def make_mot_transforms(image_set, args):
normalize = T.Compose([
T.ToTensor(),
T.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
scales = [480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800]
if image_set == 'train' and not args.eval:
return T.Compose([
T.RandomHorizontalFlip(),
T.RandomSelect(
T.RandomResize(scales, max_size=1333),
T.Compose([
T.RandomResize([800, 1000, 1200]),
# T.RandomSizeCrop(384, 600),
T.RandomSizeCrop_MOT(800, 1200),
T.RandomResize(scales, max_size=1333),
])
),
normalize,
])
if image_set == 'trainall' and not args.eval:
return T.Compose([
T.RandomHorizontalFlip(),
T.RandomSelect(
T.RandomResize(scales, max_size=1333),
T.Compose([
T.RandomResize([800, 1000, 1200]),
# T.RandomSizeCrop(384, 600),
T.RandomSizeCrop_MOT(800, 1200),
T.RandomResize(scales, max_size=1333),
])
),
normalize,
])
if image_set == 'val' or args.eval:
return T.Compose([
T.RandomResize([800], max_size=1333),
normalize,
])
if image_set == 'test' or args.eval:
return T.Compose([
T.RandomResize([800], max_size=1333),
normalize,
])
raise ValueError(f'unknown {image_set}')
def make_mot_transforms(image_set):
normalize = T.Compose([
T.ToTensor(),
T.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
# scales = [480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800]
scales = [480, 512, 544]
if image_set == 'train':
return T.Compose([
T.RandomHorizontalFlip(),
T.RandomSelect(
T.RandomResize(scales, max_size=960),
T.Compose([
T.RandomResize([400, 500, 600]),
# T.RandomSizeCrop(384, 600),
T.RandomSizeCrop_MOT(384, 600),
T.RandomResize(scales, max_size=960),
])),
normalize,
])
if image_set == 'val':
return T.Compose([
T.RandomResize([540], max_size=960),
normalize,
])
if image_set == 'test':
return T.Compose([
T.RandomResize([540], max_size=960),
normalize,
])
raise ValueError(f'unknown {image_set}')