feed_data.pd.imgid
f_pd_ignore = tf.placeholder(tf.int32)
feed_dict[f_pd_ignore] = feed_data.pd.ignore
f_pd_score = tf.placeholder(tf.float32)
feed_dict[f_pd_score] = feed_data.pd.score
ap, pr = ops.average_precision3d(
iou_threshold=f_iou,
groundtruth_bbox=f_gt_bbox,
groundtruth_imageid=f_gt_imgid,
groundtruth_ignore=f_gt_ignore,
prediction_bbox=f_pd_bbox,
prediction_imageid=f_pd_imgid,
prediction_ignore=f_pd_ignore,
prediction_score=f_pd_score,
num_recall_points=self.metadata.NumberOfPrecisionRecallPoints())
scalar_metrics = {'ap': ap}
curve_metrics = {'pr': pr}
return scalar_metrics, curve_metrics, feed_dict
def _GetAP(self, gt_bbox, gt_imgid, pd_bbox, pd_imgid, pd_score):
g = tf.Graph()
with g.as_default():
iou, pr = ops.average_precision3d(
iou_threshold=0.5,
groundtruth_bbox=gt_bbox,
groundtruth_imageid=gt_imgid,
groundtruth_ignore=tf.zeros_like(gt_imgid, dtype=tf.int32),
prediction_bbox=pd_bbox,
prediction_imageid=pd_imgid,
prediction_score=pd_score,
prediction_ignore=tf.zeros_like(pd_imgid, dtype=tf.int32),
num_recall_points=41,
algorithm='KITTI')
with self.session(graph=g) as sess:
val = sess.run([iou, pr])
return val
def _BuildMetric(self, feed_data, classid):
"""Construct tensors and the feed_dict for KITTI metric op.
Args:
feed_data: a NestedMap returned by _GetData()
classid: integer. Unused in this implementation.
Returns:
A tuple of 3 dicts:
- scalar_metrics: a dict mapping all the metric names to fetch tensors.
- curves: a dict mapping all the curve names to fetch tensors.
- feed_dict: a dict mapping the tensors in feed_tensors to feed values.
"""
if feed_data is None:
dummy_scalar = tf.constant(np.nan)
dummy_calibration = tf.constant(np.nan)
dummy_curve = tf.zeros(
[self.metadata.NumberOfPrecisionRecallPoints(), 2], tf.float32)
scalar_metrics = {'ap': dummy_scalar}
curve_metrics = {'pr': dummy_curve}
calibration_metrics = {'calibrations': dummy_calibration}
return py_utils.NestedMap(feed_dict={},
scalar_metrics=scalar_metrics,
curve_metrics=curve_metrics,
calibration_metrics=calibration_metrics)
feed_dict = {}
f_iou = tf.placeholder(tf.float32)
feed_dict[f_iou] = feed_data.iou_threshold
f_gt_bbox = tf.placeholder(tf.float32)
feed_dict[f_gt_bbox] = feed_data.gt.bbox
f_gt_imgid = tf.placeholder(tf.int32)
feed_dict[f_gt_imgid] = feed_data.gt.imgid
f_gt_ignore = tf.placeholder(tf.int32)
feed_dict[f_gt_ignore] = feed_data.gt.ignore
f_pd_bbox = tf.placeholder(tf.float32)
feed_dict[f_pd_bbox] = feed_data.pd.bbox
f_pd_imgid = tf.placeholder(tf.int32)
feed_dict[f_pd_imgid] = feed_data.pd.imgid
f_pd_ignore = tf.placeholder(tf.int32)
feed_dict[f_pd_ignore] = feed_data.pd.ignore
f_pd_score = tf.placeholder(tf.float32)
feed_dict[f_pd_score] = feed_data.pd.score
# TODO(shlens): The third returned argument contain statistics for measuring
# the calibration error. Use it.
ap, pr, calibration = ops.average_precision3d(
iou_threshold=f_iou,
groundtruth_bbox=f_gt_bbox,
groundtruth_imageid=f_gt_imgid,
groundtruth_ignore=f_gt_ignore,
prediction_bbox=f_pd_bbox,
prediction_imageid=f_pd_imgid,
prediction_ignore=f_pd_ignore,
prediction_score=f_pd_score,
num_recall_points=self.metadata.NumberOfPrecisionRecallPoints())
scalar_metrics = {'ap': ap}
curve_metrics = {'pr': pr}
calibration_metrics = {'calibrations': calibration}
return py_utils.NestedMap(feed_dict=feed_dict,
scalar_metrics=scalar_metrics,
curve_metrics=curve_metrics,
calibration_metrics=calibration_metrics)