Functions — github.com/Pixel-Talk/EHM-Tracker

hub / github.com/Pixel-Talk/EHM-Tracker / functions

Functions581 in github.com/Pixel-Talk/EHM-Tracker

⨍ Functions 581◇ Types & classes 83

↓ 131 callersMethodto

(self, device)

src/modules/pixie/pixie_encoder.py:134

↓ 65 callersMethodget

(self, key, default_vaule)

src/configs/base_config.py:21

↓ 62 callersMethodto

(self, device)

src/utils/graphics.py:231

↓ 59 callersMethodget

(self, img_bgr, **kwargs)

src/utils/face_analysis_diy.py:39

↓ 55 callersMethoditems

(self)

src/modules/flame/FLAME.py:378

↓ 40 callersMethodkeys

(self)

src/modules/flame/FLAME.py:375

↓ 40 callersMethodload

(self, key_name, type='image', **kwargs)

src/utils/lmdb.py:45

↓ 34 callersMethodexists

(self, key_name)

src/utils/lmdb.py:112

↓ 23 callersMethodwrite

(self, image)

src/utils/video.py:108

↓ 21 callersMethodupdate

Update the faces properties when vertex masks are changed

src/modules/flame/FLAME.py:411

↓ 15 callersMethodclose

(self, )

src/utils/lmdb.py:152

↓ 15 callersMethodkeys

(self, )

src/utils/lmdb.py:146

↓ 15 callersFunctionto_np

(array, dtype=np.float32)

src/modules/smplx/lbs.py:452

↓ 15 callersFunctionto_tensor

(array, dtype=torch.float32)

src/modules/smplx/lbs.py:448

↓ 14 callersFunctiondraw_landmarks

(landmarks, image, color=(0, 255, 0), radius=2, thickness=-1, viz_index=False)

src/utils/draw.py:21

↓ 14 callersFunctionload_config

(config_fp)

src/utils/io.py:124

↓ 13 callersMethoddump

(self, key_name, payload, type='image', encode_jpeg=True)

src/utils/lmdb.py:73

↓ 13 callersMethodrun

(self, *args)

src/modules/base/onnx_model.py:20

↓ 13 callersMethodtransform_points_screen

(self, points, with_xyflip = True, **kwargs)

src/utils/graphics.py:201

↓ 12 callersFunctionwrite_dict_pkl

(path, a_dict:dict)

src/utils/io.py:275

↓ 11 callersMethoddecompose_code

Convert a flattened parameter vector to a dictionary of parameters

src/modules/pixie/pixie.py:191

↓ 11 callersFunctionto_tensor

( array: Union[Array, Tensor], dtype=torch.float32 )

src/modules/mano/MANO.py:22

↓ 10 callersMethodencode

Encode images to smplx parameters Args: data: dict key: image_type (body/head/hand) value:

src/modules/pixie/pixie.py:240

↓ 10 callersFunctioninstantiate_from_config

(config)

src/utils/helper.py:148

↓ 10 callersFunctionload_dict_pkl

(path, encoding='')

src/utils/io.py:282

↓ 10 callersFunctionmask_from_points

Create a mask of supplied size from supplied points :param size: tuple of output mask size :param points: array of [x, y] points :returns

src/utils/draw.py:148

↓ 9 callersMethodbatch_orth_proj

X is N x num_verts x 3 camera is N x 3, each stands for s, x, y

src/modules/smplx/SMPLX.py:211

↓ 9 callersFunctioncrop_image_by_bbox

(img, bbox, lmk=None, dsize=512, scale=1.1, angle=None, flag_rot=False, **kwargs)

src/utils/crop.py:281

↓ 9 callersMethoddecompose_code

Convert a flattened parameter vector to a dictionary of parameters

src/modules/pixie/pixie_encoder.py:145

↓ 8 callersMethodwarmup

(self)

src/utils/landmark_runner.py:78

↓ 7 callersMethod__init__

(self, in_channels, out_channels, bilinear=True)

src/modules/pixie/models/resnet.py:227

↓ 7 callersFunction_transform_pts

conduct similarity or affine transformation to the pts pts: Nx2 ndarray M: 2x3 matrix or 3x3 matrix return: Nx2

src/utils/crop.py:31

↓ 7 callersFunctionexec_cmd

(cmd)

src/utils/video.py:15

↓ 7 callersFunctionimage2tensor

(image, norm=True)

src/utils/helper.py:248

↓ 7 callersFunctionto_np

(array, dtype=np.float32)

src/modules/flame/FLAME.py:41

↓ 7 callersFunctionto_np

(array, dtype=np.float32)

src/modules/mano/MANO.py:31

↓ 7 callersFunctionto_tensor

(array, dtype=torch.float32)

src/modules/flame/FLAME.py:36

↓ 7 callersFunctionvertices2landmarks

Calculates landmarks by barycentric interpolation Parameters ---------- vertices: torch.tensor BxVx3, dtype = torch.float32

src/modules/smplx/lbs.py:108

↓ 6 callersMethod__init__

(self, feature_channels=[256, 512, 1024, 2048], fpn_out=256)

src/modules/human_matting/stylematte.py:51

↓ 6 callersFunctioncrop_image

(img, pts: np.ndarray, **kwargs)

src/utils/crop.py:378

↓ 6 callersMethodeval

(self)

src/modules/pixie/pixie_encoder.py:140

↓ 6 callersFunctionload_onnx_model

(model_config, device=None)

src/utils/helper.py:179

↓ 6 callersFunctionperspective_projection

This function computes the perspective projection of a set of points. Input: points (bs, N, 3): 3D points rotation (bs, 3, 3)

src/modules/pixie/utils/util.py:38

↓ 6 callersMethodpose_abs2rel

change absolute pose to relative pose Basic knowledge for SMPLX kinematic tree: absolute pose = parent pose * relative pose

src/modules/smplx/SMPLX.py:278

↓ 5 callersFunctionlbs

Performs Linear Blend Skinning with the given shape and pose parameters Parameters ---------- betas : torch.tensor BxNB

src/modules/smplx/lbs.py:147

↓ 5 callersMethodtic

(self)

src/utils/timer.py:16

↓ 5 callersMethodtoc

(self, average=False)

src/utils/timer.py:21

↓ 5 callersFunctionvertices2joints

Calculates the 3D joint locations from the vertices Parameters ---------- J_regressor : torch.tensor JxV The regressor array tha

src/modules/smplx/lbs.py:239

↓ 5 callersFunctionvertices2landmarks

Calculates landmarks by barycentric interpolation Parameters ---------- vertices: torch.tensor BxVx3, dtype = torch.float32

src/modules/flame/lbs.py:103

↓ 4 callersMethod_make_layer

(self, block, planes, blocks, stride=1)

src/modules/pixie/models/resnet.py:47

↓ 4 callersMethodbuild_cameras_kwargs

(self, batch_size,focal_length)

src/modules/refiner/ehm_refiner.py:161

↓ 4 callersMethoddecode

Decode model parameters to smplx vertices & joints & texture Args: param_dict: smplx parameters param_type: should be

src/modules/pixie/pixie.py:491

↓ 4 callersFunctionparse_bbox_from_landmark_lite

(pts, scale=1.5, need_square=True)

src/utils/crop.py:255

↓ 4 callersMethodrandom_visualize

(self, vis_path, k=15, filter_key=None)

src/utils/lmdb.py:159

↓ 4 callersMethodrender_shape

-- rendering shape with detail normal map

src/modules/pixie/utils/renderer.py:371

↓ 4 callersMethodrun

(self, img_rgb: np.ndarray, lmk=None)

src/utils/landmark_runner.py:51

↓ 4 callersFunctionsmplx_joints_to_dwpose

(joints3d)

src/modules/refiner/smplx_utils.py:225

↓ 4 callersMethodtransform_hand_pts3d_to_image_coord

(self, X, M, img_size=512, is_left=False)

src/modules/refiner/ehm_refiner.py:114

↓ 4 callersMethodworld2uv

project vertices from world space to uv space vertices: [bz, V, 3] uv_vertices: [bz, 3, h, w]

src/modules/pixie/utils/renderer.py:518

↓ 3 callersMethod__init__

(self)

src/modules/teaser/smirk_encoder.py:15

↓ 3 callersFunction_draw_mp_kps

(image, landmarks, connections, landmark_drawing_spec, connection_drawing_spec)

src/utils/draw.py:32

↓ 3 callersFunction_elementary_basis_vector

(axis)

src/modules/pixie/utils/rotation_converter.py:283

↓ 3 callersFunction_elementary_basis_vector

(axis)

src/utils/rotation_converter.py:284

↓ 3 callersMethod_make_stage

(self, layer_config, num_inchannels, multi_scale_output=True, log=False)

src/modules/pixie/models/hrnet.py:392

↓ 3 callersMethod_make_subsample_layer

(self, in_channels=96, num_layers=3, stride=2)

src/modules/pixie/models/hrnet.py:374

↓ 3 callersMethod_make_transition_layer

( self, num_channels_pre_layer, num_channels_cur_layer)

src/modules/pixie/models/hrnet.py:300

↓ 3 callersFunctionbatch_rodrigues

Calculates the rotation matrices for a batch of rotation vectors Parameters ---------- rot_vecs: torch.tensor Nx3

src/modules/flame/lbs.py:310

↓ 3 callersFunctioncopy_parameter_from_resnet

(model, resnet_dict)

src/modules/pixie/models/resnet.py:156

↓ 3 callersFunctioncreate_backbone

(backbone_name, pretrained=True)

src/modules/teaser/smirk_encoder.py:7

↓ 3 callersMethodcvt_cam_to_persp_cam_fov

(self, wcam)

src/tracking_single_image.py:102

↓ 3 callersMethodcvt_cam_to_persp_cam_fov

(self, wcam)

src/data_prepare_pipeline.py:95

↓ 3 callersMethodexecute

(self, args: ArgumentConfig)

src/tracking_single_image.py:216

↓ 3 callersFunctionface_vertices

:param vertices: [batch size, number of vertices, 3] :param faces: [batch size, number of faces, 3] :return: [batch size, number of face

src/modules/renderer/util.py:6

↓ 3 callersFunctionfind_dynamic_lmk_idx_and_bcoords

Compute the faces, barycentric coordinates for the dynamic landmarks To do so, we first compute the rotation of the neck around the y-axis

src/modules/smplx/lbs.py:37

↓ 3 callersFunctioninference_detector

(session, oriImg)

src/modules/dwpose/onnxdet.py:97

↓ 3 callersFunctionload_obj

Ref: https://github.com/facebookresearch/pytorch3d/blob/25c065e9dafa90163e7cec873dbb324a637c68b7/pytorch3d/io/obj_io.py Load a mesh from a file-l

src/utils/io.py:222

↓ 3 callersFunctionplot_verts

Draw 68 key points Args: image: the input image kpt: (68, 3).

src/modules/pixie/utils/util.py:695

↓ 3 callersFunctionpose_abs2rel

change absolute pose to relative pose Basic knowledge for SMPLX kinematic tree: absolute pose = parent pose * relative pose Here,

src/modules/smplx/utils.py:41

↓ 3 callersMethodreselect_eyes

(self, vertices, lmk_type='lmks68')

src/modules/flame/FLAME.py:344

↓ 3 callersFunctionrot_mat_to_euler

(rot_mats)

src/modules/smplx/lbs.py:28

↓ 3 callersFunctionup_and_add

(x, y)

src/modules/human_matting/stylematte.py:46

↓ 2 callersMethod__init__

(self, image_size=256, device='cpu')

src/modules/renderer/base_renderer.py:217

↓ 2 callersMethod__init__

(self, flame_assets_dir, n_shape=300, n_exp=50, with_texture=False, add_teeth=False, with_lower_teeth=False)

src/modules/flame/FLAME.py:80

↓ 2 callersMethod__init__

(self, image_size, obj_filename, uv_size=256, rasterizer_type='standard')

src/modules/pixie/utils/renderer.py:162

↓ 2 callersMethod__init__

(self, channels = [2048, 1024, 1], last_op = None)

src/modules/pixie/models/encoders.py:31

↓ 2 callersMethod__init__

( self, focal_length=1.0, R: torch.Tensor = torch.eye(3)[None], T: torch.Tenso

src/utils/graphics.py:41

↓ 2 callersMethod_build_cameras

(self, transform_matrix, focal_length)

src/utils/graphics.py:344

↓ 2 callersFunction_get_3rd_point

To calculate the affine matrix, three pairs of points are required. This function is used to get the 3rd point, given 2D points a & b. The 3r

src/modules/dwpose/onnxpose.py:181

↓ 2 callersFunction_merge_with_weight

(p1, p2, w)

src/utils/draw.py:134

↓ 2 callersMethod_run

(self, img_rgb)

src/utils/mediapipe_runner.py:48

↓ 2 callersMethod_run

(self, inp)

src/utils/landmark_runner.py:47

↓ 2 callersMethod_run

(self, inp)

src/utils/fan_landmark_runner.py:146

↓ 2 callersFunction_transform_img

conduct similarity or affine transformation to the image, do not do border operation! img: M: 2x3 matrix or 3x3 matrix dsize: target shap

src/utils/crop.py:14

↓ 2 callersMethodadd_SHlight

sh_coeff: [bz, 9, 3]

src/modules/pixie/utils/renderer.py:317

↓ 2 callersMethodadd_directionlight

normals: [bz, nv, 3] lights: [bz, nlight, 6] returns: shading: [bz, nv, 3]

src/modules/pixie/utils/renderer.py:352

↓ 2 callersFunctionbatch_rodrigues

Calculates the rotation matrices for a batch of rotation vectors Parameters ---------- rot_vecs: torch.tensor Nx3

src/modules/smplx/lbs.py:283

↓ 2 callersFunctionblend_shapes

Calculates the per vertex displacement due to the blend shapes Parameters ---------- betas : torch.tensor Bx(num_betas) Blend s

src/modules/smplx/lbs.py:259

↓ 2 callersMethodbuild_cameras_kwargs

(self, batch_size)

src/modules/refiner/flame_refiner.py:30

↓ 2 callersFunctionbuild_minibatch

(all_frames, batch_size=1024, share_id=False)

src/utils/helper.py:327

↓ 2 callersFunctioncam2persp_cam_fov

scale by changing zdist,unchange fov Returns: w2c R, T: Rotation matrix and translation vector

src/modules/renderer/util.py:94

next →1–100 of 581, ranked by callers