aoxo commited on
Commit
f6ebf8d
·
verified ·
1 Parent(s): 73a0567

Upload mri_autoencoder.ipynb with huggingface_hub

Browse files
Files changed (1) hide show
  1. mri_autoencoder.ipynb +311 -40
mri_autoencoder.ipynb CHANGED
@@ -261,7 +261,7 @@
261
  " dcm = pydicom.dcmread(full_path)\n",
262
  " \n",
263
  " # Check image dimensions\n",
264
- " if dcm.pixel_array.shape[0] < min_size or dcm.pixel_array.shape[1] < min_size:\n",
265
  " files_to_delete.append(full_path)\n",
266
  " filtered_images += 1\n",
267
  " \n",
@@ -279,7 +279,7 @@
279
  " print(f\"Images deleted: {filtered_images}\\n\")\n",
280
  "\n",
281
  "# Usage\n",
282
- "input_dirs = [r\"D:\\TCIA_Split\\train\", r\"D:\\TCIA_Split\\val\"]\n",
283
  "filter_dicom_images(input_dirs)"
284
  ]
285
  },
@@ -928,7 +928,7 @@
928
  },
929
  {
930
  "cell_type": "code",
931
- "execution_count": null,
932
  "metadata": {},
933
  "outputs": [],
934
  "source": [
@@ -1347,9 +1347,9 @@
1347
  "# Example usage\n",
1348
  "if __name__ == \"__main__\":\n",
1349
  " # Paths to models and test image\n",
1350
- " RECONSTRUCTOR_MODEL_PATH = r\"D:/VSCODE/PreSense/small_reconstructor.pth\" # Path to your saved Reconstructor model\n",
1351
- " DENOISER_MODEL_PATH = r\"D:/VSCODE/PreSense/small_denoiser.pth\" # Path to your saved Denoiser model\n",
1352
- " TEST_DICOM_PATH = r\"D:/VSCODE/PreSense/test2.dcm\" # Replace with actual path to test DICOM \n",
1353
  " # Run inference\n",
1354
  " inference_single_image(RECONSTRUCTOR_MODEL_PATH, DENOISER_MODEL_PATH, TEST_DICOM_PATH)"
1355
  ]
@@ -1644,7 +1644,150 @@
1644
  "cell_type": "markdown",
1645
  "metadata": {},
1646
  "source": [
1647
- "### Larger Reconstructor and Denoiser U-Net (largeRD)"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1648
  ]
1649
  },
1650
  {
@@ -1691,38 +1834,41 @@
1691
  " self.bn2 = nn.BatchNorm2d(out_channels)\n",
1692
  " self.conv3 = nn.Conv2d(out_channels, out_channels, 3, padding=1)\n",
1693
  " self.bn3 = nn.BatchNorm2d(out_channels)\n",
 
 
1694
  " \n",
1695
  " def forward(self, x):\n",
1696
  " x = F.relu(self.bn1(self.conv1(x)))\n",
1697
  " x = F.relu(self.bn2(self.conv2(x)))\n",
1698
  " x = F.relu(self.bn3(self.conv3(x)))\n",
 
1699
  " return x\n",
1700
  "\n",
1701
  "class UNet(nn.Module):\n",
1702
  " def __init__(self, in_channels=1, out_channels=1):\n",
1703
  " super().__init__()\n",
1704
  " # Encoder\n",
1705
- " self.enc1 = UNetBlock(in_channels, 96)\n",
1706
- " self.enc2 = UNetBlock(96, 192)\n",
1707
- " self.enc3 = UNetBlock(192, 384)\n",
1708
- " self.enc4 = UNetBlock(384, 768)\n",
1709
- " self.enc5 = UNetBlock(768, 1536)\n",
1710
  " \n",
1711
  " # Decoder with learned upsampling (transposed convolutions)\n",
1712
  "\n",
1713
- " self.upconv5 = nn.ConvTranspose2d(1536, 768, kernel_size=2, stride=2) # Learnable upsampling\n",
1714
- " self.dec5 = UNetBlock(768 + 768, 768) # Adjust input channels after concatenation\n",
1715
  "\n",
1716
- " self.upconv4 = nn.ConvTranspose2d(768, 384, kernel_size=2, stride=2) # Learnable upsampling\n",
1717
- " self.dec4 = UNetBlock(384 + 384, 384) # Adjust input channels after concatenation\n",
1718
  "\n",
1719
- " self.upconv3 = nn.ConvTranspose2d(384, 192, kernel_size=2, stride=2) # Learnable upsampling\n",
1720
- " self.dec3 = UNetBlock(192 + 192, 192) # Adjust input channels after concatenation\n",
1721
  "\n",
1722
- " self.upconv2 = nn.ConvTranspose2d(192, 96, kernel_size=2, stride=2) # Learnable upsampling\n",
1723
- " self.dec2 = UNetBlock(96 + 96, 96) # Adjust input channels after concatenation\n",
1724
  "\n",
1725
- " self.dec1 = UNetBlock(96, out_channels) # Final output\n",
1726
  "\n",
1727
  " self.pool = nn.MaxPool2d(2, 2)\n",
1728
  " \n",
@@ -1737,30 +1883,22 @@
1737
  " # Decoder path with learned upsampling and skip connections\n",
1738
  "\n",
1739
  " d5 = self.upconv5(e5) # Learnable upsampling\n",
1740
- " # if d5.size(2) != e4.size(2) or d5.size(3) != e4.size(3):\n",
1741
- " # # Resize e4 to match d5's dimensions\n",
1742
- " # e4 = F.interpolate(e4, size=(d5.size(2), d5.size(3)), mode='nearest')\n",
1743
  " d5 = torch.cat([d5, e4], dim=1) # Concatenate with encoder features\n",
1744
  " d5 = checkpoint(self.dec5, d5)\n",
1745
  "\n",
1746
  " d4 = self.upconv4(d5) # Learnable upsampling\n",
1747
- " # if d4.size(2) != e2.size(2) or d4.size(3) != e2.size(3):\n",
1748
- " # # Resize e3 to match d4's dimensions\n",
1749
- " # e3 = F.interpolate(e3, size=(d4.size(2), d4.size(3)), mode='nearest')\n",
1750
  " d4 = torch.cat([d4, e3], dim=1) # Concatenate with encoder features\n",
1751
  " d4 = checkpoint(self.dec4, d4)\n",
1752
  "\n",
1753
  " d3 = self.upconv3(d4) # Learnable upsampling\n",
1754
- " # if d3.size(2) != e2.size(2) or d3.size(3) != e2.size(3):\n",
1755
- " # # Resize e2 to match d3's dimensions\n",
1756
- " # e2 = F.interpolate(e2, size=(d3.size(2), d3.size(3)), mode='nearest')\n",
1757
  " d3 = torch.cat([d3, e2], dim=1) # Concatenate with encoder features\n",
1758
  " d3 = checkpoint(self.dec3, d3)\n",
1759
  "\n",
1760
  " d2 = self.upconv2(d3) # Learnable upsampling\n",
1761
- " # if d2.size(2) != e1.size(2) or d2.size(3) != e1.size(3):\n",
1762
- " # # Resize e1 to match d2's dimensions\n",
1763
- " # e1 = F.interpolate(e1, size=(d2.size(2), d2.size(3)), mode='nearest')\n",
1764
  " d2 = torch.cat([d2, e1], dim=1) # Concatenate with encoder features\n",
1765
  " d2 = checkpoint(self.dec2, d2)\n",
1766
  " \n",
@@ -1818,7 +1956,6 @@
1818
  " def forward(self, x):\n",
1819
  " return self.unet(x)\n",
1820
  "\n",
1821
- "\n",
1822
  "class Denoiser(nn.Module):\n",
1823
  " def __init__(self, in_channels=1, out_channels=1):\n",
1824
  " super().__init__()\n",
@@ -1903,8 +2040,8 @@
1903
  " avg_reconstructor_train_loss = reconstructor_total_loss / len(train_dataloader)\n",
1904
  " avg_denoiser_train_loss = denoiser_total_loss / len(train_dataloader)\n",
1905
  " \n",
1906
- " avg_reconstructor_val_loss, _ = calculate_loss_and_psnr(reconstructor, val_dataloader, reconstructor_criterion)\n",
1907
- " avg_denoiser_val_loss, _ = calculate_loss_and_psnr(denoiser, val_dataloader, denoiser_criterion)\n",
1908
  " \n",
1909
  " print(f\"Epoch [{epoch+1}/{epochs}] - \"\n",
1910
  " f\"Reconstructor Train Loss: {avg_reconstructor_train_loss:.4f}, \"\n",
@@ -1916,20 +2053,154 @@
1916
  " if avg_reconstructor_val_loss < best_reconstructor_val_loss:\n",
1917
  " best_reconstructor_val_loss = avg_reconstructor_val_loss\n",
1918
  " torch.save(reconstructor.state_dict(), best_reconstructor_model_path)\n",
1919
- " print(f\"Reconstructor model saved with improved validation loss: {avg_reconstructor_val_loss:.4f}\")\n",
1920
  " \n",
1921
  " if avg_denoiser_val_loss < best_denoiser_val_loss:\n",
1922
  " best_denoiser_val_loss = avg_denoiser_val_loss\n",
1923
  " torch.save(denoiser.state_dict(), best_denoiser_model_path)\n",
1924
- " print(f\"Denoiser model saved with improved validation loss: {avg_denoiser_val_loss:.4f}\")\n",
1925
  " \n",
1926
  " return reconstructor, denoiser\n",
1927
  "\n",
1928
  "# Example usage with train and validation directories\n",
1929
  "reconstructor_model, denoiser_model = train_reconstructor_and_denoiser(\n",
1930
- " r\"D:\\TCIA_Split\\train\", r\"D:\\TCIA_Split\\val\", epochs=50, batch_size=1, grad_accumulation_steps=64\n",
1931
  ")"
1932
  ]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1933
  }
1934
  ],
1935
  "metadata": {
@@ -1948,7 +2219,7 @@
1948
  "name": "python",
1949
  "nbconvert_exporter": "python",
1950
  "pygments_lexer": "ipython3",
1951
- "version": "3.10.11"
1952
  }
1953
  },
1954
  "nbformat": 4,
 
261
  " dcm = pydicom.dcmread(full_path)\n",
262
  " \n",
263
  " # Check image dimensions\n",
264
+ " if dcm.pixel_array.shape[0] < min_size or dcm.pixel_array.shape[1] < min_size or dcm.pixel_array.shape[0] > min_size or dcm.pixel_array.shape[1] > min_size:\n",
265
  " files_to_delete.append(full_path)\n",
266
  " filtered_images += 1\n",
267
  " \n",
 
279
  " print(f\"Images deleted: {filtered_images}\\n\")\n",
280
  "\n",
281
  "# Usage\n",
282
+ "input_dirs = [\"./TCIA_Split/train\", \"./TCIA_Split/val\"]\n",
283
  "filter_dicom_images(input_dirs)"
284
  ]
285
  },
 
928
  },
929
  {
930
  "cell_type": "code",
931
+ "execution_count": 18,
932
  "metadata": {},
933
  "outputs": [],
934
  "source": [
 
1347
  "# Example usage\n",
1348
  "if __name__ == \"__main__\":\n",
1349
  " # Paths to models and test image\n",
1350
+ " RECONSTRUCTOR_MODEL_PATH = \"./small_reconstructor.pth\" # Path to your saved Reconstructor model\n",
1351
+ " DENOISER_MODEL_PATH = \"./small_denoiser.pth\" # Path to your saved Denoiser model\n",
1352
+ " TEST_DICOM_PATH = \"./test.dcm\" # Replace with actual path to test DICOM \n",
1353
  " # Run inference\n",
1354
  " inference_single_image(RECONSTRUCTOR_MODEL_PATH, DENOISER_MODEL_PATH, TEST_DICOM_PATH)"
1355
  ]
 
1644
  "cell_type": "markdown",
1645
  "metadata": {},
1646
  "source": [
1647
+ "### mediumRD Inference"
1648
+ ]
1649
+ },
1650
+ {
1651
+ "cell_type": "code",
1652
+ "execution_count": null,
1653
+ "metadata": {},
1654
+ "outputs": [],
1655
+ "source": [
1656
+ "import torch\n",
1657
+ "import pydicom\n",
1658
+ "import numpy as np\n",
1659
+ "import matplotlib.pyplot as plt\n",
1660
+ "import os\n",
1661
+ "\n",
1662
+ "# Import the models from the previous script\n",
1663
+ "# Assuming they are defined or imported correctly\n",
1664
+ "\n",
1665
+ "def load_dicom_image(dicom_path):\n",
1666
+ " \"\"\"\n",
1667
+ " Load and normalize a DICOM image\n",
1668
+ " \n",
1669
+ " Args:\n",
1670
+ " dicom_path (str): Path to the DICOM file\n",
1671
+ " \n",
1672
+ " Returns:\n",
1673
+ " torch.Tensor: Normalized image tensor\n",
1674
+ " \"\"\"\n",
1675
+ " # Read DICOM file\n",
1676
+ " dcm = pydicom.dcmread(dicom_path)\n",
1677
+ " image = dcm.pixel_array.astype(float)\n",
1678
+ " \n",
1679
+ " # Normalize image\n",
1680
+ " image = (image - image.min()) / (image.max() - image.min())\n",
1681
+ " \n",
1682
+ " # Convert to tensor\n",
1683
+ " image_tensor = torch.from_numpy(image).float().unsqueeze(0).unsqueeze(0) # Add batch and channel dimensions\n",
1684
+ " return image_tensor\n",
1685
+ "\n",
1686
+ "def calculate_psnr(output, target, max_pixel=1.0):\n",
1687
+ " \"\"\"\n",
1688
+ " Calculate Peak Signal-to-Noise Ratio (PSNR)\n",
1689
+ " \n",
1690
+ " Args:\n",
1691
+ " output (torch.Tensor): Reconstructed image\n",
1692
+ " target (torch.Tensor): Original image\n",
1693
+ " max_pixel (float): Maximum pixel value\n",
1694
+ " \n",
1695
+ " Returns:\n",
1696
+ " float: PSNR value\n",
1697
+ " \"\"\"\n",
1698
+ " # Ensure the values are in the correct range\n",
1699
+ " mse = torch.nn.functional.mse_loss(output, target)\n",
1700
+ " psnr = 20 * torch.log10(max_pixel / torch.sqrt(mse))\n",
1701
+ " return psnr.item()\n",
1702
+ "\n",
1703
+ "def visualize_reconstruction(original_image, reconstructed_image, psnr):\n",
1704
+ " \"\"\"\n",
1705
+ " Visualize original and reconstructed images\n",
1706
+ " \n",
1707
+ " Args:\n",
1708
+ " original_image (torch.Tensor): Original image tensor\n",
1709
+ " reconstructed_image (torch.Tensor): Reconstructed image tensor\n",
1710
+ " psnr (float): Peak Signal-to-Noise Ratio\n",
1711
+ " \"\"\"\n",
1712
+ " # Convert tensors to numpy for visualization\n",
1713
+ " original = original_image.squeeze().cpu().numpy()\n",
1714
+ " reconstructed = reconstructed_image.squeeze().cpu().numpy()\n",
1715
+ " \n",
1716
+ " # Create subplot\n",
1717
+ " fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))\n",
1718
+ " \n",
1719
+ " # Plot original image\n",
1720
+ " im1 = ax1.imshow(original, cmap='gray')\n",
1721
+ " ax1.set_title('Original Image')\n",
1722
+ " plt.colorbar(im1, ax=ax1)\n",
1723
+ " \n",
1724
+ " # Plot reconstructed image\n",
1725
+ " im2 = ax2.imshow(reconstructed, cmap='gray')\n",
1726
+ " ax2.set_title(f'Reconstructed Image\\nPSNR: {psnr:.2f} dB')\n",
1727
+ " plt.colorbar(im2, ax=ax2)\n",
1728
+ " \n",
1729
+ " plt.tight_layout()\n",
1730
+ " plt.show()\n",
1731
+ "\n",
1732
+ "def inference_single_image(reconstructor_model_path, denoiser_model_path, test_dicom_path):\n",
1733
+ " \"\"\"\n",
1734
+ " Perform inference on a single DICOM image using both Reconstructor and Denoiser models.\n",
1735
+ " \n",
1736
+ " Args:\n",
1737
+ " reconstructor_model_path (str): Path to the saved Reconstructor model weights\n",
1738
+ " denoiser_model_path (str): Path to the saved Denoiser model weights\n",
1739
+ " test_dicom_path (str): Path to the test DICOM file\n",
1740
+ " \"\"\"\n",
1741
+ " # Set device\n",
1742
+ " device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
1743
+ " \n",
1744
+ " # Initialize models\n",
1745
+ " reconstructor = Reconstructor().to(device)\n",
1746
+ " denoiser = Denoiser().to(device)\n",
1747
+ " \n",
1748
+ " # Load saved model weights\n",
1749
+ " reconstructor.load_state_dict(torch.load(reconstructor_model_path))\n",
1750
+ " denoiser.load_state_dict(torch.load(denoiser_model_path))\n",
1751
+ " \n",
1752
+ " reconstructor.eval()\n",
1753
+ " denoiser.eval()\n",
1754
+ " \n",
1755
+ " # Load and preprocess test image\n",
1756
+ " with torch.no_grad():\n",
1757
+ " test_image = load_dicom_image(test_dicom_path).to(device)\n",
1758
+ " \n",
1759
+ " # Perform reconstruction\n",
1760
+ " reconstructed_image = reconstructor(test_image)\n",
1761
+ " \n",
1762
+ " # Perform denoising on the reconstructed image\n",
1763
+ " denoised_image = denoiser(reconstructed_image)\n",
1764
+ " \n",
1765
+ " # Calculate PSNR for both original and denoised outputs\n",
1766
+ " psnr_reconstructed = calculate_psnr(reconstructed_image, test_image)\n",
1767
+ " psnr_denoised = calculate_psnr(denoised_image, test_image)\n",
1768
+ "\n",
1769
+ " print(f\"PSNR (Reconstructed): {psnr_reconstructed:.2f} dB\")\n",
1770
+ " print(f\"PSNR (Denoised): {psnr_denoised:.2f} dB\")\n",
1771
+ " \n",
1772
+ " # Visualize results\n",
1773
+ " visualize_reconstruction(test_image, reconstructed_image, psnr_reconstructed)\n",
1774
+ " visualize_reconstruction(test_image, denoised_image, psnr_denoised)\n",
1775
+ "\n",
1776
+ "# Example usage\n",
1777
+ "if __name__ == \"__main__\":\n",
1778
+ " # Paths to models and test image\n",
1779
+ " RECONSTRUCTOR_MODEL_PATH = \"./medium_reconstructor.pth\" # Path to your saved Reconstructor model\n",
1780
+ " DENOISER_MODEL_PATH = \"./medium_denoiser.pth\" # Path to your saved Denoiser model\n",
1781
+ " TEST_DICOM_PATH = \"./test.dcm\" # Replace with actual path to test DICOM \n",
1782
+ " # Run inference\n",
1783
+ " inference_single_image(RECONSTRUCTOR_MODEL_PATH, DENOISER_MODEL_PATH, TEST_DICOM_PATH)"
1784
+ ]
1785
+ },
1786
+ {
1787
+ "cell_type": "markdown",
1788
+ "metadata": {},
1789
+ "source": [
1790
+ "### Larger Reconstructor U-Net (largeR)"
1791
  ]
1792
  },
1793
  {
 
1834
  " self.bn2 = nn.BatchNorm2d(out_channels)\n",
1835
  " self.conv3 = nn.Conv2d(out_channels, out_channels, 3, padding=1)\n",
1836
  " self.bn3 = nn.BatchNorm2d(out_channels)\n",
1837
+ " self.conv4 = nn.Conv2d(out_channels, out_channels, 3, padding=1)\n",
1838
+ " self.bn4 = nn.BatchNorm2d(out_channels)\n",
1839
  " \n",
1840
  " def forward(self, x):\n",
1841
  " x = F.relu(self.bn1(self.conv1(x)))\n",
1842
  " x = F.relu(self.bn2(self.conv2(x)))\n",
1843
  " x = F.relu(self.bn3(self.conv3(x)))\n",
1844
+ " x = F.relu(self.bn4(self.conv4(x)))\n",
1845
  " return x\n",
1846
  "\n",
1847
  "class UNet(nn.Module):\n",
1848
  " def __init__(self, in_channels=1, out_channels=1):\n",
1849
  " super().__init__()\n",
1850
  " # Encoder\n",
1851
+ " self.enc1 = UNetBlock(in_channels, 64)\n",
1852
+ " self.enc2 = UNetBlock(64, 128)\n",
1853
+ " self.enc3 = UNetBlock(128, 256)\n",
1854
+ " self.enc4 = UNetBlock(256, 512)\n",
1855
+ " self.enc5 = UNetBlock(512, 1024)\n",
1856
  " \n",
1857
  " # Decoder with learned upsampling (transposed convolutions)\n",
1858
  "\n",
1859
+ " self.upconv5 = nn.ConvTranspose2d(1024, 512, kernel_size=2, stride=2) # Learnable upsampling\n",
1860
+ " self.dec5 = UNetBlock(512 + 512, 512) # Adjust input channels after concatenation\n",
1861
  "\n",
1862
+ " self.upconv4 = nn.ConvTranspose2d(512, 256, kernel_size=2, stride=2) # Learnable upsampling\n",
1863
+ " self.dec4 = UNetBlock(256 + 256, 256) # Adjust input channels after concatenation\n",
1864
  "\n",
1865
+ " self.upconv3 = nn.ConvTranspose2d(256, 128, kernel_size=2, stride=2) # Learnable upsampling\n",
1866
+ " self.dec3 = UNetBlock(128 + 128, 128) # Adjust input channels after concatenation\n",
1867
  "\n",
1868
+ " self.upconv2 = nn.ConvTranspose2d(128, 64, kernel_size=2, stride=2) # Learnable upsampling\n",
1869
+ " self.dec2 = UNetBlock(64 + 64, 64) # Adjust input channels after concatenation\n",
1870
  "\n",
1871
+ " self.dec1 = UNetBlock(64, out_channels) # Final output\n",
1872
  "\n",
1873
  " self.pool = nn.MaxPool2d(2, 2)\n",
1874
  " \n",
 
1883
  " # Decoder path with learned upsampling and skip connections\n",
1884
  "\n",
1885
  " d5 = self.upconv5(e5) # Learnable upsampling\n",
1886
+ "\n",
 
 
1887
  " d5 = torch.cat([d5, e4], dim=1) # Concatenate with encoder features\n",
1888
  " d5 = checkpoint(self.dec5, d5)\n",
1889
  "\n",
1890
  " d4 = self.upconv4(d5) # Learnable upsampling\n",
1891
+ "\n",
 
 
1892
  " d4 = torch.cat([d4, e3], dim=1) # Concatenate with encoder features\n",
1893
  " d4 = checkpoint(self.dec4, d4)\n",
1894
  "\n",
1895
  " d3 = self.upconv3(d4) # Learnable upsampling\n",
1896
+ "\n",
 
 
1897
  " d3 = torch.cat([d3, e2], dim=1) # Concatenate with encoder features\n",
1898
  " d3 = checkpoint(self.dec3, d3)\n",
1899
  "\n",
1900
  " d2 = self.upconv2(d3) # Learnable upsampling\n",
1901
+ "\n",
 
 
1902
  " d2 = torch.cat([d2, e1], dim=1) # Concatenate with encoder features\n",
1903
  " d2 = checkpoint(self.dec2, d2)\n",
1904
  " \n",
 
1956
  " def forward(self, x):\n",
1957
  " return self.unet(x)\n",
1958
  "\n",
 
1959
  "class Denoiser(nn.Module):\n",
1960
  " def __init__(self, in_channels=1, out_channels=1):\n",
1961
  " super().__init__()\n",
 
2040
  " avg_reconstructor_train_loss = reconstructor_total_loss / len(train_dataloader)\n",
2041
  " avg_denoiser_train_loss = denoiser_total_loss / len(train_dataloader)\n",
2042
  " \n",
2043
+ " avg_reconstructor_val_loss, avg_reconstructor_val_psnr = calculate_loss_and_psnr(reconstructor, val_dataloader, reconstructor_criterion)\n",
2044
+ " avg_denoiser_val_loss, avg_denoiser_val_psnr = calculate_loss_and_psnr(denoiser, val_dataloader, denoiser_criterion)\n",
2045
  " \n",
2046
  " print(f\"Epoch [{epoch+1}/{epochs}] - \"\n",
2047
  " f\"Reconstructor Train Loss: {avg_reconstructor_train_loss:.4f}, \"\n",
 
2053
  " if avg_reconstructor_val_loss < best_reconstructor_val_loss:\n",
2054
  " best_reconstructor_val_loss = avg_reconstructor_val_loss\n",
2055
  " torch.save(reconstructor.state_dict(), best_reconstructor_model_path)\n",
2056
+ " print(f\"Reconstructor model saved with improved validation loss: {avg_reconstructor_val_loss:.4f} and PSNR: {avg_reconstructor_val_psnr}\")\n",
2057
  " \n",
2058
  " if avg_denoiser_val_loss < best_denoiser_val_loss:\n",
2059
  " best_denoiser_val_loss = avg_denoiser_val_loss\n",
2060
  " torch.save(denoiser.state_dict(), best_denoiser_model_path)\n",
2061
+ " print(f\"Denoiser model saved with improved validation loss: {avg_denoiser_val_loss:.4f} and PSNR: {avg_denoiser_val_psnr}\")\n",
2062
  " \n",
2063
  " return reconstructor, denoiser\n",
2064
  "\n",
2065
  "# Example usage with train and validation directories\n",
2066
  "reconstructor_model, denoiser_model = train_reconstructor_and_denoiser(\n",
2067
+ " \"./TCIA_Split/train\", \"./TCIA_Split/val\", epochs=50, batch_size=8, grad_accumulation_steps=8\n",
2068
  ")"
2069
  ]
2070
+ },
2071
+ {
2072
+ "cell_type": "markdown",
2073
+ "metadata": {},
2074
+ "source": [
2075
+ "### largeR Inference"
2076
+ ]
2077
+ },
2078
+ {
2079
+ "cell_type": "code",
2080
+ "execution_count": null,
2081
+ "metadata": {},
2082
+ "outputs": [],
2083
+ "source": [
2084
+ "import torch\n",
2085
+ "import pydicom\n",
2086
+ "import numpy as np\n",
2087
+ "import matplotlib.pyplot as plt\n",
2088
+ "import os\n",
2089
+ "\n",
2090
+ "# Import the models from the previous script\n",
2091
+ "# Assuming they are defined or imported correctly\n",
2092
+ "\n",
2093
+ "def load_dicom_image(dicom_path):\n",
2094
+ " \"\"\"\n",
2095
+ " Load and normalize a DICOM image\n",
2096
+ " \n",
2097
+ " Args:\n",
2098
+ " dicom_path (str): Path to the DICOM file\n",
2099
+ " \n",
2100
+ " Returns:\n",
2101
+ " torch.Tensor: Normalized image tensor\n",
2102
+ " \"\"\"\n",
2103
+ " # Read DICOM file\n",
2104
+ " dcm = pydicom.dcmread(dicom_path)\n",
2105
+ " image = dcm.pixel_array.astype(float)\n",
2106
+ " \n",
2107
+ " # Normalize image\n",
2108
+ " image = (image - image.min()) / (image.max() - image.min())\n",
2109
+ " \n",
2110
+ " # Convert to tensor\n",
2111
+ " image_tensor = torch.from_numpy(image).float().unsqueeze(0).unsqueeze(0) # Add batch and channel dimensions\n",
2112
+ " return image_tensor\n",
2113
+ "\n",
2114
+ "def calculate_psnr(output, target, max_pixel=1.0):\n",
2115
+ " \"\"\"\n",
2116
+ " Calculate Peak Signal-to-Noise Ratio (PSNR)\n",
2117
+ " \n",
2118
+ " Args:\n",
2119
+ " output (torch.Tensor): Reconstructed image\n",
2120
+ " target (torch.Tensor): Original image\n",
2121
+ " max_pixel (float): Maximum pixel value\n",
2122
+ " \n",
2123
+ " Returns:\n",
2124
+ " float: PSNR value\n",
2125
+ " \"\"\"\n",
2126
+ " # Ensure the values are in the correct range\n",
2127
+ " mse = torch.nn.functional.mse_loss(output, target)\n",
2128
+ " psnr = 20 * torch.log10(max_pixel / torch.sqrt(mse))\n",
2129
+ " return psnr.item()\n",
2130
+ "\n",
2131
+ "def visualize_reconstruction(original_image, reconstructed_image, psnr):\n",
2132
+ " \"\"\"\n",
2133
+ " Visualize original and reconstructed images\n",
2134
+ " \n",
2135
+ " Args:\n",
2136
+ " original_image (torch.Tensor): Original image tensor\n",
2137
+ " reconstructed_image (torch.Tensor): Reconstructed image tensor\n",
2138
+ " psnr (float): Peak Signal-to-Noise Ratio\n",
2139
+ " \"\"\"\n",
2140
+ " # Convert tensors to numpy for visualization\n",
2141
+ " original = original_image.squeeze().cpu().numpy()\n",
2142
+ " reconstructed = reconstructed_image.squeeze().cpu().numpy()\n",
2143
+ " \n",
2144
+ " # Create subplot\n",
2145
+ " fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))\n",
2146
+ " \n",
2147
+ " # Plot original image\n",
2148
+ " im1 = ax1.imshow(original, cmap='gray')\n",
2149
+ " ax1.set_title('Original Image')\n",
2150
+ " plt.colorbar(im1, ax=ax1)\n",
2151
+ " \n",
2152
+ " # Plot reconstructed image\n",
2153
+ " im2 = ax2.imshow(reconstructed, cmap='gray')\n",
2154
+ " ax2.set_title(f'Reconstructed Image\\nPSNR: {psnr:.2f} dB')\n",
2155
+ " plt.colorbar(im2, ax=ax2)\n",
2156
+ " \n",
2157
+ " plt.tight_layout()\n",
2158
+ " plt.show()\n",
2159
+ "\n",
2160
+ "def inference_single_image(reconstructor_model_path, test_dicom_path):\n",
2161
+ " \"\"\"\n",
2162
+ " Perform inference on a single DICOM image using both Reconstructor and Denoiser models.\n",
2163
+ " \n",
2164
+ " Args:\n",
2165
+ " reconstructor_model_path (str): Path to the saved Reconstructor model weights\n",
2166
+ " denoiser_model_path (str): Path to the saved Denoiser model weights\n",
2167
+ " test_dicom_path (str): Path to the test DICOM file\n",
2168
+ " \"\"\"\n",
2169
+ " # Set device\n",
2170
+ " device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
2171
+ " \n",
2172
+ " # Initialize models\n",
2173
+ " reconstructor = Reconstructor().to(device)\n",
2174
+ " \n",
2175
+ " # Load saved model weights\n",
2176
+ " reconstructor.load_state_dict(torch.load(reconstructor_model_path))\n",
2177
+ " \n",
2178
+ " reconstructor.eval()\n",
2179
+ " \n",
2180
+ " # Load and preprocess test image\n",
2181
+ " with torch.no_grad():\n",
2182
+ " test_image = load_dicom_image(test_dicom_path).to(device)\n",
2183
+ " \n",
2184
+ " # Perform reconstruction\n",
2185
+ " reconstructed_image = reconstructor(test_image)\n",
2186
+ " \n",
2187
+ " \n",
2188
+ " # Calculate PSNR for both original and denoised outputs\n",
2189
+ " psnr_reconstructed = calculate_psnr(reconstructed_image, test_image)\n",
2190
+ "\n",
2191
+ " print(f\"PSNR (Reconstructed): {psnr_reconstructed:.2f} dB\")\n",
2192
+ " \n",
2193
+ " # Visualize results\n",
2194
+ " visualize_reconstruction(test_image, reconstructed_image, psnr_reconstructed)\n",
2195
+ "\n",
2196
+ "# Example usage\n",
2197
+ "if __name__ == \"__main__\":\n",
2198
+ " # Paths to models and test image\n",
2199
+ " RECONSTRUCTOR_MODEL_PATH = \"./large_reconstructor.pth\" # Path to your saved Reconstructor model\n",
2200
+ " TEST_DICOM_PATH = \"./test.dcm\" # Replace with actual path to test DICOM \n",
2201
+ " # Run inference\n",
2202
+ " inference_single_image(RECONSTRUCTOR_MODEL_PATH, TEST_DICOM_PATH)"
2203
+ ]
2204
  }
2205
  ],
2206
  "metadata": {
 
2219
  "name": "python",
2220
  "nbconvert_exporter": "python",
2221
  "pygments_lexer": "ipython3",
2222
+ "version": "3.12.3"
2223
  }
2224
  },
2225
  "nbformat": 4,