Initial Commit

app.py

@@ -0,0 +1,78 @@
+import os
+import matplotlib.pyplot as plt
+import cv2
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import numpy as np
+import gradio as gr
+from seg import U2NETP
+
+DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# Image processing utilities
+def load_image(path: str):
+    """ Loads an image from the specified path and converts it to RGB format. """
+    img = cv2.imread(path)
+    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
+    return img / 255.0
+
+def save_image(image: np.ndarray, path: str):
+    """ Saves an image to the specified path. """
+    img = (image * 255).astype(np.uint8)
+    img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
+    cv2.imwrite(path, img)
+
+# Document Segmentation Model
+class U2NETP_DocSeg(nn.Module):
+    def __init__(self, num_classes=1):
+        super(U2NETP_DocSeg, self).__init__()
+        self.u2netp = U2NETP(out_ch=num_classes)
+
+    def forward(self, x):
+        mask, *_ = self.u2netp(x)
+        return mask
+
+# Initialize the document segmentation model
+docseg = U2NETP_DocSeg(num_classes=1).to(DEVICE)
+# Load pretrained weights
+docseg_weight_path = './weights/u2netp_docseg_epoch_225_date_2026-01-02.pth'
+checkpoint = torch.load(docseg_weight_path)
+docseg.load_state_dict(checkpoint[f"model_state_dict"])
+docseg.eval()
+
+# Get segmentation mask
+def get_mask(image, confidence=0.5):
+    org_shape = image.shape[:2]
+    image_tensor = torch.from_numpy(image).float().permute(2, 0, 1).unsqueeze(0).to(DEVICE)
+    image_tensor = F.interpolate(image_tensor, size=(288, 288), mode='bilinear')
+    with torch.inference_mode():  # faster than no_grad
+        mask = docseg(image_tensor)
+        mask = (mask > confidence).float()
+        mask = F.interpolate(mask, size=org_shape, mode='bilinear')
+    return mask[0, 0]  # keep tensor
+
+def overlay_mask(image, mask):
+    image = torch.from_numpy(image).float().to(DEVICE)
+    red = torch.tensor([1.0, 0, 0], device=DEVICE).view(1, 3, 1, 1)
+    mask = mask.unsqueeze(0)  # (1, H, W)
+    mask = mask.unsqueeze(0)  # (1, 1, H, W)
+    overlay = image.permute(2, 0, 1).unsqueeze(0)
+    overlay = torch.where(mask > 0, red, overlay)
+    blended = 0.7 * image.permute(2, 0, 1).unsqueeze(0) + 0.3 * overlay
+    return blended[0].permute(1, 2, 0).cpu().numpy()
+
+def segment_image(image):
+    """ Gradio function to segment input image and return overlay. """
+    image = image.astype(np.float32) / 255.0  # Normalize to [0, 1]
+    mask = get_mask(image, confidence=0.5)
+    overlayed_image = overlay_mask(image, mask)
+    yield overlayed_image
+
+with gr.Blocks() as demo:
+    gr.Markdown("## Real-time Document Segmentation")
+    with gr.Row():
+        input_image = gr.Image(label="Input Image", type="numpy")
+        output_image = gr.Image(label="Segmentation Overlay", type="numpy")
+    
+    input_image.change(segment_image, inputs=input_image, outputs=output_image)

requirements.txt

@@ -0,0 +1,6 @@
+torch
+numpy
+matplotlib
+opencv-python
+gradio
+torchvision

seg.py

@@ -0,0 +1,682 @@
+import torch
+import torch.nn as nn
+from torchvision import models
+import torch.nn.functional as F
+import numpy as np
+
+
+class sobel_net(nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.conv_opx = nn.Conv2d(1, 1, 3, bias=False)
+        self.conv_opy = nn.Conv2d(1, 1, 3, bias=False)
+        sobel_kernelx = np.array([[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]], dtype='float32').reshape((1, 1, 3, 3))
+        sobel_kernely = np.array([[-1, -2, -1], [0, 0, 0], [1, 2, 1]], dtype='float32').reshape((1, 1, 3, 3))
+        self.conv_opx.weight.data = torch.from_numpy(sobel_kernelx)
+        self.conv_opy.weight.data = torch.from_numpy(sobel_kernely)
+
+        for p in self.parameters():
+            p.requires_grad = False
+
+    def forward(self, im):  # input rgb
+        x = (0.299 * im[:, 0, :, :] + 0.587 * im[:, 1, :, :] + 0.114 * im[:, 2, :, :]).unsqueeze(1)  # rgb2gray
+        gradx = self.conv_opx(x)
+        grady = self.conv_opy(x)
+
+        x = (gradx ** 2 + grady ** 2) ** 0.5
+        x = (x - x.min()) / (x.max() - x.min())
+        x = F.pad(x, (1, 1, 1, 1))
+
+        x = torch.cat([im, x], dim=1)
+        return x
+
+
+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
+
+
+## upsample tensor 'src' to have the same spatial size with tensor 'tar'
+def _upsample_like(src, tar):
+    src = F.interpolate(src, size=tar.shape[2:], mode='bilinear', align_corners=False)
+    return src
+
+
+### RSU-7 ###
+class RSU7(nn.Module):  # UNet07DRES(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
+
+
+### RSU-6 ###
+class RSU6(nn.Module):  # UNet06DRES(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
+
+
+### RSU-5 ###
+class RSU5(nn.Module):  # UNet05DRES(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
+
+
+### RSU-4 ###
+class RSU4(nn.Module):  # UNet04DRES(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
+
+
+### RSU-4F ###
+class RSU4F(nn.Module):  # UNet04FRES(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
+
+
+##### U^2-Net ####
+class U2NET(nn.Module):
+    def __init__(self, in_ch=3, out_ch=1):
+        super(U2NET, self).__init__()
+        self.edge = sobel_net()
+
+        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, 1)
+
+    def forward(self, x):
+        x = self.edge(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)
+
+### U^2-Net small ###
+class U2NETP(nn.Module):
+
+    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, 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 ClassifierHead(nn.Module):
+    def __init__(self, in_channels=64, channels=[512, 128], mode='avg_pool'):
+        super(ClassifierHead, self).__init__()
+        self.linears = nn.ModuleList()
+        for i, c in enumerate(channels):
+            if i == 0:
+                self.linears.append(nn.Linear(in_channels, c))
+            else:
+                self.linears.append(nn.Linear(channels[i-1], c))
+        self.cls = nn.Linear(channels[-1], 1)
+        self.available_modes = ['avg_pool', 'max_pool', 'flatten']
+        if mode not in self.available_modes:
+            raise ValueError("Mode must be one of: {}".format(self.available_modes))
+        self.mode = mode
+
+    def forward(self, x):
+        if self.mode == 'avg_pool':
+            x = F.adaptive_avg_pool2d(x, (1, 1))
+        elif self.mode == 'max_pool':
+            x = F.adaptive_max_pool2d(x, (1, 1))
+        elif self.mode == 'flatten':
+            x = torch.flatten(x, 1)
+        else:
+            raise ValueError("Unsupported mode: {}".format(self.mode))
+        # print("x shape after pooling:", x.shape)
+        x = x.view(x.size(0), -1)  # Flatten the tensor
+        for linear in self.linears:
+            x = F.relu(linear(x))
+        x = self.cls(x)
+        return x
+
+class U2NETP_v2(nn.Module):
+
+    def __init__(self, in_ch=3, out_ch=1):
+        super(U2NETP_v2, 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(out_ch * 6, 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)
+        # print(d1.shape, d2.shape, d3.shape, d4.shape, d5.shape, d6.shape)
+        d0 = self.outconv(torch.cat((d1, d2, d3, d4, d5, d6), 1))
+
+        return d0, hx6

weights/seg.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb79fdec55a5ed435dc74d8112aa9285d8213bae475022f711c709744fb19dd4
+size 4715923

weights/u2netp_docseg_epoch_225_date_2026-01-02.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b39f3eb35f0985dc168eea21d9007a8467a79a3f80baa668f2b9ff6112f31ef6
+size 14344719