def process_dir(self, dir_path, file_path, if_test=False):
datalist = [line for line in open(file_path, 'r').read().splitlines()]
pid_container = set()
for idx, item in enumerate(datalist):
img_rel_path, pid = item.split()
pid = int(pid)
pid_container.add(pid)
pid2label = {pid: label for label, pid in enumerate(pid_container)}
data = []
for idx, item in enumerate(datalist):
img_rel_path, pid = item.split()
img_path = osp.join(dir_path, img_rel_path)
img = read_image(img_path, True)
contour_path = img_path.replace('/rgb/', '/contour/')
contour_img = read_image(contour_path)
pid = int(pid)
camid = self.cam_var
self.cam_var += 1
if not if_test: pid = pid2label[pid]
# load data into memory
data.append((img_path, pid, camid, img, contour_img))
return data
def process_dir(self, dir_path, file_path, if_test=False):
datalist = [line for line in open(file_path, 'r').read().splitlines()]
# Specify data dir
sub_dir = 'train'
if if_test:
sub_dir = 'test'
dir_path = osp.join(dir_path, 'rgb', sub_dir)
pid_sample_cnts = dict()
for idx, item in enumerate(datalist):
img_rel_path, pid = item.split()
pid = int(pid)
if pid not in pid_sample_cnts:
pid_sample_cnts[pid] = 1
else:
pid_sample_cnts[pid] += 1
pid_container = set()
for pid, cnt in pid_sample_cnts.items():
if cnt >= self.at_least_num:
pid_container.add(pid)
pid2label = {pid: label for label, pid in enumerate(pid_container)}
data = []
for idx, item in enumerate(datalist):
img_rel_path, pid = item.split()
pid = int(pid)
camid = int(osp.basename(img_rel_path).split('_')
[2]) - 1 # index starts from 0
img_path = osp.join(dir_path, img_rel_path)
img = read_image(img_path, True)
contour_path = img_path.replace('/rgb/', '/contour/')
contour_img = read_image(contour_path)
if not if_test:
if pid in pid2label:
pid = pid2label[pid]
data.append((img_path, pid, camid, img, contour_img))
else:
data.append((img_path, pid, camid, img, contour_img))
return data
def process_dir(self, dir_path, file_path, if_test=False):
datalist = [line for line in open(file_path, 'r').read().splitlines()]
pid_sample_cnts = dict()
for idx, item in enumerate(datalist):
img_rel_path, pid, _ = item.split()
pid = int(pid)
if pid not in pid_sample_cnts:
pid_sample_cnts[pid] = 1
else:
pid_sample_cnts[pid] += 1
pid_container = set()
for pid, cnt in pid_sample_cnts.items():
if cnt >= self.at_least_num:
pid_container.add(pid)
pid2label = {pid: label for label, pid in enumerate(pid_container)}
data = []
for idx, item in enumerate(datalist):
img_rel_path, pid, camid = item.split()
pid, camid = int(pid), int(camid)
clothid = int(osp.basename(img_rel_path).split('_')[1])
img_path = osp.join(dir_path, img_rel_path)
img = read_image(img_path, True)
contour_path = img_path.replace('/rgb/', '/contour/').replace(
'.png', '.jpg')
try:
contour_img = read_image(contour_path)
except Exception:
continue
if not if_test:
if pid in pid2label:
pid = pid2label[pid]
data.append(
(img_path, pid, camid, clothid, img, contour_img))
else:
data.append((img_path, pid, camid, clothid, img, contour_img))
return data
def process_dir(self, dir_path, file_path, if_test=False):
datalist = [line for line in open(file_path, 'r').read().splitlines()]
pid_container = set()
for idx, item in enumerate(datalist):
img_rel_path, pid = item.split()
pid = int(pid)
pid_container.add(pid)
pid2label = {pid: label for label, pid in enumerate(pid_container)}
data = []
for idx, item in enumerate(datalist):
img_rel_path, pid = item.split()
img_path = osp.join(dir_path, img_rel_path)
img = read_image(img_path, True)
contour_path = img_path.replace('/rgb/', '/contour/')
contour_img = read_image(contour_path)
pid = int(pid)
if if_test:
camid = ROOM2CAMID[img_path.split('/')[-3]]
else:
cam = img_path.split('/')[-1].split('_')[0]
camid = ROOM2CAMID[cam]
assert 0 <= camid <= 2
if not if_test: pid = pid2label[pid]
# load data into memory
data.append((img_path, pid, camid, img, contour_img))
if not if_test:
dataset_len = len(data)
sample_factor = 0.5
sample_num = int(sample_factor * dataset_len)
random.shuffle(data)
data = data[:sample_num]
return data
def __getitem__(self, index):
img_paths, pid, camid = self.data[index]
num_imgs = len(img_paths)
if self.sample_method == 'random':
# Randomly samples seq_len images from a tracklet of length num_imgs,
# if num_imgs is smaller than seq_len, then replicates images
indices = np.arange(num_imgs)
replace = False if num_imgs >= self.seq_len else True
indices = np.random.choice(
indices, size=self.seq_len, replace=replace
)
# sort indices to keep temporal order (comment it to be order-agnostic)
indices = np.sort(indices)
elif self.sample_method == 'evenly':
# Evenly samples seq_len images from a tracklet
if num_imgs >= self.seq_len:
num_imgs -= num_imgs % self.seq_len
indices = np.arange(0, num_imgs, num_imgs / self.seq_len)
else:
# if num_imgs is smaller than seq_len, simply replicate the last image
# until the seq_len requirement is satisfied
indices = np.arange(0, num_imgs)
num_pads = self.seq_len - num_imgs
indices = np.concatenate(
[
indices,
np.ones(num_pads).astype(np.int32) * (num_imgs - 1)
]
)
assert len(indices) == self.seq_len
elif self.sample_method == 'all':
# Samples all images in a tracklet. batch_size must be set to 1
indices = np.arange(num_imgs)
else:
raise ValueError(
'Unknown sample method: {}'.format(self.sample_method)
)
imgs = []
for index in indices:
img_path = img_paths[int(index)]
img = util.read_image(img_path)
if self.transform is not None:
img = self.transform(img)
img = img.unsqueeze(0) # img must be torch.Tensor
imgs.append(img)
imgs = torch.cat(imgs, dim=0)
return imgs, pid, camid
def __getitem__(self, index):
img_path, pid, camid = self.data[index]
img = util.read_image(img_path)
if self.transform is not None:
img = self.transform(img)
return img, pid, camid, img_path