Skip to content

fixed image part extraction #19

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
TheUser0815 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fixed image part extraction #19

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 21 additions & 11 deletions lib/models/axialnet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -512,6 +512,16 @@ def __init__(self, block, block_2, layers, num_classes=2, zero_init_residual=Tru
groups=8, width_per_group=64, replace_stride_with_dilation=None,
norm_layer=None, s=0.125, img_size = 128,imgchan = 3):
super(medt_net, self).__init__()

img_axis_parts=4 #could be used as param

self.img_size = img_size
self.img_axis_parts = img_axis_parts
self.img_size_p = int(img_size / img_axis_parts)

assert not img_size % self.img_size_p, "img size {} can not be split equally into {} parts".format(img_size, img_axis_parts)
assert not self.img_size_p % 8, "img part size {} must be divisible by 8 in order to create local branch blocks".format(self.img_size_p)

if norm_layer is None:
norm_layer = nn.BatchNorm2d
self._norm_layer = norm_layer
Expand Down Expand Up @@ -567,14 +577,12 @@ def __init__(self, block, block_2, layers, num_classes=2, zero_init_residual=Tru

self.relu_p = nn.ReLU(inplace=True)

img_size_p = img_size // 4

self.layer1_p = self._make_layer(block_2, int(128 * s), layers[0], kernel_size= (img_size_p//2))
self.layer2_p = self._make_layer(block_2, int(256 * s), layers[1], stride=2, kernel_size=(img_size_p//2),
self.layer1_p = self._make_layer(block_2, int(128 * s), layers[0], kernel_size= (self.img_size_p//2))
self.layer2_p = self._make_layer(block_2, int(256 * s), layers[1], stride=2, kernel_size=(self.img_size_p//2),
dilate=replace_stride_with_dilation[0])
self.layer3_p = self._make_layer(block_2, int(512 * s), layers[2], stride=2, kernel_size=(img_size_p//4),
self.layer3_p = self._make_layer(block_2, int(512 * s), layers[2], stride=2, kernel_size=(self.img_size_p//4),
dilate=replace_stride_with_dilation[1])
self.layer4_p = self._make_layer(block_2, int(1024 * s), layers[3], stride=2, kernel_size=(img_size_p//8),
self.layer4_p = self._make_layer(block_2, int(1024 * s), layers[3], stride=2, kernel_size=(self.img_size_p//8),
dilate=replace_stride_with_dilation[2])

# Decoder
Expand Down Expand Up @@ -656,12 +664,14 @@ def _forward_impl(self, x):

# y_out = torch.ones((1,2,128,128))
x_loc = x.clone()
img_size_p = self.img_size_p
# x = F.relu(F.interpolate(self.decoder5(x) , scale_factor=(2,2), mode ='bilinear'))
#start
for i in range(0,4):
for j in range(0,4):
for i in range(0,self.img_axis_parts):
for j in range(0,self.img_axis_parts):

x_p = xin[:,:,32*i:32*(i+1),32*j:32*(j+1)]
x_p = xin[:,:,img_size_p*i:img_size_p*(i+1),img_size_p*j:img_size_p*(j+1)]

# begin patch wise
x_p = self.conv1_p(x_p)
x_p = self.bn1_p(x_p)
Expand Down Expand Up @@ -697,7 +707,7 @@ def _forward_impl(self, x):
x_p = torch.add(x_p, x1_p)
x_p = F.relu(F.interpolate(self.decoder5_p(x_p) , scale_factor=(2,2), mode ='bilinear'))

x_loc[:,:,32*i:32*(i+1),32*j:32*(j+1)] = x_p
x_loc[:,:,img_size_p*i:img_size_p*(i+1),img_size_p*j:img_size_p*(j+1)] = x_p

x = torch.add(x,x_loc)
x = F.relu(self.decoderf(x))
Expand Down Expand Up @@ -727,4 +737,4 @@ def logo(pretrained=False, **kwargs):
model = medt_net(AxialBlock,AxialBlock, [1, 2, 4, 1], s= 0.125, **kwargs)
return model

# EOF
# EOF