def __init__(self, transform, data_split):
self.transform = transform
self.data_split = data_split
self.data_path = osp.join('..', 'data', 'PW3D', 'data')
self.human_bbox_root_dir = osp.join('..', 'data', 'PW3D',
'rootnet_output',
'bbox_root_pw3d_output.json')
# SMPL joint set
self.smpl = SMPL()
self.face = self.smpl.face
self.joint_regressor = self.smpl.joint_regressor
self.vertex_num = self.smpl.vertex_num
self.joint_num = self.smpl.joint_num
self.joints_name = self.smpl.joints_name
self.skeleton = self.smpl.skeleton
self.root_joint_idx = self.smpl.root_joint_idx
self.face_kps_vertex = self.smpl.face_kps_vertex
# H36M joint set
self.h36m_joints_name = ('Pelvis', 'R_Hip', 'R_Knee', 'R_Ankle',
'L_Hip', 'L_Knee', 'L_Ankle', 'Torso', 'Neck',
'Nose', 'Head_top', 'L_Shoulder', 'L_Elbow',
'L_Wrist', 'R_Shoulder', 'R_Elbow', 'R_Wrist')
self.h36m_root_joint_idx = self.h36m_joints_name.index('Pelvis')
self.h36m_eval_joint = (1, 2, 3, 4, 5, 6, 8, 10, 11, 12, 13, 14, 15,
16)
self.h36m_joint_regressor = np.load(
osp.join('..', 'data', 'Human36M', 'J_regressor_h36m_correct.npy'))
self.datalist = self.load_data()
def __init__(self, pose_backbone, pose_net, pose2feat, mesh_backbone,
mesh_net, param_regressor):
super(Model, self).__init__()
self.pose_backbone = pose_backbone
self.pose_net = pose_net
self.pose2feat = pose2feat
self.mesh_backbone = mesh_backbone
self.mesh_net = mesh_net
self.param_regressor = param_regressor
if 'FreiHAND' in cfg.trainset_3d + cfg.trainset_2d + [cfg.testset]:
self.human_model = MANO()
self.human_model_layer = self.human_model.layer.cuda()
else:
self.human_model = SMPL()
self.human_model_layer = self.human_model.layer['neutral'].cuda()
self.root_joint_idx = self.human_model.root_joint_idx
self.mesh_face = self.human_model.face
self.joint_regressor = self.human_model.joint_regressor
if cfg.stage == 'lixel':
self.trainable_modules = [
self.pose_backbone, self.pose_net, self.pose2feat,
self.mesh_backbone, self.mesh_net
]
else:
self.trainable_modules = [self.param_regressor]
self.coord_loss = CoordLoss()
self.param_loss = ParamLoss()
self.normal_loss = NormalVectorLoss(self.mesh_face)
self.edge_loss = EdgeLengthLoss(self.mesh_face)
def __init__(self, transform, data_split):
self.transform = transform
self.data_split = data_split
self.img_dir = osp.join('..', 'data', 'MuCo', 'data')
self.annot_path = osp.join('..', 'data', 'MuCo', 'data',
'MuCo-3DHP.json')
self.smpl_param_path = osp.join('..', 'data', 'MuCo', 'data',
'smpl_param.json')
self.fitting_thr = 25 # milimeter
# MuCo joint set
self.muco_joint_num = 21
self.muco_joints_name = ('Head_top', 'Thorax', 'R_Shoulder', 'R_Elbow',
'R_Wrist', 'L_Shoulder', 'L_Elbow', 'L_Wrist',
'R_Hip', 'R_Knee', 'R_Ankle', 'L_Hip',
'L_Knee', 'L_Ankle', 'Pelvis', 'Spine',
'Head', 'R_Hand', 'L_Hand', 'R_Toe', 'L_Toe')
self.muco_flip_pairs = ((2, 5), (3, 6), (4, 7), (8, 11), (9, 12),
(10, 13), (17, 18), (19, 20))
self.muco_skeleton = ((0, 16), (16, 1), (1, 15), (15, 14), (14, 8),
(14, 11), (8, 9), (9, 10), (10, 19), (11, 12),
(12, 13), (13, 20), (1, 2), (2, 3), (3, 4),
(4, 17), (1, 5), (5, 6), (6, 7), (7, 18))
self.muco_root_joint_idx = self.muco_joints_name.index('Pelvis')
# H36M joint set
self.h36m_joint_regressor = np.load(
osp.join('..', 'data', 'Human36M', 'J_regressor_h36m_correct.npy')
) # use h36m joint regrssor (only use subset from original muco joint set)
self.h36m_flip_pairs = ((1, 4), (2, 5), (3, 6), (14, 11), (15, 12),
(16, 13))
self.h36m_joints_name = ('Pelvis', 'R_Hip', 'R_Knee', 'R_Ankle',
'L_Hip', 'L_Knee', 'L_Ankle', 'Torso', 'Neck',
'Nose', 'Head_top', 'L_Shoulder', 'L_Elbow',
'L_Wrist', 'R_Shoulder', 'R_Elbow', 'R_Wrist')
self.h36m_root_joint_idx = self.h36m_joints_name.index('Pelvis')
# SMPL joint set
self.smpl = SMPL()
self.face = self.smpl.face
self.joint_regressor = self.smpl.joint_regressor
self.vertex_num = self.smpl.vertex_num
self.joint_num = self.smpl.joint_num
self.joints_name = self.smpl.joints_name
self.flip_pairs = self.smpl.flip_pairs
self.skeleton = self.smpl.skeleton
self.root_joint_idx = self.smpl.root_joint_idx
self.face_kps_vertex = self.smpl.face_kps_vertex
self.datalist = self.load_data()
def __init__(self, transform, data_split):
self.transform = transform
self.data_split = data_split
self.img_dir = osp.join('..', 'data', 'Human36M', 'images')
self.annot_path = osp.join('..', 'data', 'Human36M', 'annotations')
self.human_bbox_root_dir = osp.join('..', 'data', 'Human36M',
'rootnet_output',
'bbox_root_human36m_output.json')
self.action_name = [
'Directions', 'Discussion', 'Eating', 'Greeting', 'Phoning',
'Posing', 'Purchases', 'Sitting', 'SittingDown', 'Smoking',
'Photo', 'Waiting', 'Walking', 'WalkDog', 'WalkTogether'
]
self.fitting_thr = 25 # milimeter
# H36M joint set
self.h36m_joint_num = 17
self.h36m_joints_name = ('Pelvis', 'R_Hip', 'R_Knee', 'R_Ankle',
'L_Hip', 'L_Knee', 'L_Ankle', 'Torso', 'Neck',
'Nose', 'Head_top', 'L_Shoulder', 'L_Elbow',
'L_Wrist', 'R_Shoulder', 'R_Elbow', 'R_Wrist')
self.h36m_flip_pairs = ((1, 4), (2, 5), (3, 6), (14, 11), (15, 12),
(16, 13))
self.h36m_skeleton = ((0, 7), (7, 8), (8, 9), (9, 10), (8, 11),
(11, 12), (12, 13), (8, 14), (14, 15), (15, 16),
(0, 1), (1, 2), (2, 3), (0, 4), (4, 5), (5, 6))
self.h36m_root_joint_idx = self.h36m_joints_name.index('Pelvis')
self.h36m_eval_joint = (1, 2, 3, 4, 5, 6, 8, 10, 11, 12, 13, 14, 15,
16)
self.h36m_joint_regressor = np.load(
osp.join('..', 'data', 'Human36M', 'J_regressor_h36m_correct.npy'))
# SMPL joint set
self.smpl = SMPL()
self.face = self.smpl.face
self.joint_regressor = self.smpl.joint_regressor
self.vertex_num = self.smpl.vertex_num
self.joint_num = self.smpl.joint_num
self.joints_name = self.smpl.joints_name
self.flip_pairs = self.smpl.flip_pairs
self.skeleton = self.smpl.skeleton
self.root_joint_idx = self.smpl.root_joint_idx
self.face_kps_vertex = self.smpl.face_kps_vertex
self.datalist = self.load_data()
def __init__(self, transform, data_split):
self.transform = transform
self.data_split = 'train' if data_split == 'train' else 'val'
self.img_path = osp.join('..', 'data', 'MSCOCO', 'images')
self.annot_path = osp.join('..', 'data', 'MSCOCO', 'annotations')
self.rootnet_output_path = osp.join('..', 'data', 'MSCOCO',
'rootnet_output',
'bbox_root_coco_output.json')
self.fitting_thr = 3.0 # pixel in cfg.output_hm_shape space
# mscoco skeleton
self.coco_joint_num = 18 # original: 17, manually added pelvis
self.coco_joints_name = ('Nose', 'L_Eye', 'R_Eye', 'L_Ear', 'R_Ear',
'L_Shoulder', 'R_Shoulder', 'L_Elbow',
'R_Elbow', 'L_Wrist', 'R_Wrist', 'L_Hip',
'R_Hip', 'L_Knee', 'R_Knee', 'L_Ankle',
'R_Ankle', 'Pelvis')
self.coco_skeleton = ((1, 2), (0, 1), (0, 2), (2, 4), (1, 3), (6, 8),
(8, 10), (5, 7), (7, 9), (12, 14), (14, 16),
(11, 13), (13, 15), (5, 6), (11, 12))
self.coco_flip_pairs = ((1, 2), (3, 4), (5, 6), (7, 8), (9, 10),
(11, 12), (13, 14), (15, 16))
self.coco_joint_regressor = np.load(
osp.join('..', 'data', 'MSCOCO', 'J_regressor_coco_hip_smpl.npy'))
# smpl skeleton
self.smpl = SMPL()
self.face = self.smpl.face
self.joint_regressor = self.smpl.joint_regressor
self.vertex_num = self.smpl.vertex_num
self.joint_num = self.smpl.joint_num
self.joints_name = self.smpl.joints_name
self.flip_pairs = self.smpl.flip_pairs
self.skeleton = self.smpl.skeleton
self.root_joint_idx = self.smpl.root_joint_idx
self.face_kps_vertex = self.smpl.face_kps_vertex
self.datalist = self.load_data()
def __init__(self,
step_size=1e-2,
batch_size=66,
num_iters=100,
focal_length=5000,
device=torch.device('cuda')):
# Store options
self.device = device
self.focal_length = focal_length
self.step_size = step_size
# Ignore the the following joints for the fitting process
ign_joints = ['OP Neck', 'OP RHip', 'OP LHip', 'Right Hip', 'Left Hip']
self.ign_joints = [constants.JOINT_IDS[i] for i in ign_joints]
self.num_iters = num_iters
# GMM pose prior
self.pose_prior = MaxMixturePrior(prior_folder='data',
num_gaussians=8,
dtype=torch.float32).to(device)
# Load SMPL model
self.smpl = SMPL(config.SMPL_MODEL_DIR,
batch_size=batch_size,
create_transl=False).to(self.device)
sys.path.append('..')
import os
import cv2
import ipdb
import torch
import config
import pickle
import numpy as np
from tqdm import tqdm
from utils.smpl import SMPL
from utils.geometry import batch_rodrigues
# --- predefined variables
device = torch.device('cpu')
smpl_male = SMPL(config.SMPL_MODEL_DIR, gender='male',
create_transl=False).to(device)
smpl_female = SMPL(config.SMPL_MODEL_DIR, gender='female',
create_transl=False).to(device)
def projection(smpl, smpl_trans, camPose, camIntrinsics):
"""
projection annoted 3D joints to 2D, so that we can obtain GT 2D joints.
"""
smpl += smpl_trans
smpl = np.concatenate([smpl, np.ones((49, 1))], axis=1)
smpl = np.dot(smpl, camPose.T)[:, :3]
smpl /= smpl[:, np.newaxis, -1]
smpl = np.dot(smpl, camIntrinsics.T)
return smpl[:, :2]
#import ipdb
#ipdb.set_trace()
joints = (joints + 1) / 2 * 500 + 500
for idx in range(joints.shape[0]):
pt = joints[idx]
cv2.circle(image, (pt[0], pt[1]), 4, np.array([197, 27, 125]).tolist(), -1)
cv2.imwrite('tmp.jpg', image)
if __name__ == '__main__':
idx = 0
mosh_file = '/data/neutrMosh/neutrSMPL_H3.6/S5/Directions 1_cam1_aligned.pkl'
images = '/data/human36m_full_raw/images/S5_Directions_1.54138969_{:06d}.jpg'.format(idx+1)
mosh = pkl.load(open(mosh_file, 'rb'), encoding='latin1')
pose = mosh['new_poses'][idx]
shape = mosh['betas']
pose = torch.from_numpy(pose).float().unsqueeze(0)
shape = torch.from_numpy(shape).float().unsqueeze(0)
smpl = SMPL(config.SMPL_MODEL_DIR,
batch_size=1,
create_transl=False).cpu()
smpl_out = smpl(betas=shape, body_pose=pose[:,3:], global_orient=pose[:, :3])
vertices = smpl_out.vertices
joints = smpl_out.joints
scatter_pose(cv2.imread(images), joints[0])