From fe96aa91af94c2c8b650ae8d3215a3b497af7469 Mon Sep 17 00:00:00 2001
From: Otto Brinkhaus <brinkhaus@gmail.com>
Date: Thu, 19 Oct 2023 16:57:10 +0200
Subject: [PATCH] fix: don't use pixels from exclusion mask as seed pixels

---
 decimer_segmentation/complete_structure.py | 33 ++++++++++++++--------
 1 file changed, 22 insertions(+), 11 deletions(-)

diff --git a/decimer_segmentation/complete_structure.py b/decimer_segmentation/complete_structure.py
index ef5bcea..4a593fd 100644
--- a/decimer_segmentation/complete_structure.py
+++ b/decimer_segmentation/complete_structure.py
@@ -286,7 +286,11 @@ def get_mask_center(mask_array: np.array) -> Tuple[int, int]:
             return None, None
 
 
-def get_seeds(image_array: np.array, mask_array: np.array) -> List[Tuple[int, int]]:
+def get_seeds(
+    image_array: np.array,
+    mask_array: np.array,
+    exclusion_mask: np.array,
+) -> List[Tuple[int, int]]:
     """
     This function takes an array that represents an image and a mask.
     It returns a list of tuples with indices of seeds in the structure
@@ -295,6 +299,7 @@ def get_seeds(image_array: np.array, mask_array: np.array) -> List[Tuple[int, in
     Args:
         image_array (np.array): Image
         mask_array (np.array): Mask
+        exclusion_mask (np.array): Exclusion mask
 
     Returns:
         List[Tuple[int, int]]: [(x,y), (x,y), ...]
@@ -312,32 +317,36 @@ def get_seeds(image_array: np.array, mask_array: np.array) -> List[Tuple[int, in
                 if not mask_array[y_center, x_center + n]:
                     up = False
                 if not image_array[y_center, x_center + n]:
-                    seed_pixels.append((x_center + n, y_center))
-                    up = False
+                    if not exclusion_mask[y_center, x_center + n]:
+                        seed_pixels.append((x_center + n, y_center))
+                        up = False
         # Check for seeds below center
         if down:
             if x_center - n >= 0:
                 if not mask_array[y_center, x_center - n]:
                     down = False
                 if not image_array[y_center, x_center - n]:
-                    seed_pixels.append((x_center - n, y_center))
-                    down = False
+                    if not exclusion_mask[y_center, x_center - n]:
+                        seed_pixels.append((x_center - n, y_center))
+                        down = False
         # Check for seeds left from center
         if left:
             if y_center + n < image_array.shape[0]:
                 if not mask_array[y_center + n, x_center]:
                     left = False
                 if not image_array[y_center + n, x_center]:
-                    seed_pixels.append((x_center, y_center + n))
-                    left = False
+                    if not exclusion_mask[y_center + n, x_center]:
+                        seed_pixels.append((x_center, y_center + n))
+                        left = False
         # Check for seeds right from center
         if right:
             if y_center - n >= 0:
                 if not mask_array[y_center - n, x_center]:
                     right = False
                 if not image_array[y_center - n, x_center]:
-                    seed_pixels.append((x_center, y_center - n))
-                    right = False
+                    if not exclusion_mask[y_center - n, x_center]:
+                        seed_pixels.append((x_center, y_center - n))
+                        right = False
     return seed_pixels
 
 
@@ -450,7 +459,9 @@ def expansion_coordination(
     the mask expansion. It returns the expanded mask. The purpose of this function is
     wrapping up the expansion procedure in a map function.
     """
-    seed_pixels = get_seeds(image_array, mask_array)
+    seed_pixels = get_seeds(image_array,
+                            mask_array,
+                            exclusion_mask)
     if seed_pixels != []:
         mask_array = expand_masks(image_array, seed_pixels, mask_array, exclusion_mask)
     else:
@@ -498,7 +509,7 @@ def complete_structure_mask(
 
     if mask_array.size != 0:
         # Binarization of input image
-        binarized_image_array = binarize_image(image_array, threshold=0.85)
+        binarized_image_array = binarize_image(image_array, threshold=0.72)
         # Apply dilation with a resolution-dependent kernel to the image
         blur_factor = (
             int(image_array.shape[1] / 185) if image_array.shape[1] / 185 >= 2 else 2