Functions — github.com/CyrilSterling/EVFlowNet-pytorch

CodeHub MCPcopy Create free account

hub / github.com/CyrilSterling/EVFlowNet-pytorch / functions

Functions33 in github.com/CyrilSterling/EVFlowNet-pytorch

⨍ Functions 33 ◇ Types & classes 5

↓ 8 callersFunctiondrawImageTitle

↓ 7 callersFunctiongeneral_conv2d

a general convolution layer which includes a conv2d, a relu and a batch_normalize

src/basic_layers.py:61

↓ 5 callersFunctioncharbonnier_loss

Robust Charbonnier loss, as defined in equation (4) of the paper.

src/losses.py:33

↓ 4 callersFunctionconfigs

src/config.py:3

↓ 3 callersFunctionflow_viz_np

(flow_x, flow_y)

src/vis_utils.py:52

↓ 3 callersFunctionprop_flow

(x_flow, y_flow, x_indices, y_indices, x_mask, y_mask, scale_factor=1.0)

src/eval_utils.py:50

↓ 2 callersFunction_save_events

(args, events, image_times, event_count_images,

data/extract_rosbag_to_npy.py:20

↓ 2 callersFunctionfilter_events

r'''Removes all events with timestamp lower than the specified one Args: events (list): the list of events in form of (x, y, t, p)

data/extract_rosbag_to_npy.py:90

↓ 2 callersFunctionwarp_images_with_flow

Generates a prediction of an image given the optical flow, as in Spatial Transformer Networks.

src/losses.py:6

↓ 1 callersMethod__init__

(self, channels, layers=2, do_batch_norm=False)

src/basic_layers.py:8

↓ 1 callersMethod_read_events

(self, event_count_images, event_time_images, n

src/data_loader.py:140

↓ 1 callersFunctioncompute_photometric_loss

Multi-scale photometric loss, as defined in equation (3) of the paper.

src/losses.py:65

↓ 1 callersFunctioncompute_smoothness_loss

Local smoothness loss, as defined in equation (5) of the paper. The neighborhood here is defined as the 8-connected region around each pixel.

src/losses.py:41

↓ 1 callersFunctionestimate_corresponding_gt_flow

(x_flow_in, y_flow_in, gt_timestamps,

src/eval_utils.py:94

↓ 1 callersFunctionflow_error_dense

(flow_gt, flow_pred, event_img, is_car=False)

src/eval_utils.py:12

↓ 1 callersFunctionmain

data/extract_rosbag_to_npy.py:105

↓ 1 callersFunctionmain

src/train.py:14

↓ 1 callersFunctionmain

src/test.py:190

↓ 1 callersMethodread_file_paths

return: event_data_paths,paths of event data (left and right in one folder is two) n_ima: the sum number of event pictures in every p

src/data_loader.py:170

↓ 1 callersFunctiontest

(args, EVFlowNet_model, EventDataLoder)

Method__getitem__

src/data_loader.py:30

src/EVFlowNet.py:8

(self, smoothness_weight, weight_decay_weight=1e-4)

src/losses.py:94

(self, in_channels, out_channels, ksize=3, do_batch_norm=False)

src/basic_layers.py:27

(self, data_folder_path, split, count_only=False, time_only=False, skip_frames=False)

src/data_loader.py:21

src/data_loader.py:137

Function_bytes_feature

data/extract_rosbag_to_npy.py:17

Function_int64_feature

data/extract_rosbag_to_npy.py:14

Functiondraw_color_wheel_np

(width, height)

src/vis_utils.py:24

src/EVFlowNet.py:31

(self, flow_dict, prev_image, next_image, EVFlowNet_model)

src/losses.py:99

(self,input_res)

src/basic_layers.py:18

src/basic_layers.py:51