File size: 10,983 Bytes
fe64bad |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 |
# # -*- coding: utf-8 -*-
#
# # Max-Planck-Gesellschaft zur Förderung der Wissenschaften e.V. (MPG) is
# # holder of all proprietary rights on this computer program.
# # You can only use this computer program if you have closed
# # a license agreement with MPG or you get the right to use the computer
# # program from someone who is authorized to grant you that right.
# # Any use of the computer program without a valid license is prohibited and
# # liable to prosecution.
# #
# # Copyright©2019 Max-Planck-Gesellschaft zur Förderung
# # der Wissenschaften e.V. (MPG). acting on behalf of its Max Planck Institute
# # for Intelligent Systems. All rights reserved.
# #
# # Contact: ps-license@tuebingen.mpg.de
#
# import torch
# import joblib
# import numpy as np
# import os.path as osp
# from torch.utils.data import Dataset
#
# # from lib.core.config import VIBE_DB_DIR
# VIBE_DB_DIR = '../VIBE/data/vibe_db'
# # from lib.data_utils.img_utils import split_into_chunks
#
# def split_into_chunks(vid_names, seqlen, stride):
# video_start_end_indices = []
#
# video_names, group = np.unique(vid_names, return_index=True)
# perm = np.argsort(group)
# video_names, group = video_names[perm], group[perm]
#
# indices = np.split(np.arange(0, vid_names.shape[0]), group[1:])
#
# for idx in range(len(video_names)):
# indexes = indices[idx]
# if indexes.shape[0] < seqlen:
# continue
# chunks = view_as_windows(indexes, (seqlen,), step=stride)
# start_finish = chunks[:, (0, -1)].tolist()
# video_start_end_indices += start_finish
#
# return video_start_end_indices
#
# class AMASS(Dataset):
# def __init__(self, seqlen):
# self.seqlen = seqlen
#
# self.stride = seqlen
#
# self.db = self.load_db()
# self.vid_indices = split_into_chunks(self.db['vid_name'], self.seqlen, self.stride)
# del self.db['vid_name']
# print(f'AMASS dataset number of videos: {len(self.vid_indices)}')
#
# def __len__(self):
# return len(self.vid_indices)
#
# def __getitem__(self, index):
# return self.get_single_item(index)
#
# def load_db(self):
# db_file = osp.join(VIBE_DB_DIR, 'amass_db.pt')
# db = joblib.load(db_file)
# return db
#
# def get_single_item(self, index):
# start_index, end_index = self.vid_indices[index]
# thetas = self.db['theta'][start_index:end_index+1]
#
# cam = np.array([1., 0., 0.])[None, ...]
# cam = np.repeat(cam, thetas.shape[0], axis=0)
# theta = np.concatenate([cam, thetas], axis=-1)
#
# target = {
# 'theta': torch.from_numpy(theta).float(), # cam, pose and shape
# }
# return target
import os
import numpy as np
import joblib
from .dataset import Dataset
from src.config import ROT_CONVENTION_TO_ROT_NUMBER
from src import config
from PIL import Image
import sys
sys.path.append('')
# action2motion_joints = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 21, 24, 38]
# change 0 and 8
action2motion_joints = [8, 1, 2, 3, 4, 5, 6, 7, 0, 9, 10, 11, 12, 13, 14, 21, 24, 38] # [18,]
from src.utils.action_label_to_idx import action_label_to_idx, idx_to_action_label
def get_z(cam_s, cam_pos, joints, img_size, flength):
"""
Solves for the depth offset of the model to approx. orth with persp camera.
"""
# Translate the model itself: Solve the best z that maps to orth_proj points
joints_orth_target = (cam_s * (joints[:, :2] + cam_pos) + 1) * 0.5 * img_size
height3d = np.linalg.norm(np.max(joints[:, :2], axis=0) - np.min(joints[:, :2], axis=0))
height2d = np.linalg.norm(np.max(joints_orth_target, axis=0) - np.min(joints_orth_target, axis=0))
tz = np.array(flength * (height3d / height2d))
return float(tz)
def get_trans_from_vibe(vibe, use_z=True):
alltrans = []
for t in range(vibe["joints3d"].shape[0]):
# Convert crop cam to orig cam
# No need! Because `convert_crop_cam_to_orig_img` from demoutils of vibe
# does this already for us :)
# Its format is: [sx, sy, tx, ty]
cam_orig = vibe["orig_cam"][t]
x = cam_orig[2]
y = cam_orig[3]
if use_z:
z = get_z(cam_s=cam_orig[0], # TODO: There are two scales instead of 1.
cam_pos=cam_orig[2:4],
joints=vibe['joints3d'][t],
img_size=480,
flength=500)
# z = 500 / (0.5 * 480 * cam_orig[0])
else:
z = 0
trans = [x, y, z]
alltrans.append(trans)
alltrans = np.array(alltrans)
return alltrans - alltrans[0]
class AMASS(Dataset):
dataname = "amass"
def __init__(self, datapath="data/amass/amass_30fps_legacy_db.pt", split="train", use_z=1, **kwargs):
assert '_db.pt' in datapath
self.datapath = datapath.replace('_db.pt', '_{}.pt'.format(split))
assert os.path.exists(self.datapath)
print('datapath used by amass is [{}]'.format(self.datapath))
super().__init__(**kwargs)
self.dataname = "amass"
# FIXME - hardcoded:
self.rot_convention = 'legacy'
self.use_betas = False
self.use_gender = False
self.use_body_features = False
if 'clip_preprocess' in kwargs.keys():
self.clip_preprocess = kwargs['clip_preprocess']
self.use_z = (use_z != 0)
# keep_actions = [6, 7, 8, 9, 22, 23, 24, 38, 80, 93, 99, 100, 102]
dummy_class = [0]
genders = config.GENDERS
self.num_classes = len(dummy_class)
self.db = self.load_db()
self._joints3d = []
self._poses = []
self._num_frames_in_video = []
self._actions = []
self._betas = []
self._genders = []
self._heights = []
self._masses = []
self._clip_images = []
self._clip_texts = []
self._clip_pathes = []
self._actions_cat = []
self.clip_label_text = "text_raw_labels" # "text_proc_labels"
seq_len = 100
n_sequences = len(self.db['thetas'])
# split sequences
for seq_idx in range(n_sequences):
n_sub_seq = self.db['thetas'][seq_idx].shape[0] // seq_len
if n_sub_seq == 0: continue
n_frames_in_use = n_sub_seq * seq_len
joints3d = np.split(self.db['joints3d'][seq_idx][:n_frames_in_use], n_sub_seq)
poses = np.split(self.db['thetas'][seq_idx][:n_frames_in_use], n_sub_seq)
self._joints3d.extend(joints3d)
self._poses.extend(poses)
self._num_frames_in_video.extend([seq_len] * n_sub_seq)
if 'action_cat' in self.db:
self._actions_cat.extend(np.split(self.db['action_cat'][seq_idx][:n_frames_in_use], n_sub_seq))
if self.use_betas:
self._betas.extend(np.split(self.db['betas'][seq_idx][:n_frames_in_use], n_sub_seq))
if self.use_gender:
self._genders.extend([str(self.db['genders'][seq_idx]).replace("b'female'", "female").replace("b'male'",
"male")] * n_sub_seq)
if self.use_body_features:
self._heights.extend([self.db['heights'][seq_idx]] * n_sub_seq)
self._masses.extend([self.db['masses'][seq_idx]] * n_sub_seq)
if 'clip_images' in self.db.keys():
images = [np.squeeze(e) for e in np.split(self.db['clip_images'][seq_idx][:n_sub_seq], n_sub_seq)]
processed_images = [self.clip_preprocess(Image.fromarray(img)) for img in images]
self._clip_images.extend(processed_images)
if self.clip_label_text in self.db:
self._clip_texts.extend(np.split(self.db[self.clip_label_text][seq_idx][:n_frames_in_use], n_sub_seq))
if 'clip_pathes' in self.db:
self._clip_pathes.extend(np.split(self.db['clip_pathes'][seq_idx][:n_sub_seq], n_sub_seq))
if 'clip_images_emb' in self.db.keys():
self._clip_images_emb.extend(np.split(self.db['clip_images_emb'][seq_idx][:n_sub_seq], n_sub_seq))
actions = [0] * n_sub_seq
self._actions.extend(actions)
assert len(self._num_frames_in_video) == len(self._poses) == len(self._joints3d) == len(self._actions)
if self.use_betas:
assert len(self._poses) == len(self._betas)
if self.use_gender:
assert len(self._poses) == len(self._genders)
if 'clip_images' in self.db.keys():
assert len(self._poses) == len(self._clip_images)
self._actions = np.array(self._actions)
self._num_frames_in_video = np.array(self._num_frames_in_video)
N = len(self._poses)
# same set for training and testing
self._train = np.arange(N)
self._test = np.arange(N)
self._action_to_label = {x: i for i, x in enumerate(dummy_class)}
self._label_to_action = {i: x for i, x in enumerate(dummy_class)}
self._gender_to_label = {x: i for i, x in enumerate(genders)}
self._label_to_gender = {i: x for i, x in enumerate(genders)}
self._action_classes = idx_to_action_label
def load_db(self):
# Load amass dataset encoded to a .db file
# The loaded data is structured:
# {
# 'theta': [data_size, 82] (float64) (structured [pose(72), betas(10)])
# 'vid_name': [data_size] (str)
# }
# data_size should be [16275369]
db_file = self.datapath
db = joblib.load(db_file)
if 'clip_images' in db and db['clip_images'][0] is None: # No images added
del db['clip_images']
return db
def _load_joints3D(self, ind, frame_ix):
joints3D = self._joints3d[ind][frame_ix]
return joints3D
def _load_rotvec(self, ind, frame_ix):
pose = self._poses[ind][frame_ix, :].reshape(-1, ROT_CONVENTION_TO_ROT_NUMBER[self.rot_convention] + 1,
3) # +1 for global orientation
return pose
def _load_betas(self, ind, frame_ix):
betas = self._betas[ind][frame_ix].transpose((1, 0))
return betas
def _load_gender(self, ind, frame_ix):
gender = self._gender_to_label[self._genders[ind]]
return gender
def _load_body_features(self, ind, frame_ix):
return {'mass': float(self._masses[ind]), 'height': float(self._heights[ind])}
if __name__ == "__main__":
dataset = AMASS()
|