Bootstrap method working!

Author: feldnerd

.DS_Store

@@ -120,4 +120,100 @@ def get_right_min_image_combinations(number_image_pairs_selected, image_data, im
     image_combination_data = np.array(image_combination_data)
     image_combination_labels = np.array(image_combination_labels)
 
-    return image_combination_data, image_combination_labels
+    return image_combination_data, image_combination_labels
+
+
+# Randomly select a defined number of image paires from a given dataset
+# Always store the minimum image on the left
+def get_left_min_image_combinations_separate(number_image_pairs_selected, image_data, image_labels):
+    
+    num_images_total = len(image_labels)
+    
+    image_combination_data_A = []
+    image_combination_data_B = []
+    
+    image_combination_labels = []
+    
+    for i in range(number_image_pairs_selected):
+        
+        # Draw two image indices from a uniform random distribution
+        random_index_A = random.randint(0, num_images_total-1)
+        random_index_B = random.randint(0, num_images_total-1)
+        
+        # Randomly choose two images from the dataset
+        image_A = image_data[random_index_A]
+        image_B = image_data[random_index_B]
+        
+        # Find the minimum between the two labels
+        label_A = image_labels[random_index_A]
+        label_B = image_labels[random_index_B]
+        minimum_label = min(label_A, label_B)
+        
+        if label_A < label_B:
+            # Here we want to exclude the images where they are the same on both sides
+            image_combination_data_A.append(image_A)
+            image_combination_data_B.append(image_B)
+            
+            image_combination_labels.append(minimum_label)
+        elif label_A > label_B: 
+            # Here we want to exclude the images where they are the same on both sides
+            image_combination_data_A.append(image_B)
+            image_combination_data_B.append(image_A)
+            image_combination_labels.append(minimum_label)
+        
+    # Convert image data and labels lists to numpy arrays
+    image_combination_data_A = np.array(image_combination_data_A)
+    image_combination_data_B = np.array(image_combination_data_B)
+    
+    image_combination_labels = np.array(image_combination_labels)
+    
+    return image_combination_data_A, image_combination_data_B, image_combination_labels
+
+
+# Randomly select a defined number of image paires from a given dataset
+# Always store the minimum image on the left
+def get_right_min_image_combinations_separate(number_image_pairs_selected, image_data, image_labels):
+    
+    num_images_total = len(image_labels)
+    
+    image_combination_data_A = []
+    image_combination_data_B = []
+    
+    image_combination_labels = []
+    
+    for i in range(number_image_pairs_selected):
+        
+        # Draw two image indices from a uniform random distribution
+        random_index_A = random.randint(0, num_images_total-1)
+        random_index_B = random.randint(0, num_images_total-1)
+        
+        # Randomly choose two images from the dataset
+        image_A = image_data[random_index_A]
+        image_B = image_data[random_index_B]
+        
+        # Find the minimum between the two labels
+        label_A = image_labels[random_index_A]
+        label_B = image_labels[random_index_B]
+        minimum_label = min(label_A, label_B)
+        
+        if label_A > label_B:
+            # Here we want to exclude the images where they are the same on both sides
+            image_combination_data_A.append(image_A)
+            image_combination_data_B.append(image_B)
+            
+            image_combination_labels.append(minimum_label)
+            
+        elif label_A < label_B: 
+            # Here we want to exclude the images where they are the same on both sides
+            image_combination_data_A.append(image_B)
+            image_combination_data_B.append(image_A)
+            
+            image_combination_labels.append(minimum_label)
+        
+    # Convert image data and labels lists to numpy arrays
+    image_combination_data_A = np.array(image_combination_data_A)
+    image_combination_data_B = np.array(image_combination_data_B)
+    
+    image_combination_labels = np.array(image_combination_labels)
+    
+    return image_combination_data_A, image_combination_data_B, image_combination_labels