:param inputdata: :param is_training: :param name: :param scale: :return:

stack conv and activation in vgg16 :param input_tensor: :param k_size: :param out_dims: :param name:

:param name: :param inputdata: :return:

Packing the tensorflow conv2d function. :param name: op name :param inputdata: A 4D tensorflow tensor which ust have known nu

:param input_tensor: :param previous_feats_tensor: :param out_channels_nums: :param kernel_size: :param prev

* Constructor. Using config file to setup lanenet model. Mainly defined object are as follows: * 1.Init mnn model file path * 2.Init lanenet model p

:param input_tensor: :param kernel_size: :param name: :param depth_multiplier: :param padding: :para

:param input_tensor: :param name: :param reuse :return:

conv block in attention refine :param input_tensor: :param k_size: :param output_channels: :param stride:

conv block in attention refine :param input_tensor: :param k_size: :param output_channels: :param stride:

:param name: :param inputdata: :param kernel_size: :param stride: :param padding: :param data_format

:param inputdata: :param axis: :param name: :return:

conv block in attention refine :param input_tensor: :param k_size: :param output_channels: :param stride:

dataset feed pipline input :param batch_size: :return: A tuple (images, labels), where: * images is a flo

conv block in attention refine :param input_tensor: :param k_size: :param output_channels: :param stride:

:param value: :return:

Performs central crops of the given image :param image: :param crop_height: :param crop_width: :return:

calculate lanenet loss for training :param input_tensor: :param binary_label: :param instance_label: :param n

:param key: :param create_if_not_exist: :return:

:param key: :param value: :param create_if_not_exist: :return:

:param input_tensor: :param name: :param reuse: :return:

:param f_obj: :return:

---- hyperrectangle helpers ---- */

:param binary_seg_result: :param instance_seg_result: :param min_area_threshold: :param source_image: :param

inserts the item. if dist_sq is >= 0, then do an ordered insert */ TODO make the ordering code use heapsort */

:param name: :param inputdata: :return:

:param input_tensor: :param e: :param name :return:

:param input_tensor: :param output_channels: :param e: :param name :return:

Calculate the average gradient for each shared variable across all towers. Note that this function provides a synchronization point across all

:return:

:return:

:param onehot_labels: :param logits: :param classes_weights: :return:

Calculate gradients for single GPU :param images: images for training :param binary_labels: binary labels corresponding to im

:param warmup_steps: :param name: :return:

:param warmup_steps: :param name: :return:

connect components analysis to remove the small components :param image: :return:

conv block in attention refine :param input_tensor: :param k_size: :param output_channels: :param stride:

conv block in attention refine :param input_tensor: :param k_size: :param output_channels: :param stride:

dbscan cluster :param embedding_image_feats: :return:

Generate training example index file, split source file into 0.85, 0.1, 0.05 for training, testing and validation. Each image folder

get lane embedding features according the binary seg result :param binary_seg_ret: :param instance_seg_ret: :return:

if the net is used for training or not :return:

if the net is used for training or not :return:

if the net is used for training or not :return:

if the net is used for training or not :return:

if the net is used for training or not :return:

if the net is used for training or not :return:

if the net is used for training or not :return:

if the net is used for training or not :return:

Check if source data complete :return:

Check if the training sample index file is complete :return:

:param config_file_path :return:

:return:

morphological process to fill the hole in the binary segmentation result :param image: :param kernel_size: :return:

:param onehot_labels: :param logits: :param classes_weights: :param gamma: :return:

:return:

:param input_tensor: :param name: :return:

:param binary_seg_result: :param instance_seg_result: :return:

:param name: :param inputdata: :param kernel_size: :param stride: :param padding: :param data_format

:param detail_output: :param semantic_output: :param name: :return:

:param input_tensor: :param name: :return:

:param input_tensor: :param name: :return:

:param input_tensor: :param name: :return:

:param input_tensor: :param name: :param reuse: :return:

:param input_tensor: :param name: :param reuse: :return:

:param input_tensor: :param name: :param prepare_data_for_booster: :return:

:param ckpt_file_path: :param pb_file_path: :return:

Packing the tensorflow conv2d function. :param name: op name :param inputdata: A 4D tensorflow tensor which ust have known nu

discriminative loss :param prediction: inference of network :param correct_label: instance label :param feature_dim: feature dimensio

:param name: :param inputdata: :param keep_prob: :param noise_shape: :return:

:param src_dir: :param weights_path: :param save_dir: :return:

generate sample index file :param src_dir: :param b_gt_image_dir: :param i_gt_image_dir: :param image_dir: :return:

:return:

Generate tensorflow records file :return:

:return:

:return:

:return:

:return: