Darknet Convolution2D followed by BatchNormalization and LeakyReLU.

Get image identifiers for corresponding list of dataset identifies. Parameters ---------- voc_path : str Path to VOCdevkit direct

Compose arbitrarily many functions, evaluated left to right. Reference: https://mathieularose.com/function-composition-in-python/

Convert final layer features to bounding box parameters. Parameters ---------- feats : tensor Final convolutional layer features.

Process all given ids and adds them to given datasets.

Wrapper to set Darknet weight regularizer for Convolution2D.

Bottleneck block of 3x3, 1x1, 3x3 convolutions.

Get path to image annotation for given year and image id.

Get path to image for given year and image id.

Create YOLO_V2 model CNN body in Keras.

Evaluate YOLO model on given input batch and return filtered boxes.

Bottleneck block of 3x3, 1x1, 3x3, 1x1, 3x3 convolutions.

returns the body of the model and the model # Params: load_pretrained: whether or not to load the pretrained model or initialize all we

Generate first 18 conv layers of Darknet-19.

Draw bounding boxes on image. Draw bounding boxes with class name and optional box score on image. Args: image: An `array` of shape

Find detector in YOLO where ground truth box should appear. Parameters ---------- true_boxes : array List of ground truth boxes i

Process selected Pascal VOC dataset to generate TFRecords files. Parameters ---------- name : string Name of resulting dataset 't

Locate files for train and test sets and then generate TFRecords.

Convert Pascal VOC ground truth to TFExample protobuf. Parameters ---------- image_data : bytes Encoded image bytes. boxes :

Draw bounding boxes on image data

loads the anchors from a file

Get object bounding boxes annotations for given image. Parameters ---------- voc_path : str Path to VOCdevkit directory. year

loads the classes

Return list of random colors for number of classes given.

Precompute detectors_mask and matching_true_boxes for training. Detectors mask is 1 for each spatial position in the final conv layer and

Get image data as uint8 array for given image. Parameters ---------- voc_path : str Path to VOCdevkit directory. year : str

Process Pascal VOC annotations.

processes the data

Decode image at given path.

retrain/fine-tune the model logs training with tensorboard saves training weights in current directory best weights according to v

Convert all config sections to have unique names. Adds unique suffixes to config sections for compability with configparser.

Convert YOLO box predictions to bounding box corners.

Filter YOLO boxes based on object and class confidence.

Generate Darknet-19 model for Imagenet classification.

Thin wrapper for Tensorflow space_to_depth with block_size=2.

Determine space_to_depth output shape for block_size=2. Note: For Lambda with TensorFlow backend, output shape may not be needed.

Generate a complete YOLO_v2 localization model.

YOLO localization loss function. Parameters ---------- yolo_output : tensor Final convolutional layer features. true_boxes :