Python nets.inception_resnet_v2.inception_resnet_v2_base() Examples

The following are 27 code examples of nets.inception_resnet_v2.inception_resnet_v2_base(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module nets.inception_resnet_v2 , or try the search function .
Example #1
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From ros_tensorflow with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #2
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From multilabel-image-classification-tensorflow with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #3
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From mtl-ssl with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay, trainable=self._is_training)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #4
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From models with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #5
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From g-tensorflow-models with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #6
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From MAX-Object-Detector with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #7
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #8
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #9
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From Elphas with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #10
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From MBMD with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #11
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #12
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From hands-detection with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #13
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From moveo_ros with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #14
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From BMW-TensorFlow-Training-GUI with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #15
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From DOTA_models with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #16
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From Gun-Detector with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #17
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From tensorflow with BSD 2-Clause "Simplified" License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #18
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From Hands-On-Machine-Learning-with-OpenCV-4 with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #19
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From Traffic-Rule-Violation-Detection-System with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #20
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From yolo_v2 with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #21
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From HereIsWally with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #22
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From garbage-object-detection-tensorflow with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #23
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From cartoonify with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map 
Example #24
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From Person-Detection-and-Tracking with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #25
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From ros_people_object_detection_tensorflow with Apache License 2.0 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #26
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From vehicle_counting_tensorflow with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm],
                          is_training=self._train_batch_norm):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          return inception_resnet_v2.inception_resnet_v2_base(
              preprocessed_inputs, final_endpoint='PreAuxLogits',
              scope=scope, output_stride=self._first_stage_features_stride,
              align_feature_maps=True) 
Example #27
Source File: faster_rcnn_inception_resnet_v2_feature_extractor.py    From object_detector_app with MIT License 5 votes vote down vote up
def _extract_proposal_features(self, preprocessed_inputs, scope):
    """Extracts first stage RPN features.

    Extracts features using the first half of the Inception Resnet v2 network.
    We construct the network in `align_feature_maps=True` mode, which means
    that all VALID paddings in the network are changed to SAME padding so that
    the feature maps are aligned.

    Args:
      preprocessed_inputs: A [batch, height, width, channels] float32 tensor
        representing a batch of images.
      scope: A scope name.

    Returns:
      rpn_feature_map: A tensor with shape [batch, height, width, depth]
    Raises:
      InvalidArgumentError: If the spatial size of `preprocessed_inputs`
        (height or width) is less than 33.
      ValueError: If the created network is missing the required activation.
    """
    if len(preprocessed_inputs.get_shape().as_list()) != 4:
      raise ValueError('`preprocessed_inputs` must be 4 dimensional, got a '
                       'tensor of shape %s' % preprocessed_inputs.get_shape())

    with slim.arg_scope(inception_resnet_v2.inception_resnet_v2_arg_scope(
        weight_decay=self._weight_decay)):
      # Forces is_training to False to disable batch norm update.
      with slim.arg_scope([slim.batch_norm], is_training=False):
        with tf.variable_scope('InceptionResnetV2',
                               reuse=self._reuse_weights) as scope:
          rpn_feature_map, _ = (
              inception_resnet_v2.inception_resnet_v2_base(
                  preprocessed_inputs, final_endpoint='PreAuxLogits',
                  scope=scope, output_stride=self._first_stage_features_stride,
                  align_feature_maps=True))
    return rpn_feature_map