Functions — github.com/Dengzhi-USTC/A-robust-registration-loss

hub / github.com/Dengzhi-USTC/A-robust-registration-loss / functions

Functions384 in github.com/Dengzhi-USTC/A-robust-registration-loss

⨍ Functions 384◇ Types & classes 63

↓ 29 callersMethodtransform

(self, tensor)

code/exps_deep_learning/fmr/se_math/transforms.py:188

↓ 16 callersMethod__init__

(self, args)

code/exps_deep_learning/dcp/model.py:463

↓ 16 callersMethodclone

(self)

code/exps_deep_learning/fmr/se_math/mesh.py:19

↓ 16 callersFunctiontransform_point_cloud

(point_cloud, rotation, translation)

code/utils.py:32

↓ 13 callersMethodsave

Save model checkpoint to file Args: model: Torch model optimizer: Torch optimizer step (int): Step, model

code/exps_deep_learning/rpm/common/torch.py:109

↓ 12 callersFunctionnpmat2euler

(mats, seq='zyx')

code/utils.py:70

↓ 8 callersMethod__len__

(self)

code/exps_deep_learning/pre_dataloader.py:183

↓ 7 callersFunctionindex_points

Array indexing, i.e. retrieves relevant points based on indices Args: points: input points data_loader, [B, N, C] idx: sample ind

code/exps_deep_learning/rpm/models/pointnet_util.py:51

↓ 7 callersFunctionsinc2

sinc2: t -> (1 - cos(t)) / (t**2)

code/LieAlgebra/sinc.py:91

↓ 7 callersMethodupdate

(self, loss, pred_src_transformed_final, data)

code/exps_deep_learning/rpm/Train_RPM.py:314

↓ 6 callersFunctionchamfer_dist

points_x : batchsize * M * 3 points_y : batchsize * N * 3 output : batchsize * M, batchsize * N

code/loss.py:236

↓ 6 callersMethodload

Loads saved model from file Args: save_path: Path to saved model (.pth). If a directory is provided instead, model-best.pth is us

code/exps_deep_learning/rpm/common/torch.py:131

↓ 6 callersFunctionmat

(x)

code/exps_deep_learning/fmr/se_math/so3.py:16

↓ 6 callersFunctionmat

(x)

code/LieAlgebra/so3.py:17

↓ 6 callersFunctionto_numpy

Wrapper around .detach().cpu().numpy()

code/exps_deep_learning/rpm/common/torch.py:24

↓ 6 callersMethodtransform_R_T

(self, points, R, T)

code/exps_deep_learning/pre_dataloader.py:41

↓ 5 callersFunctionclones

(module, N)

code/exps_deep_learning/dcp/model.py:23

↓ 5 callersFunctionsinc1

sinc1: t -> sin(t)/t

code/LieAlgebra/sinc.py:5

↓ 5 callersFunctionsinc2

sinc2: t -> (1 - cos(t)) / (t**2)

code/exps_deep_learning/fmr/se_math/sinc.py:96

↓ 4 callersFunctionRandom_uniform_distribution_lines_batch_efficient_resample

( r, centers, N, vertices1, vertices2, device='cpu')

code/loss.py:415

↓ 4 callersFunctionSample_neighs

(points, num_sample=5000, num_neigh=3, device='cpu')

code/loss.py:473

↓ 4 callersFunctionSelect_tensors

(data, idx_batch, idx_faces)

code/loss.py:31

↓ 4 callersMethodbackward

(ctx, grad_output)

code/exps_deep_learning/fmr/se_math/se3.py:147

↓ 4 callersFunctioncal_loss_intersection_batch_whole_median_pts_lines

(s_m, s_n,

code/loss.py:170

↓ 4 callersMethodeval

(self, loss_ignore_gt=False)

code/exps_deep_learning/rpm/Train_RPM.py:141

↓ 4 callersFunctiongenerate_datasets_human

(DCP=False, FMR=False)

code/exps_deep_learning/pre_dataloader.py:190

↓ 4 callersFunctionsinc1

sinc1: t -> sin(t)/t

code/exps_deep_learning/fmr/se_math/sinc.py:6

↓ 4 callersFunctionsinc3

sinc3: t -> (t - sin(t)) / (t**3)

code/LieAlgebra/sinc.py:120

↓ 4 callersFunctionsquare_distance

Calculate Euclid distance between each two points. src^T * dst = xn * xm + yn * ym + zn * zm； sum(src^2, dim=-1) = xn*xn + yn*yn + zn*

code/exps_deep_learning/rpm/models/pointnet_util.py:29

↓ 4 callersMethodtrain

(self, model, trainloader, optimizer, device,

code/exps_deep_learning/fmr/model.py:564

↓ 3 callersFunctionBatch_index_select

(data, idx)

code/utils.py:83

↓ 3 callersFunctionSample_points_normals

(points, normals, npoints)

code/utils.py:388

↓ 3 callersMethod__init__

(self, nch_input, nch_layers, b_shared=True,

code/exps_deep_learning/fmr/model.py:88

↓ 3 callersFunctionangle

Compute angle between 2 vectors For robustness, we use the same formulation as in PPFNet, i.e. angle(v1, v2) = atan2(cross(v1, v2), dot(v

code/exps_deep_learning/rpm/models/pointnet_util.py:173

↓ 3 callersFunctionangle_difference

Calculate angle between each pair of vectors. Assumes points are l2-normalized to unit length. Input: src: source points, [B, N, C]

code/exps_deep_learning/rpm/models/pointnet_util.py:11

↓ 3 callersFunctiondict_all_to_device

Sends everything into a certain device

code/exps_deep_learning/rpm/common/torch.py:17

↓ 3 callersMethodestimate_t

give two point clouds, estimate the T by using IC algorithm :param p0: point cloud :param p1: point cloud :param maxi

code/exps_deep_learning/fmr/model.py:186

↓ 3 callersFunctionfarthest_point_sample

Input: xyz: pointcloud data, [B, N, 3] npoint: number of samples Return: centroids: sampled pointcloud index, [B, npo

code/utils.py:275

↓ 3 callersFunctionindex_points

Input: points: input points data, [B, N, C] idx: sample index data, [B, S] Return: new_points:, indexed points data,

code/utils.py:233

↓ 3 callersFunctionquat2mat

(quat)

code/utils.py:52

↓ 3 callersFunctionsinc3

sinc3: t -> (t - sin(t)) / (t**3)

code/exps_deep_learning/fmr/se_math/sinc.py:126

↓ 2 callersMethodTransform

(self)

code/loss.py:455

↓ 2 callersMethod__init__

PointNet based Parameter prediction network Args: weights_dim: Number of weights to predict (excluding beta), should be something

code/exps_deep_learning/rpm/models/feature_nets.py:16

↓ 2 callersMethod_update_checkpoints_file

(self)

code/exps_deep_learning/rpm/common/torch.py:100

↓ 2 callersFunctionbatch_inverse

M(n) -> M(n); x -> x^-1

code/exps_deep_learning/fmr/se_math/invmat.py:6

↓ 2 callersFunctionbatch_inverse_dx

backward

code/exps_deep_learning/fmr/se_math/invmat.py:16

↓ 2 callersFunctioncal_intersection_batch2_points_with_line

(point_neis, line)

code/loss.py:68

↓ 2 callersFunctioncal_intersection_batch2_rand_lines

(points, line)

code/loss.py:319

↓ 2 callersMethodchamfer_loss

(self, a, b)

code/exps_deep_learning/fmr/model.py:442

↓ 2 callersMethodclose

(self)

code/exps_deep_learning/dcp/Train_DCP.py:43

↓ 2 callersMethodcompute_loss

(self, solver, data, device, mode='train', maxiter=5)

code/exps_deep_learning/fmr/model.py:504

↓ 2 callersFunctioncompute_sqrdis_map_2

points_x : batchsize * M * 3 points_y : batchsize * N * 3 output : batchsize * M * N

code/loss.py:38

↓ 2 callersMethodcreate_model

(self)

code/exps_deep_learning/fmr/model.py:754

↓ 2 callersFunctiondict_all_to_device

Sends everything into a certain device

code/utils.py:12

↓ 2 callersFunctiondict_all_to_device

Sends everything into a certain device

code/exps_deep_learning/fmr/model.py:27

↓ 2 callersFunctionexp

(x)

code/exps_deep_learning/fmr/se_math/se3.py:57

↓ 2 callersFunctionexp

(x)

code/exps_deep_learning/fmr/se_math/so3.py:61

↓ 2 callersFunctionexp

(x)

code/LieAlgebra/se3.py:57

↓ 2 callersFunctionexp

(x)

code/LieAlgebra/so3.py:62

↓ 2 callersFunctionfarthest_point_sample

Iterative farthest point sampling Args: xyz: pointcloud data_loader, [B, N, C] npoint: number of samples Returns: cen

code/exps_deep_learning/rpm/models/pointnet_util.py:71

↓ 2 callersFunctiongenerate_bbox

(vertices)

code/loss.py:325

↓ 2 callersFunctiongenerate_depth_mesh

(img, mask, idx=0)

code/utils.py:110

↓ 2 callersFunctiongenerate_mesh_by_bbox

(bbox, device='cpu')

code/loss.py:354

↓ 2 callersMethodplot

(self, fig=None, ax=None, *args, **kwargs)

code/exps_deep_learning/fmr/se_math/mesh.py:49

↓ 2 callersFunctionquery_ball_point

Grouping layer in PointNet++. Inputs: radius: local region radius nsample: max sample number in local region xyz: all po

code/exps_deep_learning/rpm/models/pointnet_util.py:96

↓ 2 callersMethodre_load

(self)

code/exps_deep_learning/rpm/Train_RPM.py:453

↓ 2 callersMethodrot_x

(self)

code/exps_deep_learning/fmr/se_math/mesh.py:88

↓ 2 callersFunctionrpmnet_arguments

Arguments used for both training and testing

code/exps_deep_learning/rpm/arguments.py:5

↓ 2 callersFunctionsave_checkpoint

(state, filename, suffix)

code/exps_deep_learning/fmr/Train_FMR.py:253

↓ 2 callersFunctionsave_pred_gt_obj

(V_src, V_pred, V_gt, V_tgt_trans, paths_src, paths_pred, paths_gt, paths_gt_pred)

code/exps_deep_learning/fmr/model.py:34

↓ 2 callersFunctiontest_one_epoch

(args, net, test_loader, save_results=None, epoch=0)

code/exps_deep_learning/dcp/Train_DCP.py:62

↓ 2 callersMethodtrain

(self)

code/exps_deep_learning/rpm/Train_RPM.py:45

↓ 1 callersFunctionBatch_index_select

code/exps_deep_learning/pre_dataloader.py:17

↓ 1 callersFunctionRandom_uniform_distribution_lines_batch_efficient

(r, centers,

code/loss.py:384

↓ 1 callersMethodSave_eval_results

(self, idx, pred_transforms, data)

code/exps_deep_learning/rpm/Train_RPM.py:85

↓ 1 callersFunctionWelsch1

(x, c)

code/loss.py:20

↓ 1 callersMethod__init__

(self, args: argparse.Namespace)

code/exps_deep_learning/rpm/models/rpmnet.py:161

↓ 1 callersMethod__init__

(self, args: argparse.Namespace)

code/exps_deep_learning/rpm/models/rpmnet copy.py:161

↓ 1 callersMethod__init__

(self, args: argparse.Namespace)

code/exps_deep_learning/rpm/models/rpmnet copy 2.py:161

↓ 1 callersFunction_fix_modelnet_broken_off

(filepath)

code/exps_deep_learning/fmr/se_math/mesh.py:148

↓ 1 callersFunction_load_off

read Geomview OFF file.

code/exps_deep_learning/fmr/se_math/mesh.py:118

↓ 1 callersFunction_mlp_layers

[B, Cin, N] -> [B, Cout, N] or [B, Cin] -> [B, Cout]

code/exps_deep_learning/fmr/model.py:57

↓ 1 callersMethod_remove_old_checkpoints

(self)

code/exps_deep_learning/rpm/common/torch.py:89

↓ 1 callersMethod_save_checkpoint

(self, step, model, optimizer, score)

code/exps_deep_learning/rpm/common/torch.py:68

↓ 1 callersFunctionadjust_learning_rate

Sets the learning rate to the initial LR decayed by 10 every 2 epochs

code/test_demo_optimized_Lie_Algebra.py:15

↓ 1 callersMethodapply_transform

(self, p0, x)

code/exps_deep_learning/fmr/se_math/transforms.py:177

↓ 1 callersMethodapprox_Jac

(self, p0, f0, dt)

code/exps_deep_learning/fmr/model.py:408

↓ 1 callersFunctionattention

(query, key, value, mask=None, dropout=None)

code/exps_deep_learning/dcp/model.py:27

↓ 1 callersMethodbackward

(ctx, grad_output)

code/exps_deep_learning/fmr/se_math/invmat.py:94

↓ 1 callersMethodbackward

(ctx, grad_output)

code/LieAlgebra/se3.py:170

↓ 1 callersFunctionbatch_pinv_dx

returns y = (x'*x)^-1 * x' and dy/dx.

code/exps_deep_learning/fmr/se_math/invmat.py:42

↓ 1 callersFunctionbtrace

(X)

code/exps_deep_learning/fmr/se_math/so3.py:83

↓ 1 callersFunctionbtrace

(X)

code/LieAlgebra/so3.py:84

↓ 1 callersMethodcal_gt_loss

(self, pred_transforms, data)

code/exps_deep_learning/rpm/Train_RPM.py:282

↓ 1 callersFunctioncal_intersection_batch2_step1

(points, line)

code/loss.py:265

↓ 1 callersFunctioncal_intersection_batch2_step2

(center2A_1, center2B_1, center2C_1, S)

code/loss.py:302

↓ 1 callersFunctioncal_loss

(data, rotation_ab_pred, translation_ab_pred, device)

code/exps_deep_learning/dcp/Train_DCP.py:233

↓ 1 callersMethodcal_loss

(self, pred_transforms, endpoints, data, r

code/exps_deep_learning/rpm/Train_RPM.py:184

↓ 1 callersFunctioncal_loss_intersection_batch_m_n_median_pts_lines

( m, n, points_informations_1, points_informations_2, label_intersection_sum_2_2, nface1, nfac

code/loss.py:115

next →1–100 of 384, ranked by callers