Initial commit

merics_code/MS-SSIM_MAE_PSNR.py

+import os
+import cv2
+import tensorflow as tf
+import pandas as pd
+import numpy as np
+
+#Vanilla CVDM
+folder_a = r'path to\images\VanillaCVDM_inference'
+folder_b = r'path to\images\BioSR_high_resolution_images'
+
+
+#CVDM+N
+# folder_a=r'path to\images\naturalized_BioSR_high_resolution_images'
+# folder_b=r'path to\images\CVDM+N_inference'
+
+#NCVDM
+# folder_a=r'path to\images\naturalized_BioSR_high_resolution_images'
+# folder_b=r'path to\images\NCVDM_inference'
+
+
+metrics = []
+
+# Loop through the images in the first folder
+for filename in os.listdir(folder_a):
+
+    img_path_a = os.path.join(folder_a, filename)
+    img_path_b = os.path.join(folder_b, filename)
+
+    if os.path.isfile(img_path_b):
+        # Read the grayscale TIFF images
+        img_a = cv2.imread(img_path_a, cv2.IMREAD_UNCHANGED)
+        img_b = cv2.imread(img_path_b, cv2.IMREAD_UNCHANGED)
+        if img_a is None:
+            print(f"Error: Could not load {img_path_a}")
+            continue
+        if img_b is None:
+            print(f"Error: Could not load {img_path_b}")
+            continue
+
+        img_a = img_a.astype('float32')
+        img_b = img_b.astype('float32')
+
+        # Min-Max normalization
+        img_a_min, img_a_max = img_a.min(), img_a.max()
+        img_b_min, img_b_max = img_b.min(), img_b.max()
+         #normalize image to [0, 255] 
+        img_a = ((img_a - img_a_min) / (img_a_max - img_a_min)) * 255  
+        img_b = ((img_b - img_b_min) / (img_b_max - img_b_min)) * 255
+        # Adds a channel dimension
+        img_a = tf.expand_dims(img_a, axis=-1)  
+        img_b = tf.expand_dims(img_b, axis=-1)  
+
+        # Calculate MS-SSIM, PSNR and MAE
+        msssim_value = tf.image.ssim_multiscale(img_a, img_b, max_val=255.0).numpy()
+        psnr_value = tf.image.psnr(img_a, img_b, max_val=255).numpy()
+        mae_value = np.mean(np.abs(img_a - img_b))
+        # Append the result to the list
+        metrics.append({
+            "filename": filename,
+            "MS-SSIM": msssim_value,
+            "MAE": mae_value,
+            "PSNR": psnr_value
+        })
+
+# Create a DataFrame from the results
+results_df = pd.DataFrame(metrics)
+
+# Save results to a CSV file
+results_df.to_csv(r"path to\CVDM+N.csv", index=False)

merics_code/Pearson_correlation.py

+import pandas as pd
+
+# Load the Excel file
+file_path = r"path to\VanillaCVDM_all_metrics.xlsx"  # Replace with your file path 
+df = pd.read_excel(file_path, header=None)  
+
+# Set the first row as column headers
+df.columns = df.iloc[0]  
+df = df[1:]  # Remove the first row from the data
+
+# # Debugging: Check the processed data
+# print("\nProcessed DataFrame:")
+# print(df.head())
+
+# Calculate the Pearson correlation coefficient matrix
+correlation_matrix = df.corr(method='pearson').round(3)
+
+# Display the correlation matrix
+print("\nPearson Correlation Coefficient Matrix:")
+print(correlation_matrix)
+
+# Optionally, save the correlation matrix to a CSV file
+#correlation_matrix.to_csv(r"path to\VanillaCVDM_correlation_matrix.csv", index=True)

merics_code/RMS contrast.py

+#RMS contrast
+import os
+import cv2
+import numpy as np
+import pandas as pd
+
+# Function to normalize image to [0, 255] using Min-Max normalization
+def min_max_normalize(image):
+    return ((image - np.min(image)) / (np.max(image) - np.min(image)))*255
+
+# Function to calculate RMS contrast
+def calculate_rms_contrast(image):
+
+    # Calculate the mean intensity
+    mean_intensity = np.mean(image)
+
+    # Calculate RMS contrast
+    rms_contrast = np.sqrt(np.mean((image - mean_intensity) ** 2))
+
+    return rms_contrast
+
+
+#Vanilla CVDM
+image_folder = r'path to\images\VanillaCVDM_inference'  # Replace with your folder path
+output_csv = r"path to \RMS_contrast_CVDM.csv"
+
+#CVDM+N
+# image_folder = =r'path to\images\CVDM+N_inference'  # Replace with your folder path
+# output_csv = r"path to\RMS_contrast_CVDM+N.csv"
+
+#NCVDM
+# image_folder = r'path to\images\NCVDM_inference'  # Replace with your folder path
+# output_csv = r"path to \RMS_contrast_NCVDM.csv"
+
+
+
+
+# List to store results
+results = []
+
+# Loop through all files in the folder
+for filename in os.listdir(image_folder):
+    file_path = os.path.join(image_folder, filename)
+
+    # Check if the file is an image (optional: add specific extensions if needed)
+    if filename.lower().endswith(( ".tif", ".tiff")):
+        try:
+            # Read the image
+            image = cv2.imread(file_path,cv2.IMREAD_UNCHANGED)
+            image=image.astype('float32')
+            image=min_max_normalize(image)
+            # Ensure the image was loaded successfully
+            if image is not None:
+                # Calculate RMS contrast
+                rms_contrast = calculate_rms_contrast(image)
+
+                # Append the result to the list
+                results.append({"filename": filename, "RMS Contrast": rms_contrast})
+            else:
+                print(f"Could not read image: {filename}")
+        except Exception as e:
+            print(f"Error processing file {filename}: {e}")
+
+# Save results to a CSV file
+if results:
+    df = pd.DataFrame(results)
+    df.to_csv(output_csv, index=False)
+    print(f"RMS contrast results saved to {output_csv}")
+else:
+    print("No valid images found in the folder.")