Skip to content

Commit

Permalink
Add files via upload
Browse files Browse the repository at this point in the history
  • Loading branch information
@Haozhe-Xing
Haozhe-Xing authored Apr 24, 2023
1 parent 00ae8c3 commit acb05c8
Show file tree
Hide file tree
Showing 2 changed files with 3 additions and 145 deletions.
6 changes: 3 additions & 3 deletions config.py
View file
Original file line number Diff line number Diff line change
@@ -1,10 +1,10 @@
import os

root = '/home/et21-xinghz/deep_learning_project'
datasets_root = os.path.join(root,'SODTR_1/data/Camouflaged_Object_Detection/COD10K')
root = ''
datasets_root = os.path.join(root,'COD10K')
cod_training_root = os.path.join(datasets_root, 'TrainDataset1')
chameleon_path = os.path.join(datasets_root, 'TestDataset/CHAMELEON')
camo_path = os.path.join(datasets_root, 'TestDataset/CAMO')
cod10k_path = os.path.join(datasets_root, 'TestDataset/COD10K')
nc4k_path = os.path.join(datasets_root, 'TestDataset/NC4K')
pvtv2_checkpoint_dir = '/home/et21-xinghz/deep_learning_project/PVTv2_Seg/checkpoint/'
pvtv2_checkpoint_dir = 'PVTv2_Seg/checkpoint/'
142 changes: 0 additions & 142 deletions new_infer.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -79,111 +79,6 @@ def main(exp_name,net,scale,results_path,pth_path):
end = time.time()
print("Total Testing Time: {}".format(str(datetime.timedelta(seconds=int(end - start)))))

def inference(exp_name,net,scale,img_dict,save_path,pth_path):
check_mkdir(save_path)
img_transform = transforms.Compose([
transforms.Resize((scale, scale)),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
to_pil = transforms.ToPILImage()
net.load_state_dict(torch.load(pth_path))
print('Load {} succeed!'.format(exp_name + '.pth'))
net.eval()
with torch.no_grad():
start = time.time()
for name, root in img_dict.items():
time_list = []
image_path = root
check_mkdir(os.path.join(save_path, name))
img_list = [os.path.splitext(f)[0] for f in os.listdir(image_path) if f.endswith('jpg')]
for idx, img_name in enumerate(img_list):
img = Image.open(os.path.join(image_path, img_name + '.jpg')).convert('RGB')
w, h = img.size
img_var = Variable(img_transform(img).unsqueeze(0)).cuda()
start_each = time.time()
predictions = net(img_var)
prediction = predictions[-1]
prediction = torch.sigmoid(prediction)
time_each = time.time() - start_each
time_list.append(time_each)

prediction = np.array(transforms.Resize((h, w))(to_pil(prediction.data.squeeze(0).cpu())))
Image.fromarray(prediction).convert('L').save(
os.path.join(save_path, name, img_name + '.png'))


print(('{}'.format(exp_name)))
print("{}'s average Time Is : {:.3f} s".format(name, mean(time_list)))
print("{}'s average Time Is : {:.1f} fps".format(name, 1 / mean(time_list)))

end = time.time()
print("Total Testing Time: {}".format(str(datetime.timedelta(seconds=int(end - start)))))


def evaluation_VCOD(exp_name,net,scale,results_path,pth_path):
check_mkdir(results_path)
VCOD_testdateset = {

'MoCA': '/home/et21-xinghz/deep_learning_project/SODTR_1/data/Camouflaged_Object_Detection/MoCA_Video/TestDataset',
#'CAD': '/home/et21-xinghz/deep_learning_project/SODTR_1/data/Camouflaged_Object_Detection/CAD',
# 'VCOD_animals':'/home/et21-xinghz/deep_learning_project/VideoCOD/VCOD_frame/animals',
# 'VCOD_artist':'/home/et21-xinghz/deep_learning_project/VideoCOD/VCOD_frame/artist',
# 'VCOD_solider':'/home/et21-xinghz/deep_learning_project/VideoCOD/VCOD_frame/solider'
}
#results = OrderedDict()
img_transform = transforms.Compose([
transforms.Resize((scale,scale)),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])
to_pil = transforms.ToPILImage()
net.load_state_dict(torch.load(pth_path))
print('Load {} succeed!'.format(exp_name+'.pth'))
net.eval()
with torch.no_grad():
start = time.time()
for name, root in VCOD_testdateset.items():
cal_total_seg_metrics = CalTotalMetric()
time_list = []
check_mkdir(os.path.join(results_path, exp_name, name))
for scene in os.listdir(root):
scene_path = os.path.join(root,scene)
save_path = os.path.join(results_path,exp_name,name,scene)
check_mkdir(save_path)
image_path = os.path.join(scene_path, 'Imgs')
mask_path = os.path.join(scene_path, 'GT')
img_list = [os.path.splitext(f)[0] for f in os.listdir(mask_path) if f.endswith('png')]
for idx, img_name in enumerate(img_list):
img = Image.open(os.path.join(image_path, img_name + '.jpg')).convert('RGB')

mask = np.array(Image.open(os.path.join(mask_path, img_name + '.png')).convert('L'))

w, h = img.size
img_var = Variable(img_transform(img).unsqueeze(0)).cuda()

start_each = time.time()

predictions = net(img_var)
prediction = predictions[-1]
prediction = torch.sigmoid(prediction)
time_each = time.time() - start_each
time_list.append(time_each)

prediction = np.array(transforms.Resize((h, w))(to_pil(prediction.data.squeeze(0).cpu())))
Image.fromarray(prediction).convert('L').save(
os.path.join(save_path, img_name + '.png'))

cal_total_seg_metrics.step(prediction, mask, mask_path)
print(('{}'.format(exp_name)))
print("{}'s average Time Is : {:.3f} s".format(name, mean(time_list)))
print("{}'s average Time Is : {:.1f} fps".format(name, 1 / mean(time_list)))
results = cal_total_seg_metrics.get_results()

print(results)

end = time.time()
print("Total Testing Time: {}".format(str(datetime.timedelta(seconds=int(end - start)))))

def evluation_with_resultspath(results_path,path_excel):
print(results_path)
Expand Down Expand Up @@ -220,43 +115,6 @@ def evluation_with_resultspath(results_path,path_excel):
print(results)
excel_logger(row_data=results, dataset_name=name, method_name=exp_name)

def evluation_with_resultspath_VCOD(results_path,path_excel):
print(results_path)
_, exp_name = os.path.split(results_path)
exp_name = exp_name + 'pseudo'
to_test = OrderedDict([
('MoCA-Mask', '/home/et21-xinghz/deep_learning_project/SODTR_1/data/Camouflaged_Object_Detection/MoCA_Video/TestDataset'),
#('CAD', '/home/et21-xinghz/deep_learning_project/SODTR_1/data/Camouflaged_Object_Detection/CAD')
])
excel_logger = MetricExcelRecorder(xlsx_path=path_excel, dataset_names=[name for name, root in to_test.items()])

for name, root in to_test.items():
# if not os.path.exists(os.path.join(results_path,name)):
# continue
cal_total_seg_metrics = CalTotalMetric()
for scene in os.listdir(root):
mask_path = os.path.join(root, scene, 'GT')
if exp_name == 'Ours_long_termpseudo' or exp_name == 'Ours_short_termpseudo':
results_scene_dir = os.path.join(results_path, scene, 'Pred')
else:
results_scene_dir = os.path.join(results_path, scene)
#results_scene_dir = os.path.join(results_path, scene)
mask_list = [os.path.splitext(f)[0] for f in os.listdir(mask_path) if f.endswith('png')]
for idx, img_name in enumerate(mask_list):
if not os.path.exists(os.path.join(results_scene_dir, img_name + '.png')):
continue
print(os.path.join(results_scene_dir,img_name + '.png'))
prediction = Image.open(os.path.join(results_scene_dir, img_name + '.png')).convert('L')

mask = Image.open(os.path.join(mask_path, img_name + '.png')).convert('L')
if not prediction.size == mask.size:
mask = mask.resize(prediction.size)
cal_total_seg_metrics.step(np.array(prediction), np.array(mask), mask_path)

results = cal_total_seg_metrics.get_results()
print(results)
excel_logger(row_data=results, dataset_name=name, method_name=exp_name)

def evaluation_COD(exp_name,net,scale,results_path,pth_path):
main(exp_name,net,scale,results_path,pth_path)

Expand Down

0 comments on commit acb05c8

Please sign in to comment.