New crop functionality

app.py

@@ -73,7 +73,7 @@ def build_model(hypar,device):
 def crop_signature(original_image_path, mask, padding=32):
     """
     Crop the signature from the original image using the provided mask.
-    
+
     :param original_image_path: The file path of the original image.
     :param mask: The binary mask of the signature.
     :param padding: Padding to add around the bounding box of the signature.
@@ -82,34 +82,34 @@ def crop_signature(original_image_path, mask, padding=32):
     # Convert the mask to a binary image
     _, binary_mask = cv2.threshold(mask, 127, 255, cv2.THRESH_BINARY)
 
-    # Optionally, dilate the mask here if needed
-    kernel = np.ones((5,5), np.uint8)
-    dilated_mask = cv2.dilate(binary_mask, kernel, iterations=1)
+    # Find contours from the binary mask
+    contours, _ = cv2.findContours(binary_mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
 
-    # Find contours from the binary (possibly dilated) mask
-    contours, _ = cv2.findContours(dilated_mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
-    
     # Open the original image
     original_image = Image.open(original_image_path).convert("RGB")
-    
+
     # If contours are found, proceed to crop
     if contours:
-        # Assume the largest contour is the signature
-        x, y, w, h = cv2.boundingRect(max(contours, key=cv2.contourArea))
+        # Find the combined bounding box of all contours
+        min_x, min_y = original_image.width, original_image.height
+        max_x = max_y = 0
 
-        # Here you can manually adjust the y coordinate to include more space above
-        y_adjusted = max(y - padding - 10, 0)  # Adjust 10 pixels more if needed
+        for contour in contours:
+            x, y, w, h = cv2.boundingRect(contour)
+            min_x, min_y = min(min_x, x), min(min_y, y)
+            max_x, max_y = max(max_x, x + w), max(max_y, y + h)
 
         # Add padding to the bounding box
-        x_padded = max(x - padding, 0)
-        w_padded = min(w + 2 * padding, original_image.width - x_padded)
-        h_padded = min(h + padding + (y - y_adjusted), original_image.height - y_adjusted)  # Include the adjustment
+        x_padded = max(min_x - padding, 0)
+        y_padded = max(min_y - padding, 0)
+        w_padded = min(max_x + padding, original_image.width) - x_padded
+        h_padded = min(max_y + padding, original_image.height) - y_padded
+
+        # Crop the original image using the combined bounding box with padding
+        cropped_image = original_image.crop((x_padded, y_padded, x_padded + w_padded, y_padded + h_padded))
 
-        # Crop the original image and mask using the bounding box with padding
-        cropped_image = original_image.crop((x_padded, y_adjusted, x_padded + w_padded, y_adjusted + h_padded))
-        
         return cropped_image
-    
+
     # If no contours are found, return the original image
     return original_image