def get_validation_set(opt, spatial_transform, temporal_transform,
target_transform):
assert opt.dataset in ['jester', 'egogesture', 'nv']
if opt.dataset == 'jester':
validation_data = Jester(opt.video_path,
opt.annotation_path,
'validation',
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'egogesture':
validation_data = EgoGesture(opt.video_path,
opt.annotation_path,
'validation',
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'nv':
validation_data = NV(opt.video_path,
opt.annotation_path,
'validation',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
return validation_data
def get_test_set(opt, spatial_transform, temporal_transform, target_transform):
assert opt.dataset in ['jester', 'egogesture', 'nv', 'sc', 'ems']
assert opt.test_subset in ['val', 'test']
if opt.test_subset == 'val':
subset = 'validation'
else:
subset = 'testing'
if opt.dataset == 'ems':
test_data = EMS(opt.video_path,
opt.annotation_path,
subset,
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'sc':
test_data = Jester(opt.video_path,
opt.annotation_path,
subset,
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'jester':
test_data = Jester(opt.video_path,
opt.annotation_path,
subset,
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'egogesture':
test_data = EgoGesture(opt.video_path,
opt.annotation_path,
subset,
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'nv':
test_data = NV(opt.video_path,
opt.annotation_path,
'validation',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
return test_data
def get_training_set(opt, spatial_transform, temporal_transform,
target_transform):
assert opt.dataset in [
'kinetics', 'jester', 'ucf101', 'egogesture', 'nvgesture'
]
if opt.train_validate:
subset = ['training', 'validation']
else:
subset = 'training'
if opt.dataset == 'kinetics':
training_data = Kinetics(opt.video_path,
opt.annotation_path,
'training',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration)
elif opt.dataset == 'jester':
training_data = Jester(opt.video_path,
opt.annotation_path,
'training',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration)
elif opt.dataset == 'ucf101':
training_data = UCF101(opt.video_path,
opt.annotation_path,
'training',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration)
elif opt.dataset == 'egogesture':
training_data = EgoGesture(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
elif opt.dataset == 'nvgesture':
training_data = NV(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
return training_data
def get_training_set(opt, spatial_transform, temporal_transform,
target_transform):
assert opt.dataset in ['jester', 'nv', 'ipn']
if opt.train_validate:
subset = ['training', 'validation']
else:
subset = 'training'
if opt.dataset == 'jester':
training_data = Jester(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
elif opt.dataset == 'egogesture':
training_data = EgoGesture(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
elif opt.dataset == 'nv':
training_data = NV(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
elif opt.dataset == 'ipn':
training_data = IPN(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
elif opt.dataset == 'denso':
training_data = Denso(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality,
no_subject_crop=opt.no_scrop)
elif opt.dataset == 'AHG':
training_data = AHG(opt.video_path,
opt.annotation_path,
subset,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality,
no_subject_crop=opt.no_scrop)
return training_data
def get_validation_set(opt, spatial_transform, temporal_transform,
target_transform):
assert opt.dataset in ['jester', 'nv', 'ipn']
if opt.dataset == 'jester':
validation_data = Jester(opt.video_path,
opt.annotation_path,
'validation',
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'egogesture':
validation_data = EgoGesture(opt.video_path,
opt.annotation_path,
'validation',
opt.n_val_samples,
spatial_transform,
temporal_transform,
target_transform,
modality=opt.modality,
sample_duration=opt.sample_duration)
elif opt.dataset == 'nv':
validation_data = NV(opt.video_path,
opt.annotation_path,
'validation',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
elif opt.dataset == 'ipn':
validation_data = IPN(opt.video_path,
opt.annotation_path,
'validation',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality)
elif opt.dataset == 'denso':
validation_data = Denso(opt.video_path,
opt.annotation_path,
'validation',
true_valid=opt.true_valid,
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality,
no_subject_crop=opt.no_scrop)
elif opt.dataset == 'AHG':
validation_data = AHG(opt.video_path,
opt.annotation_path,
'validation',
spatial_transform=spatial_transform,
temporal_transform=temporal_transform,
target_transform=target_transform,
sample_duration=opt.sample_duration,
modality=opt.modality,
no_subject_crop=opt.no_scrop)
return validation_data