MCPcopy Create free account

hub / github.com/Lilac-Lee/Neural_Scene_Flow_Prior / functions

Functions70 in github.com/Lilac-Lee/Neural_Scene_Flow_Prior

↓ 3 callersMethodavg
(self)
utils.py:82
↓ 2 callersFunction_handle_pointcloud_input
If points is an instance of Pointclouds, retrieve the padded points tensor along with the number of points per batch and the padded normals.
loss.py:27
↓ 2 callersMethodas_3d_bbox
r"""Calculate the 8 bounding box corners. Args: None Returns: Numpy array of shape (8,3) Co
preprocess_sf_nuscenes.py:334
↓ 2 callersFunctionflow_to_rgb
Creates a RGB representation of an optical flow (borrowed/modified from flowpy). Args: flow: scene flow. flow[..., 0] shoul
visualize.py:93
↓ 2 callersMethodget_avg
(self, key)
utils.py:63
↓ 2 callersFunctionget_object
(label)
preprocess_sf_nuscenes.py:231
↓ 2 callersFunctionmy_chamfer_fn
Chamfer distance between two pointclouds x and y. Args: x: FloatTensor of shape (N, P1, D) or a Pointclouds object representing
loss.py:65
↓ 2 callersMethodprint
(self, key=None)
utils.py:56
↓ 2 callersFunctionshow_flows
(pc1, pc2, flow, inverse=False)
visualize.py:21
↓ 2 callersMethodstep
(self, metrics)
utils.py:116
↓ 2 callersMethodtic
(self, key)
utils.py:50
↓ 2 callersMethodtoc
(self, key)
utils.py:53
↓ 1 callersMethod_init_is_better
(self, mode, min_delta, percentage)
utils.py:135
↓ 1 callersFunction_validate_chamfer_reduction_inputs
Check the requested reductions are valid. Args: batch_reduction: Reduction operation to apply for the loss across the batch,
loss.py:10
↓ 1 callersFunctionattach_coord
Attach the flow value to the coordinate tooltip. It allows you to see on the same figure, the RGB value of the pixel and the underlying value of
visualize.py:214
↓ 1 callersFunctioncalibration_pattern
Generates a calibration pattern to add as a legend to the scene flow plots. Args: pixel_size: Radius of the square test pattern.
visualize.py:154
↓ 1 callersFunctioncustom_draw_geometry_with_key_callback
(pcds)
visualize.py:262
↓ 1 callersFunctiondraw_alpha_rectangle
Alpha blend colored rectangle into image. Corner coords given as (x,y) tuples
preprocess_sf_nuscenes.py:537
↓ 1 callersFunctionget_lidar_sweeps
Get the lidar sweep from the given sweep_directory. ​ Args: sweep_directory: path to middle lidar sweep. sweep_index: index of
preprocess_sf_argoverse.py:210
↓ 1 callersFunctionlabel_is_closeby
Check if 3d cuboid pt (in egovehicle frame) is within range from egovehicle to prevent plot overcrowding.
preprocess_sf_nuscenes.py:530
↓ 1 callersFunctionmake_colorwheel
Creates a colorwheel (borrowed/modified from flowpy). A colorwheel defines the transitions between the six primary hues: Red(255, 0, 0), Yel
visualize.py:64
↓ 1 callersFunctionoptimize_neural_prior
(options, data_loader)
optimization.py:179
↓ 1 callersMethodreset
(self)
utils.py:85
↓ 1 callersFunctionscene_flow_metrics
(pred, labels)
utils.py:12
↓ 1 callersFunctionsolver
( pc1: torch.Tensor, pc2: torch.Tensor, flow: torch.Tensor, options: argparse.Namespace, n
optimization.py:28
↓ 1 callersMethodtic
(self)
utils.py:71
↓ 1 callersMethodtoc
(self, average=True)
utils.py:74
↓ 1 callersFunctionuv_coord_is_valid
Check if 2d-point lies within 3-channel color image boundaries
preprocess_sf_nuscenes.py:524
Method__getitem__
(self, index)
preprocess_sf_argoverse.py:55
Method__getitem__
(self, index)
preprocess_sf_nuscenes.py:71
Method__getitem__
(self, index)
data.py:76
Method__getitem__
(self, index)
data.py:136
Method__getitem__
(self, index)
data.py:182
Method__getitem__
(self, index)
data.py:282
Method__init__
( self, dataset_path='', partition='val', remove_ground=True,
preprocess_sf_argoverse.py:20
Method__init__
( self, save_fi_name, nusc, scene_tokens, channel = 'LIDAR_TOP',
preprocess_sf_nuscenes.py:30
Method__init__
Create an ObjectLabelRecord. Args: quaternion: Numpy vector representing quaternion, box/cuboid orientation transla
preprocess_sf_nuscenes.py:273
Method__init__
(self)
utils.py:47
Method__init__
(self)
utils.py:68
Method__init__
(self, gen)
utils.py:94
Method__init__
(self, mode='min', min_delta=0, patience=10, percentage=False)
utils.py:103
Method__init__
(self, options, partition='test')
data.py:10
Method__init__
(self, options, train=False)
data.py:121
Method__init__
(self, options, partition='val', width=1)
data.py:167
Method__init__
(self, options, partition="val", width=1)
data.py:253
Method__init__
(self, dim_x=3, filter_size=128, act_fn='relu', layer_size=8)
model.py:5
Method__iter__
(self)
utils.py:97
Method__len__
(self)
preprocess_sf_argoverse.py:198
Method__len__
(self)
preprocess_sf_nuscenes.py:227
Method__len__
(self)
data.py:73
Method__len__
(self)
data.py:133
Method__len__
(self)
data.py:179
Method__len__
(self)
data.py:347
Functionadd_config
(parser)
config.py:36
Methodas_2d_bbox
Construct a 2D bounding box from this label. Length is x, width is y, and z is height Alternatively could write code like::
preprocess_sf_nuscenes.py:309
Methodas_3d_bbox_ego
r"""Calculate the 8 bounding box corners. Args: None Returns: Numpy array of shape (8,3) Co
preprocess_sf_nuscenes.py:368
Functionattach_calibration_pattern
Attach a calibration pattern to axes. This function uses calibration_pattern to generate a figure. Args: calibration_pattern_kwargs
visualize.py:180
Functioncapture_depth
(vis)
visualize.py:274
Functioncapture_image
(vis)
visualize.py:280
Functionchange_background_to_black
(vis)
visualize.py:263
Methoddraw_rect
(selected_corners: np.ndarray, color: Tuple[int, int, int])
preprocess_sf_nuscenes.py:465
Methodforward
points -> features [B, N, 3] -> [B, K]
model.py:27
Functioninit_dirs
(options)
config.py:17
Functionmove_hue_on_S_axis
(hues, factors)
visualize.py:137
Functionmove_hue_on_V_axis
(hues, factors)
visualize.py:135
Functionnew_format_coord
(x, y)
visualize.py:238
Methodrender_clip_frustum_cv2
r"""We bring the 3D points into each camera, and do the clipping there. Renders box using OpenCV2. Roughly based on https://github
preprocess_sf_nuscenes.py:418
Functionset_deterministic_seeds
(options)
config.py:29
Functionstr2bool
(v)
config.py:8
Methodtotal
(self)
utils.py:79