Fix AI tools and My Library, add U2net background removal
- Fix My Library: Add CSS styling for library browser, items now visible - Integrate My Library into Shapes tool with tabbed interface - Improve AI Inpaint: Add transform mode for scaling/sizing selections - Add helpful guidance explaining inpaint vs transform modes - Add U2net as alternative background removal (avoids rembg issues) - Create U2net model definition and download script - Improve Caddyfile with multiple options and troubleshooting guide Note: Brush Select (AI Paint) tool was already implemented and working. https://claude.ai/code/session_01CLedz6CanT9t46KBvng3vz
This commit is contained in:
Claude
+138
-20
@@ -128,53 +128,171 @@ async def inpaint_base64(request: InpaintRequest):
|
||||
raise HTTPException(status_code=500, detail=str(e))
|
||||
|
||||
|
||||
class RemoveBackgroundRequestV2(BaseModel):
|
||||
image: str # Base64 encoded image
|
||||
model: Optional[str] = "auto" # "auto", "u2net", "rembg", "birefnet"
|
||||
|
||||
|
||||
@router.post("/remove-background-base64")
|
||||
async def remove_background_base64(request: RemoveBackgroundRequest):
|
||||
"""
|
||||
Remove background from a base64 encoded image using rembg with BiRefNet.
|
||||
BiRefNet is state-of-the-art for background removal (better than u2net).
|
||||
Remove background from a base64 encoded image.
|
||||
Tries multiple methods: U2Net (direct), rembg with BiRefNet, rembg default.
|
||||
Returns base64 encoded PNG with transparent background.
|
||||
Used by miniPaint frontend.
|
||||
"""
|
||||
try:
|
||||
from rembg import remove, new_session
|
||||
except ImportError:
|
||||
raise HTTPException(
|
||||
status_code=500,
|
||||
detail="rembg not installed. Run: pip install rembg"
|
||||
)
|
||||
|
||||
try:
|
||||
# Decode base64 image
|
||||
image_bytes = base64.b64decode(request.image)
|
||||
img = Image.open(BytesIO(image_bytes)).convert('RGB')
|
||||
|
||||
# Use BiRefNet model for best quality (state-of-the-art)
|
||||
# Falls back to default model if BiRefNet not available
|
||||
result_bytes = None
|
||||
method_used = None
|
||||
|
||||
# Try U2Net first (direct implementation, no rembg dependency issues)
|
||||
try:
|
||||
session = new_session("birefnet-general")
|
||||
result_bytes = remove(image_bytes, session=session)
|
||||
except Exception:
|
||||
# Fallback to default model
|
||||
result_bytes = remove(image_bytes)
|
||||
result_bytes = await _remove_background_u2net(img)
|
||||
method_used = "u2net"
|
||||
except Exception as e:
|
||||
print(f"U2Net failed: {e}")
|
||||
|
||||
# Fall back to rembg if U2Net failed
|
||||
if result_bytes is None:
|
||||
try:
|
||||
from rembg import remove, new_session
|
||||
try:
|
||||
session = new_session("birefnet-general")
|
||||
result_bytes = remove(image_bytes, session=session)
|
||||
method_used = "birefnet"
|
||||
except Exception:
|
||||
result_bytes = remove(image_bytes)
|
||||
method_used = "rembg-default"
|
||||
except ImportError:
|
||||
pass
|
||||
except Exception as e:
|
||||
print(f"rembg failed: {e}")
|
||||
|
||||
if result_bytes is None:
|
||||
raise HTTPException(
|
||||
status_code=500,
|
||||
detail="No background removal method available. Install u2net or rembg."
|
||||
)
|
||||
|
||||
# Convert result to base64
|
||||
result_b64 = base64.b64encode(result_bytes).decode('utf-8')
|
||||
|
||||
# Get dimensions
|
||||
img = Image.open(BytesIO(result_bytes))
|
||||
result_img = Image.open(BytesIO(result_bytes))
|
||||
|
||||
return {
|
||||
"result": result_b64,
|
||||
"width": img.width,
|
||||
"height": img.height
|
||||
"width": result_img.width,
|
||||
"height": result_img.height,
|
||||
"method": method_used
|
||||
}
|
||||
|
||||
except HTTPException:
|
||||
raise
|
||||
except Exception as e:
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
raise HTTPException(status_code=500, detail=str(e))
|
||||
|
||||
|
||||
# Global U2Net model cache
|
||||
_u2net_model = None
|
||||
|
||||
|
||||
async def _remove_background_u2net(img: Image.Image) -> bytes:
|
||||
"""
|
||||
Remove background using U2Net model directly.
|
||||
This avoids rembg dependency issues while providing good quality.
|
||||
"""
|
||||
global _u2net_model
|
||||
|
||||
import torch
|
||||
from pathlib import Path
|
||||
|
||||
# Check for U2Net model
|
||||
models_dir = Path('/app/data/models')
|
||||
u2net_path = models_dir / 'u2net.pth'
|
||||
|
||||
# Also check alternative names
|
||||
if not u2net_path.exists():
|
||||
for alt_name in ['u2net.onnx', 'u2netp.pth', 'u2net_human_seg.pth']:
|
||||
alt_path = models_dir / alt_name
|
||||
if alt_path.exists():
|
||||
u2net_path = alt_path
|
||||
break
|
||||
|
||||
if not u2net_path.exists():
|
||||
raise FileNotFoundError(
|
||||
f"U2Net model not found at {u2net_path}. "
|
||||
"Download from: https://github.com/xuebinqin/U-2-Net"
|
||||
)
|
||||
|
||||
# Load model if not cached
|
||||
if _u2net_model is None:
|
||||
print(f"Loading U2Net model from {u2net_path}")
|
||||
|
||||
if str(u2net_path).endswith('.onnx'):
|
||||
# Use ONNX runtime
|
||||
import onnxruntime as ort
|
||||
_u2net_model = ort.InferenceSession(str(u2net_path))
|
||||
else:
|
||||
# Use PyTorch
|
||||
from app.services.u2net_model import U2NET
|
||||
_u2net_model = U2NET(3, 1)
|
||||
_u2net_model.load_state_dict(torch.load(str(u2net_path), map_location='cpu'))
|
||||
_u2net_model.eval()
|
||||
|
||||
print("U2Net model loaded")
|
||||
|
||||
# Preprocess image
|
||||
img_np = np.array(img)
|
||||
original_size = img.size
|
||||
|
||||
# Resize to model input size
|
||||
input_size = 320
|
||||
img_resized = img.resize((input_size, input_size), Image.Resampling.BILINEAR)
|
||||
img_np = np.array(img_resized).astype(np.float32)
|
||||
|
||||
# Normalize
|
||||
img_np = img_np / 255.0
|
||||
img_np = (img_np - [0.485, 0.456, 0.406]) / [0.229, 0.224, 0.225]
|
||||
img_np = img_np.transpose(2, 0, 1) # HWC to CHW
|
||||
img_np = np.expand_dims(img_np, 0) # Add batch dimension
|
||||
|
||||
# Run inference
|
||||
if hasattr(_u2net_model, 'run'):
|
||||
# ONNX runtime
|
||||
input_name = _u2net_model.get_inputs()[0].name
|
||||
outputs = _u2net_model.run(None, {input_name: img_np})
|
||||
mask = outputs[0][0, 0]
|
||||
else:
|
||||
# PyTorch
|
||||
with torch.no_grad():
|
||||
input_tensor = torch.from_numpy(img_np).float()
|
||||
d1, d2, d3, d4, d5, d6, d7 = _u2net_model(input_tensor)
|
||||
mask = d1[0, 0].numpy()
|
||||
|
||||
# Post-process mask
|
||||
mask = (mask - mask.min()) / (mask.max() - mask.min() + 1e-8)
|
||||
mask = (mask * 255).astype(np.uint8)
|
||||
|
||||
# Resize mask back to original size
|
||||
mask_img = Image.fromarray(mask).resize(original_size, Image.Resampling.BILINEAR)
|
||||
|
||||
# Apply mask to original image
|
||||
result = img.convert('RGBA')
|
||||
result.putalpha(mask_img)
|
||||
|
||||
# Save to bytes
|
||||
buffer = BytesIO()
|
||||
result.save(buffer, format='PNG')
|
||||
return buffer.getvalue()
|
||||
|
||||
|
||||
@router.post("/remove-background")
|
||||
async def remove_background(
|
||||
project_id: Optional[int] = Form(None),
|
||||
|
||||
@@ -0,0 +1,500 @@
|
||||
"""
|
||||
U2Net Model Definition for Background Removal
|
||||
Based on: https://github.com/xuebinqin/U-2-Net
|
||||
|
||||
This is a simplified implementation that works with the standard U2Net weights.
|
||||
"""
|
||||
|
||||
import torch
|
||||
import torch.nn as nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
|
||||
class REBNCONV(nn.Module):
|
||||
def __init__(self, in_ch=3, out_ch=3, dirate=1):
|
||||
super(REBNCONV, self).__init__()
|
||||
|
||||
self.conv_s1 = nn.Conv2d(in_ch, out_ch, 3, padding=1*dirate, dilation=1*dirate)
|
||||
self.bn_s1 = nn.BatchNorm2d(out_ch)
|
||||
self.relu_s1 = nn.ReLU(inplace=True)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
xout = self.relu_s1(self.bn_s1(self.conv_s1(hx)))
|
||||
return xout
|
||||
|
||||
|
||||
def _upsample_like(src, tar):
|
||||
src = F.interpolate(src, size=tar.shape[2:], mode='bilinear', align_corners=False)
|
||||
return src
|
||||
|
||||
|
||||
class RSU7(nn.Module):
|
||||
def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
|
||||
super(RSU7, self).__init__()
|
||||
|
||||
self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
|
||||
|
||||
self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
|
||||
self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool5 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv6 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
|
||||
self.rebnconv7 = REBNCONV(mid_ch, mid_ch, dirate=2)
|
||||
|
||||
self.rebnconv6d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv5d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv4d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
hxin = self.rebnconvin(hx)
|
||||
|
||||
hx1 = self.rebnconv1(hxin)
|
||||
hx = self.pool1(hx1)
|
||||
|
||||
hx2 = self.rebnconv2(hx)
|
||||
hx = self.pool2(hx2)
|
||||
|
||||
hx3 = self.rebnconv3(hx)
|
||||
hx = self.pool3(hx3)
|
||||
|
||||
hx4 = self.rebnconv4(hx)
|
||||
hx = self.pool4(hx4)
|
||||
|
||||
hx5 = self.rebnconv5(hx)
|
||||
hx = self.pool5(hx5)
|
||||
|
||||
hx6 = self.rebnconv6(hx)
|
||||
|
||||
hx7 = self.rebnconv7(hx6)
|
||||
|
||||
hx6d = self.rebnconv6d(torch.cat((hx7, hx6), 1))
|
||||
hx6dup = _upsample_like(hx6d, hx5)
|
||||
|
||||
hx5d = self.rebnconv5d(torch.cat((hx6dup, hx5), 1))
|
||||
hx5dup = _upsample_like(hx5d, hx4)
|
||||
|
||||
hx4d = self.rebnconv4d(torch.cat((hx5dup, hx4), 1))
|
||||
hx4dup = _upsample_like(hx4d, hx3)
|
||||
|
||||
hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1))
|
||||
hx3dup = _upsample_like(hx3d, hx2)
|
||||
|
||||
hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
|
||||
hx2dup = _upsample_like(hx2d, hx1)
|
||||
|
||||
hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
|
||||
|
||||
return hx1d + hxin
|
||||
|
||||
|
||||
class RSU6(nn.Module):
|
||||
def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
|
||||
super(RSU6, self).__init__()
|
||||
|
||||
self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
|
||||
|
||||
self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
|
||||
self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
|
||||
self.rebnconv6 = REBNCONV(mid_ch, mid_ch, dirate=2)
|
||||
|
||||
self.rebnconv5d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv4d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
hxin = self.rebnconvin(hx)
|
||||
|
||||
hx1 = self.rebnconv1(hxin)
|
||||
hx = self.pool1(hx1)
|
||||
|
||||
hx2 = self.rebnconv2(hx)
|
||||
hx = self.pool2(hx2)
|
||||
|
||||
hx3 = self.rebnconv3(hx)
|
||||
hx = self.pool3(hx3)
|
||||
|
||||
hx4 = self.rebnconv4(hx)
|
||||
hx = self.pool4(hx4)
|
||||
|
||||
hx5 = self.rebnconv5(hx)
|
||||
|
||||
hx6 = self.rebnconv6(hx5)
|
||||
|
||||
hx5d = self.rebnconv5d(torch.cat((hx6, hx5), 1))
|
||||
hx5dup = _upsample_like(hx5d, hx4)
|
||||
|
||||
hx4d = self.rebnconv4d(torch.cat((hx5dup, hx4), 1))
|
||||
hx4dup = _upsample_like(hx4d, hx3)
|
||||
|
||||
hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1))
|
||||
hx3dup = _upsample_like(hx3d, hx2)
|
||||
|
||||
hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
|
||||
hx2dup = _upsample_like(hx2d, hx1)
|
||||
|
||||
hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
|
||||
|
||||
return hx1d + hxin
|
||||
|
||||
|
||||
class RSU5(nn.Module):
|
||||
def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
|
||||
super(RSU5, self).__init__()
|
||||
|
||||
self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
|
||||
|
||||
self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
|
||||
self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool3 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
|
||||
self.rebnconv5 = REBNCONV(mid_ch, mid_ch, dirate=2)
|
||||
|
||||
self.rebnconv4d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
hxin = self.rebnconvin(hx)
|
||||
|
||||
hx1 = self.rebnconv1(hxin)
|
||||
hx = self.pool1(hx1)
|
||||
|
||||
hx2 = self.rebnconv2(hx)
|
||||
hx = self.pool2(hx2)
|
||||
|
||||
hx3 = self.rebnconv3(hx)
|
||||
hx = self.pool3(hx3)
|
||||
|
||||
hx4 = self.rebnconv4(hx)
|
||||
|
||||
hx5 = self.rebnconv5(hx4)
|
||||
|
||||
hx4d = self.rebnconv4d(torch.cat((hx5, hx4), 1))
|
||||
hx4dup = _upsample_like(hx4d, hx3)
|
||||
|
||||
hx3d = self.rebnconv3d(torch.cat((hx4dup, hx3), 1))
|
||||
hx3dup = _upsample_like(hx3d, hx2)
|
||||
|
||||
hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
|
||||
hx2dup = _upsample_like(hx2d, hx1)
|
||||
|
||||
hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
|
||||
|
||||
return hx1d + hxin
|
||||
|
||||
|
||||
class RSU4(nn.Module):
|
||||
def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
|
||||
super(RSU4, self).__init__()
|
||||
|
||||
self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
|
||||
|
||||
self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
|
||||
self.pool1 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
self.pool2 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=1)
|
||||
|
||||
self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=2)
|
||||
|
||||
self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=1)
|
||||
self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
hxin = self.rebnconvin(hx)
|
||||
|
||||
hx1 = self.rebnconv1(hxin)
|
||||
hx = self.pool1(hx1)
|
||||
|
||||
hx2 = self.rebnconv2(hx)
|
||||
hx = self.pool2(hx2)
|
||||
|
||||
hx3 = self.rebnconv3(hx)
|
||||
|
||||
hx4 = self.rebnconv4(hx3)
|
||||
|
||||
hx3d = self.rebnconv3d(torch.cat((hx4, hx3), 1))
|
||||
hx3dup = _upsample_like(hx3d, hx2)
|
||||
|
||||
hx2d = self.rebnconv2d(torch.cat((hx3dup, hx2), 1))
|
||||
hx2dup = _upsample_like(hx2d, hx1)
|
||||
|
||||
hx1d = self.rebnconv1d(torch.cat((hx2dup, hx1), 1))
|
||||
|
||||
return hx1d + hxin
|
||||
|
||||
|
||||
class RSU4F(nn.Module):
|
||||
def __init__(self, in_ch=3, mid_ch=12, out_ch=3):
|
||||
super(RSU4F, self).__init__()
|
||||
|
||||
self.rebnconvin = REBNCONV(in_ch, out_ch, dirate=1)
|
||||
|
||||
self.rebnconv1 = REBNCONV(out_ch, mid_ch, dirate=1)
|
||||
self.rebnconv2 = REBNCONV(mid_ch, mid_ch, dirate=2)
|
||||
self.rebnconv3 = REBNCONV(mid_ch, mid_ch, dirate=4)
|
||||
|
||||
self.rebnconv4 = REBNCONV(mid_ch, mid_ch, dirate=8)
|
||||
|
||||
self.rebnconv3d = REBNCONV(mid_ch*2, mid_ch, dirate=4)
|
||||
self.rebnconv2d = REBNCONV(mid_ch*2, mid_ch, dirate=2)
|
||||
self.rebnconv1d = REBNCONV(mid_ch*2, out_ch, dirate=1)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
hxin = self.rebnconvin(hx)
|
||||
|
||||
hx1 = self.rebnconv1(hxin)
|
||||
hx2 = self.rebnconv2(hx1)
|
||||
hx3 = self.rebnconv3(hx2)
|
||||
|
||||
hx4 = self.rebnconv4(hx3)
|
||||
|
||||
hx3d = self.rebnconv3d(torch.cat((hx4, hx3), 1))
|
||||
hx2d = self.rebnconv2d(torch.cat((hx3d, hx2), 1))
|
||||
hx1d = self.rebnconv1d(torch.cat((hx2d, hx1), 1))
|
||||
|
||||
return hx1d + hxin
|
||||
|
||||
|
||||
class U2NET(nn.Module):
|
||||
def __init__(self, in_ch=3, out_ch=1):
|
||||
super(U2NET, self).__init__()
|
||||
|
||||
self.stage1 = RSU7(in_ch, 32, 64)
|
||||
self.pool12 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage2 = RSU6(64, 32, 128)
|
||||
self.pool23 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage3 = RSU5(128, 64, 256)
|
||||
self.pool34 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage4 = RSU4(256, 128, 512)
|
||||
self.pool45 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage5 = RSU4F(512, 256, 512)
|
||||
self.pool56 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage6 = RSU4F(512, 256, 512)
|
||||
|
||||
# decoder
|
||||
self.stage5d = RSU4F(1024, 256, 512)
|
||||
self.stage4d = RSU4(1024, 128, 256)
|
||||
self.stage3d = RSU5(512, 64, 128)
|
||||
self.stage2d = RSU6(256, 32, 64)
|
||||
self.stage1d = RSU7(128, 16, 64)
|
||||
|
||||
self.side1 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
self.side2 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
self.side3 = nn.Conv2d(128, out_ch, 3, padding=1)
|
||||
self.side4 = nn.Conv2d(256, out_ch, 3, padding=1)
|
||||
self.side5 = nn.Conv2d(512, out_ch, 3, padding=1)
|
||||
self.side6 = nn.Conv2d(512, out_ch, 3, padding=1)
|
||||
|
||||
self.outconv = nn.Conv2d(6*out_ch, out_ch, 1)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
|
||||
# stage 1
|
||||
hx1 = self.stage1(hx)
|
||||
hx = self.pool12(hx1)
|
||||
|
||||
# stage 2
|
||||
hx2 = self.stage2(hx)
|
||||
hx = self.pool23(hx2)
|
||||
|
||||
# stage 3
|
||||
hx3 = self.stage3(hx)
|
||||
hx = self.pool34(hx3)
|
||||
|
||||
# stage 4
|
||||
hx4 = self.stage4(hx)
|
||||
hx = self.pool45(hx4)
|
||||
|
||||
# stage 5
|
||||
hx5 = self.stage5(hx)
|
||||
hx = self.pool56(hx5)
|
||||
|
||||
# stage 6
|
||||
hx6 = self.stage6(hx)
|
||||
hx6up = _upsample_like(hx6, hx5)
|
||||
|
||||
# decoder
|
||||
hx5d = self.stage5d(torch.cat((hx6up, hx5), 1))
|
||||
hx5dup = _upsample_like(hx5d, hx4)
|
||||
|
||||
hx4d = self.stage4d(torch.cat((hx5dup, hx4), 1))
|
||||
hx4dup = _upsample_like(hx4d, hx3)
|
||||
|
||||
hx3d = self.stage3d(torch.cat((hx4dup, hx3), 1))
|
||||
hx3dup = _upsample_like(hx3d, hx2)
|
||||
|
||||
hx2d = self.stage2d(torch.cat((hx3dup, hx2), 1))
|
||||
hx2dup = _upsample_like(hx2d, hx1)
|
||||
|
||||
hx1d = self.stage1d(torch.cat((hx2dup, hx1), 1))
|
||||
|
||||
# side output
|
||||
d1 = self.side1(hx1d)
|
||||
|
||||
d2 = self.side2(hx2d)
|
||||
d2 = _upsample_like(d2, d1)
|
||||
|
||||
d3 = self.side3(hx3d)
|
||||
d3 = _upsample_like(d3, d1)
|
||||
|
||||
d4 = self.side4(hx4d)
|
||||
d4 = _upsample_like(d4, d1)
|
||||
|
||||
d5 = self.side5(hx5d)
|
||||
d5 = _upsample_like(d5, d1)
|
||||
|
||||
d6 = self.side6(hx6)
|
||||
d6 = _upsample_like(d6, d1)
|
||||
|
||||
d0 = self.outconv(torch.cat((d1, d2, d3, d4, d5, d6), 1))
|
||||
|
||||
return torch.sigmoid(d0), torch.sigmoid(d1), torch.sigmoid(d2), torch.sigmoid(d3), torch.sigmoid(d4), torch.sigmoid(d5), torch.sigmoid(d6)
|
||||
|
||||
|
||||
class U2NETP(nn.Module):
|
||||
"""Smaller/faster U2Net variant (u2netp)"""
|
||||
|
||||
def __init__(self, in_ch=3, out_ch=1):
|
||||
super(U2NETP, self).__init__()
|
||||
|
||||
self.stage1 = RSU7(in_ch, 16, 64)
|
||||
self.pool12 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage2 = RSU6(64, 16, 64)
|
||||
self.pool23 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage3 = RSU5(64, 16, 64)
|
||||
self.pool34 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage4 = RSU4(64, 16, 64)
|
||||
self.pool45 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage5 = RSU4F(64, 16, 64)
|
||||
self.pool56 = nn.MaxPool2d(2, stride=2, ceil_mode=True)
|
||||
|
||||
self.stage6 = RSU4F(64, 16, 64)
|
||||
|
||||
# decoder
|
||||
self.stage5d = RSU4F(128, 16, 64)
|
||||
self.stage4d = RSU4(128, 16, 64)
|
||||
self.stage3d = RSU5(128, 16, 64)
|
||||
self.stage2d = RSU6(128, 16, 64)
|
||||
self.stage1d = RSU7(128, 16, 64)
|
||||
|
||||
self.side1 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
self.side2 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
self.side3 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
self.side4 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
self.side5 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
self.side6 = nn.Conv2d(64, out_ch, 3, padding=1)
|
||||
|
||||
self.outconv = nn.Conv2d(6*out_ch, out_ch, 1)
|
||||
|
||||
def forward(self, x):
|
||||
hx = x
|
||||
|
||||
hx1 = self.stage1(hx)
|
||||
hx = self.pool12(hx1)
|
||||
|
||||
hx2 = self.stage2(hx)
|
||||
hx = self.pool23(hx2)
|
||||
|
||||
hx3 = self.stage3(hx)
|
||||
hx = self.pool34(hx3)
|
||||
|
||||
hx4 = self.stage4(hx)
|
||||
hx = self.pool45(hx4)
|
||||
|
||||
hx5 = self.stage5(hx)
|
||||
hx = self.pool56(hx5)
|
||||
|
||||
hx6 = self.stage6(hx)
|
||||
hx6up = _upsample_like(hx6, hx5)
|
||||
|
||||
hx5d = self.stage5d(torch.cat((hx6up, hx5), 1))
|
||||
hx5dup = _upsample_like(hx5d, hx4)
|
||||
|
||||
hx4d = self.stage4d(torch.cat((hx5dup, hx4), 1))
|
||||
hx4dup = _upsample_like(hx4d, hx3)
|
||||
|
||||
hx3d = self.stage3d(torch.cat((hx4dup, hx3), 1))
|
||||
hx3dup = _upsample_like(hx3d, hx2)
|
||||
|
||||
hx2d = self.stage2d(torch.cat((hx3dup, hx2), 1))
|
||||
hx2dup = _upsample_like(hx2d, hx1)
|
||||
|
||||
hx1d = self.stage1d(torch.cat((hx2dup, hx1), 1))
|
||||
|
||||
d1 = self.side1(hx1d)
|
||||
|
||||
d2 = self.side2(hx2d)
|
||||
d2 = _upsample_like(d2, d1)
|
||||
|
||||
d3 = self.side3(hx3d)
|
||||
d3 = _upsample_like(d3, d1)
|
||||
|
||||
d4 = self.side4(hx4d)
|
||||
d4 = _upsample_like(d4, d1)
|
||||
|
||||
d5 = self.side5(hx5d)
|
||||
d5 = _upsample_like(d5, d1)
|
||||
|
||||
d6 = self.side6(hx6)
|
||||
d6 = _upsample_like(d6, d1)
|
||||
|
||||
d0 = self.outconv(torch.cat((d1, d2, d3, d4, d5, d6), 1))
|
||||
|
||||
return torch.sigmoid(d0), torch.sigmoid(d1), torch.sigmoid(d2), torch.sigmoid(d3), torch.sigmoid(d4), torch.sigmoid(d5), torch.sigmoid(d6)
|
||||
Reference in New Issue
Block a user