Closer with contours

interpretter_notes.py

@@ -130,4 +130,13 @@ array([[1907,  887],
 """
 >>> cv.boundingRect(c[0])
 (399, 340, 5, 3)
-"""
+
+>>> get_coordinates_for_bb_simple(results["segmentation"], 1)
+((399, 300), (538, 392))
+>>> make_new_bounding_box(cv.boundingRect(c[0]), cv.boundingRect(c[1]))
+(399, 300, 140, 93)
+>>> cv.boundingRect(c[0])
+(399, 340, 5, 3)
+>>> cv.boundingRect(c[1])
+(409, 300, 130, 93)
+"""

understand.py

@@ -103,9 +103,9 @@ def get_coordinates_for_bb_simple(map_to_use, label_id):
     else:
         mask = (map_to_use.numpy() == label_id)
     
-    x, y = np.where(mask==True)
-    x_max, x_min = max(x), min(x)
-    y_max, y_min = max(y), min(y)
+    y_vals, x_vals = np.where(mask==True)
+    x_max, x_min = max(x_vals), min(x_vals)
+    y_max, y_min = max(y_vals), min(y_vals)
     return (x_min, y_min), (x_max, y_max)
 
 def make_simple_box(left_top, right_bottom, map_size):
@@ -123,7 +123,7 @@ def make_simple_box(left_top, right_bottom, map_size):
     plt.show()
 
 
-def test(map_to_use, label_id):
+def map_bounding_box_draw(map_to_use, label_id, img_obj=TEST_IMAGE):
     """
     map_to_use: You have to pass in `results["segmentation"]`
     """
@@ -137,14 +137,15 @@ def test(map_to_use, label_id):
     left_x, top_y = lt
     right_x, bottom_y = rb
     
-    mask[left_x:right_x, top_y] = .5
-    mask[left_x:right_x, bottom_y] = .5
-    mask[left_x, top_y:bottom_y] = .5
-    mask[right_x, top_y:bottom_y] = .5
+    mask[top_y, left_x:right_x] = .5
+    mask[bottom_y, left_x:right_x] = .5
+    mask[ top_y:bottom_y, left_x] = .5
+    mask[ top_y:bottom_y, right_x] = .5
 
     visual_mask = (mask* 255).astype(np.uint8)
     visual_mask = Image.fromarray(visual_mask)
-    plt.imshow(visual_mask)
+    plt.imshow(img_obj)
+    plt.imshow(visual_mask, alpha=0.25)
     plt.show()
 
 def contour_map(map_to_use, label_id):
@@ -166,4 +167,66 @@ def contour_map(map_to_use, label_id):
 
 # Idea for determining if close
 #   https://dsp.stackexchange.com/questions/2564/opencv-c-connect-nearby-contours-based-on-distance-between-them
-#   Bing Search: cv determine if 2 contours belong together
+#   Bing Search: cv determine if 2 contours belong together
+
+def find_if_close(contour1, contour2, c_dist=50):
+    """
+    Source: https://dsp.stackexchange.com/questions/2564/opencv-c-connect-nearby-contours-based-on-distance-between-them
+
+    """
+    row1, row2 = contour1.shape[0], contour2.shape[0]
+    for i in range(row1):
+        for j in range(row2):
+            dist = np.linalg.norm(contour1[i]-contour2[j])
+            if abs(dist) < c_dist:
+                return True
+            elif i == (row1-1) and j == (row2-1):
+                return False
+
+
+def make_new_bounding_box(bb1, bb2):
+    x1, y1, w1, h1 = bb1
+    x2, y2, w2, h2 = bb2
+    new_x = min(x1, x2)
+    new_y = min(y1, y2)
+    new_w = abs(max(x1+w1, x2+w2) - new_x)
+    new_h = abs(max(y1+h1, y2+h2) - new_y)
+
+    return (new_x, new_y, new_w, new_h)
+
+def map_bounding_box_draw(map_to_use, label_id, img_obj=TEST_IMAGE, v="cv"):
+    """
+    map_to_use: You have to pass in `results["segmentation"]`
+    v: version of bounding box
+        cv, coord
+    """
+    if torch.cuda.is_available():
+        mask = (map_to_use.cpu().numpy() == label_id)
+    else:
+        mask = (map_to_use.numpy() == label_id)
+    
+    if v == "cv":
+        c, v = contour_map(map_to_use, label_id)
+        x, y, w, h = make_new_bounding_box(cv.boundingRect(c[0]), cv.boundingRect(c[1]))
+        lt = (x, y)
+        rb = (x + w, y + h)
+        left_x, top_y = lt
+        right_x, bottom_y = rb
+    elif v == "coord":
+        lt, rb = get_coordinates_for_bb_simple(map_to_use, label_id)
+        left_x, top_y = lt
+        right_x, bottom_y = rb
+    else:
+        print(f"Not available `v` command {v}")
+        return
+    
+    mask[top_y, left_x:right_x] = .5
+    mask[bottom_y, left_x:right_x] = .5
+    mask[ top_y:bottom_y, left_x] = .5
+    mask[ top_y:bottom_y, right_x] = .5
+
+    visual_mask = (mask* 255).astype(np.uint8)
+    visual_mask = Image.fromarray(visual_mask)
+    plt.imshow(img_obj)
+    plt.imshow(visual_mask, alpha=0.25)
+    plt.show(block=False)