annotation tool added

Author: cccaaannn

.gitignore

@@ -1,5 +1,4 @@
 .vscode
-test.py
 dist
 imagepreprocessing.egg-info
 build

imagepreprocessing/__init__.py

@@ -1 +1 @@
-from .imagepreprocessing import create_training_data_keras, make_prediction_from_directory_keras, make_prediction_from_array_keras, create_training_data_yolo, create_cfg_file_yolo, make_prediction_from_directory_yolo, create_confusion_matrix, train_test_split
+from .imagepreprocessing import create_training_data_keras, make_prediction_from_directory_keras, make_prediction_from_array_keras, create_training_data_yolo, yolo_annotation_tool, create_cfg_file_yolo, make_prediction_from_directory_yolo, create_confusion_matrix, train_test_split

imagepreprocessing/convert_annotations.py

@@ -0,0 +1,70 @@
+
+def __convert_annotations_opencv_to_yolo(image_width, image_height, all_points):
+    """
+    0 x1 y1 x2 y2 -> 0 centerX centerY width height
+    converts opencv image coordinates to yolo style annotations label of the images has to be included
+    """
+    yolo_points = []
+    for points in all_points:
+        label = points[0]
+        x1 = points[1][0]
+        y1 = points[1][1]
+        x2 = points[2][0]
+        y2 = points[2][1]
+
+        w = abs(x1-x2)
+        h = abs(y1-y2)
+        
+        if(x1>x2):
+            cx = x1-w/2
+        else:
+            cx = x2-w/2
+        
+        if(y1>y2):
+            cy = y1 - h/2
+        else:
+            cy = y2 - h/2
+            
+        cx = cx/image_width
+        cy = cy/image_height
+        
+        w = abs(x1-x2)/image_width
+        h = abs(y1-y2)/image_height
+
+        yolo_points.append((label, cx,cy,w,h))
+
+    return yolo_points
+
+
+def __convert_annotations_yolo_to_opencv(image_width, image_height, all_points):
+    """
+    0 centerX centerY width height -> 0 x1 y1 x2 y2
+    converts yolo style annotations to opencv image coordinates so you can draw rectangles
+    it also returns label of the image as first element of the list
+    """
+
+    # most of the time this values will be reed from a file and an empty list can cause some problems so I used try catch 
+    try:
+        opencv_points = []
+        for points in all_points:
+            # percent to abs pixel positions
+            mx = image_width * float(points[1])
+            my = image_height * float(points[2])
+
+            w = image_width * float(points[3])
+            h = image_height * float(points[4])
+
+            # it should be like this in a regular coodinate system but on computer y coordinates are reversed 
+            # 0,0 is top left instead of bottom left
+            # soly = int(my + h/2) sagy = int(my - h/2) 
+
+            top_left_y = int(my - h/2)
+            top_left_x = int(mx - w/2)
+            bottom_rigth_y = int(my + h/2)
+            bottom_right_x = int(mx + w/2)
+            opencv_points.append((int(points[0]), top_left_x, top_left_y, bottom_right_x, bottom_rigth_y))
+    except(IndexError):
+        pass
+
+    return opencv_points
+

imagepreprocessing/file_operations.py

@@ -0,0 +1,17 @@
+
+def __read_from_file(file_name):
+    try:
+        with open(file_name,'r', encoding='utf-8') as file:
+            content = file.read()
+            return content
+    except (OSError, IOError) as e:
+        print(e)
+
+def __write_to_file(to_write, file_name, write_mode="w"):
+    try:
+        with open(file_name,write_mode, encoding='utf-8') as file:
+            for item in to_write:
+                file.write(item.__str__())
+                file.write("\n")
+    except (OSError, IOError) as e:
+        print(e)