Helper function for preparing groundtruth data for target assignment. In order to be consistent with the model.DetectionModel interface, grou

Returns the classification loss for a model.

Extracts features dict from input dict.

Returns numpy arrays belonging to class with index `class_index`. Args: detected_boxes: A numpy array containing detected boxes. dete

Extracts labels dict from input dict.

Loads labelmap from the labelmap path. Args: labelmap_path: Path to the labelmap. Returns: A dictionary mapping class name to class nume

Converts predictions from the second stage box classifier to detections. Args: refined_box_encodings: a 3-D float tensor with shape

Implementations must override this method. Args: image_features: A list of float tensors of shape [batch_size, height_i, width_i, c

Converts one or more images from RGB to Grayscale. Outputs a tensor of the same `DType` and rank as `images`. The size of the last dimension of

Maps image resize method from enumeration type to TensorFlow. Args: resize_method: The resize_method attribute of keep_aspect_ratio_resizer or

Updates dictionary with update content. Args: initial_dict: initial dictionary. update: updated dictionary.

Updates input configuration to reflect a new input path. The input_config object is updated in place, and hence not returned. Args: input_co

Adds detections for a single image to be used for evaluation. Args: image_id: A unique string/integer identifier for the image. detec

Evaluates VRD as being tp, fp from a single image. Args: detected_box_tuples: A numpy array of structures with shape [N,], repres

Evaluates detections as being tp, fp or weighted from a single image. The evaluation is done in two stages: 1. All detections are matched to

Creates an eval `input` function for `Estimator`. Args: eval_config: An eval_pb2.EvalConfig. eval_input_config: An input_reader_pb2.InputRe

Creates a TFRecord file from examples. Args: output_filename: Path to where output file is saved. num_shards: Number of shards for output f

Adds a bounding box to an image. Bounding box coordinates can be specified in either absolute (pixel) or normalized coordinates by setting the us

Adds a bounding box to an image. Bounding box coordinates can be specified in either absolute (pixel) or normalized coordinates by setting the us

Draws bounding boxes, masks, and keypoints on batch of image tensors. Args: images: A 4D uint8 image tensor of shape [N, H, W, C]. If C > 3, ad

Draws bounding boxes, masks, and keypoints on batch of image tensors. Args: images: A 4D uint8 image tensor of shape [N, H, W, C]. If C > 3, ad

Draws mask on an image. Args: image: uint8 numpy array with shape (img_height, img_height, 3) mask: a uint8 numpy array of shape (img_heigh

Draws mask on an image. Args: image: uint8 numpy array with shape (img_height, img_height, 3) mask: a uint8 numpy array of shape (img_heigh

Expands a row containing bounding boxes from CSV file. Args: csv_row: a single row of Open Images released groundtruth file. Returns:

Expands a row containing bounding boxes from CSV file. Args: csv_row: a single row of Open Images released groundtruth file. Returns:

Inserts multiple ones into a shape vector. Inserts an all-1 vector of length num_dims at position start_dim into a shape. Can be combined with tf

Filter to keep only boxes and masks with score exceeding a given threshold. This op keeps the collection of boxes and masks whose corresponding sco

Filter to keep only boxes with score exceeding a given threshold. This op keeps the collection of boxes whose corresponding scores are greater th

Filters out the variables matching the filter_regex. Filter out the variables whose name matches the any of the regular expressions in filter_reg

Returns a dictionary of eval metric ops to use with `tf.EstimatorSpec`. Note that once value_op is called, the detections and groundtruth added v

Determines the locations of the traffic light in the image Args: image: camera image Returns: list of boundi

Returns the optimizer type for training. Args: train_config: A train_pb2.TrainConfig. Returns: The type of the optimizer

Function to quantize weights and activation of the default graph.

Returns boolean column indicator where True means the colum is ignored. Returns: column_indicator: boolean vector which is True for all ign

Compute pairwise intersection areas between boxes. Args: boxes1: a numpy array with shape [N, 4] holding N boxes boxes2: a numpy array with

Compute pairwise intersection areas between masks. Args: masks1: a numpy array with shape [N, height, width] holding N masks. Masks value

Compute pairwise intersection areas between boxes. Args: boxlist1: BoxList holding N boxes boxlist2: BoxList holding M boxes scope: nam

Implement the Kalman Filter, including the predict and the update stages, with the measurement z

Loads label map proto. Args: path: path to StringIntLabelMap proto text file. Returns: a StringIntLabelMapProto

Constructs the object detection model. Args: features: Dictionary of feature tensors, returned from `input_fn`. labels: Dictionary of

Returns the number of anchors per spatial location. Returns: a list of integers, one for each expected feature map to be passed to th

Returns number (int32 scalar tensor) of matched columns.

Returns number (int32 scalar tensor) of matched columns.

Returns number (int32 scalar tensor) of unmatched columns.

Pipeline function for detection and tracking

Implment only the predict stage. This is used for unmatched detections and unmatched tracks

Randomly adjusts brightness. Makes sure the output image is still between 0 and 255. Args: image: rank 3 float32 tensor contains 1 image ->

Randomly adjusts contrast. Makes sure the output image is still between 0 and 255. Args: image: rank 3 float32 tensor contains 1 image -> [h

Randomly adjusts hue. Makes sure the output image is still between 0 and 255. Args: image: rank 3 float32 tensor contains 1 image -> [height

Randomly adjusts saturation. Makes sure the output image is still between 0 and 255. Args: image: rank 3 float32 tensor contains 1 image ->

Randomly crops the image. Given the input image and its bounding boxes, this op randomly crops a subimage. Given a user-provided set of input co

Convenience function for setting box coordinates. Args: boxes: a tensor of shape [N, 4] representing box corners Raises: ValueEr

Sort boxes and associated fields according to a scalar field. A common use case is reordering the boxes according to descending scores. Args:

Random crop preprocessing with default parameters as in SSD paper. Liu et al., SSD: Single shot multibox detector. For further information on ran

Unstacks all tensors in `tensor_dict` along 0th dimension. Unstacks tensor from the tensor dict along 0th dimension and returns a tensor_dict con

Export detections of a single image to COCO format. This function converts detections represented as numpy arrays to dictionaries that can be ing

Export detection masks of a single image to COCO format. This function converts detections represented as numpy arrays to dictionaries that can b

Export groundtruth of a single image to COCO format. This function converts groundtruth detection annotations represented as numpy arrays to dict

Returns true if COCO Eval is configured to evaluate in agnostic mode.

Fetches dictionary holding category information given category id. Args: category_id: integer id Returns: dictionary holding 'id'

Returns list of valid category ids.

Initializes the LookupTensor handler. Simply calls a vocabulary (most often, a label mapping) lookup. Args: tensor_key: the name of th

COCOWrapper constructor. See http://mscoco.org/dataset/#format for a description of the format. By default, the coco.COCO class constructor r

Constructor. Args: categories: A list of dicts, each of which has the following keys - 'id': (required) an integer id uniquely iden

Adds output nodes for detection boxes and scores. Adds the following nodes for output tensors - * num_detections: float32 tensor of shape [batc