Merge pull request #16 from Koldim2001/new_visualize_and_converter

New visualize and converter

Author: Koldim2001

README.md

@@ -172,7 +172,8 @@ Possible arguments of the ```visualize_results``` function:
 | random_object_colors    | bool            | False         | If true, colors for each object are selected randomly.                                        |
 | show_confidences        | bool            | False         | If true and show_class=True, confidences near class are visualized.                           |
 | axis_off                | bool            | True          | If true, axis is turned off in the final visualization.                                       |
-| show_classes_list       | list            | []            | If empty, visualize all classes. Otherwise, visualize only classes in the list.               |
+| show_classes_list       | list            | []            | If empty, visualize all classes. Otherwise, visualize only classes in the list.              |
+| list_of_class_colors      | list      | None         | A list of tuples representing the colors for each class in BGR format. If provided, these colors will be used for displaying the classes instead of random colors. |
 | return_image_array      | bool            | False         | If True, the function returns the image (BGR np.array) instead of displaying it.              |
 
 

patched_yolo_infer/README.md

@@ -7,6 +7,12 @@ The library also provides a sleek customization of the visualization of the infe
 **Model Support**: The library offers support for multiple ultralytics deep learning [models](https://docs.ultralytics.com/models/), such as YOLOv8, YOLOv8-seg, YOLOv9, YOLOv9-seg, YOLOv10, FastSAM, and RTDETR. Users can select from pre-trained options or utilize custom-trained models to best meet their task requirements.
 
 
+__Explanation of how Patch-Based-Inference works:__
+
+<p align="center">
+  <img width="600" alt="patched_inf_explanation" src="https://github.com/Koldim2001/YOLO-Patch-Based-Inference/blob/main/readme_content/patched_inf_explanation.gif?raw=true">
+</p>
+
 ## Installation
 You can install the library via pip:
 
@@ -141,6 +147,7 @@ Visualizes custom results of object detection or segmentation on an image.
 - **show_confidences** (*bool*): If true and show_class=True, confidences near class are visualized. Default is False.
 - **axis_off** (*bool*): If true, axis is turned off in the final visualization. Default is True.
 - **show_classes_list** (*list*): If empty, visualize all classes. Otherwise, visualize only classes in the list.
+- **list_of_class_colors** (*list*) A list of tuples representing the colors for each class in BGR format. If provided, these colors will be used for displaying the classes instead of random colors.
 - **return_image_array** (*bool*): If True, the function returns the image (BGR np.array) instead of displaying it. Default is False.
 
 

patched_yolo_infer/__init__.py

@@ -1,5 +1,10 @@
-from .functions_extra import visualize_results_usual_yolo_inference, get_crops, visualize_results
+from .functions_extra import (
+    visualize_results_usual_yolo_inference,
+    get_crops,
+    visualize_results,
+    create_masks_from_polygons,
+)
 
 from .nodes.MakeCropsDetectThem import MakeCropsDetectThem
 from .nodes.CombineDetections import CombineDetections
-from .elements.CropElement import CropElement
+from .elements.CropElement import CropElement

patched_yolo_infer/elements/CropElement.py

@@ -55,7 +55,7 @@ def calculate_inference(self, model, imgsz=640, conf=0.35, iou=0.7, segment=Fals
                 self.polygons = [mask.astype(np.uint16) for mask in pred.masks.xy]
             else:
                 # Get the masks
-                self.detected_masks = pred.masks.data.cpu().numpy()
+                self.detected_masks = pred.masks.data.cpu().numpy().astype(np.uint8)
 
 
     def calculate_real_values(self):
@@ -88,7 +88,7 @@ def calculate_real_values(self):
                             x_start_global:x_start_global+self.crop.shape[1]] = mask_resized
 
                 # Append the masked image to the list of detected_masks_real
-                self.detected_masks_real.append(black_image)
+                self.detected_masks_real.append(black_image.astype(np.uint8))
 
         if self.polygons is not None:
             # Adjust the mask coordinates
@@ -116,7 +116,7 @@ def resize_results(self):
             # Resize mask
             mask_resized = cv2.resize(mask, (self.source_image.shape[1], self.source_image.shape[0]),
                                     interpolation=cv2.INTER_NEAREST)
-            resized_masks.append(mask_resized)
+            resized_masks.append(mask_resized.astype(np.uint8))
 
 
         for polygon in self.detected_polygons_real:

patched_yolo_infer/functions_extra.py

@@ -27,6 +27,7 @@ def visualize_results_usual_yolo_inference(
     show_confidences=False,
     axis_off=True,
     show_classes_list=[],
+    list_of_class_colors=None,
     return_image_array=False,
     inference_extra_args=None,
 ):
@@ -44,8 +45,8 @@ def visualize_results_usual_yolo_inference(
         show_class (bool): Whether to show class labels. Default is True.
         fill_mask (bool): Whether to fill the segmented regions with color. Default is False.
         alpha (float): The transparency of filled masks. Default is 0.3.
-        color_class_background (tuple): The background bgr color for class labels. Default is (0, 0, 255) (red).
-        color_class_text (tuple): The text color for class labels. Default is (255, 255, 255) (white).
+        color_class_background (tuple): The background BGR color for class labels. Default is (0, 0, 255) (red).
+        color_class_text (tuple): The text BGR color for class labels. Default is (255, 255, 255) (white).
         thickness (int): The thickness of bounding box and text. Default is 4.
         font: The font type for class labels. Default is cv2.FONT_HERSHEY_SIMPLEX.
         font_scale (float): The scale factor for font size. Default is 1.5.
@@ -56,6 +57,9 @@ def visualize_results_usual_yolo_inference(
         axis_off (bool): If True, axis is turned off in the final visualization.
         show_classes_list (list): If empty, visualize all classes. Otherwise, visualize only classes in the list.
         inference_extra_args (dict/None): Dictionary with extra ultralytics inference parameters.
+        list_of_class_colors (list/None): A list of tuples representing the colors for each class in BGR format. If provided,  
+                    these colors will be used for displaying the classes instead of random colors. The number of tuples 
+                    in the list must match the number of possible classes in the network.
         return_image_array (bool): If True, the function returns the image bgr array instead of displaying it. 
                                    Default is False.
 
@@ -106,10 +110,12 @@ def visualize_results_usual_yolo_inference(
 
             if random_object_colors:
                 color = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255))
-            else:
+            elif list_of_class_colors is None:
                 # Assign color according to class
                 random.seed(int(classes[i] + delta_colors))
                 color = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255))
+            else:
+                color = list_of_class_colors[classes[i]]
 
             box = boxes[i]
             x_min, y_min, x_max, y_max = box
@@ -278,6 +284,7 @@ def visualize_results(
     show_confidences=False,
     axis_off=True,
     show_classes_list=[],
+    list_of_class_colors=None,
     return_image_array=False
 ):
     """
@@ -295,8 +302,8 @@ def visualize_results(
         show_class (bool): Whether to show class labels. Default is True.
         fill_mask (bool): Whether to fill the segmented regions with color. Default is False.
         alpha (float): The transparency of filled masks. Default is 0.3.
-        color_class_background (tuple): The background bgr color for class labels. Default is (0, 0, 255) (red).
-        color_class_text (tuple): The text color for class labels. Default is (255, 255, 255) (white).
+        color_class_background (tuple): The background BGR color for class labels. Default is (0, 0, 255) (red).
+        color_class_text (tuple): The text BGR color for class labels. Default is (255, 255, 255) (white).
         thickness (int): The thickness of bounding box and text. Default is 4.
         font: The font type for class labels. Default is cv2.FONT_HERSHEY_SIMPLEX.
         font_scale (float): The scale factor for font size. Default is 1.5.
@@ -306,8 +313,10 @@ def visualize_results(
         show_confidences (bool): If true and show_class=True, confidences near class are visualized. Default is False.
         axis_off (bool): If true, axis is turned off in the final visualization. Default is True.
         show_classes_list (list): If empty, visualize all classes. Otherwise, visualize only classes in the list.
-        return_image_array (bool): If True, the function returns the image bgr array instead of displaying it. 
-                                   Default is False.
+        list_of_class_colors (list/None): A list of tuples representing the colors for each class in BGR format. If provided,  
+                    these colors will be used for displaying the classes instead of random colors. The number of tuples 
+                    in the list must match the number of possible classes in the network.
+        return_image_array (bool): If True, the function returns the image bgr array instead of displaying it. Default is False.
                                    
     Returns:
         None/np.array
@@ -332,10 +341,12 @@ def visualize_results(
 
         if random_object_colors:
             color = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255))
-        else:
+        elif list_of_class_colors is None:
             # Assign color according to class
             random.seed(int(classes_ids[i] + delta_colors))
             color = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255))
+        else:
+            color = list_of_class_colors[classes_ids[i]]
 
         box = boxes[i]
         x_min, y_min, x_max, y_max = box
@@ -409,3 +420,43 @@ def visualize_results(
         if axis_off:
             plt.axis('off')
         plt.show()
+
+
+def create_masks_from_polygons(polygons, image):
+    """
+    Create binary masks from a list of polygons.
+
+    This function takes a list of polygons and an image, and generates binary masks
+    where each mask corresponds to one polygon. The masks are boolean arrays with
+    the same dimensions as the input image, where the regions covered by the polygons
+    are marked as True.
+
+    Parameters:
+    polygons (list of numpy.ndarray): A list of polygons, where each polygon is
+        represented as a numpy array of shape (N, 2) containing N (x, y) coordinates.
+    image (numpy.ndarray): The input image, used to determine the dimensions of the masks.
+
+    Returns:
+    list of numpy.ndarray: A list of binary masks, where each mask is a boolean
+        numpy array of the same dimensions as the input image.
+    """
+    # Get the dimensions of the image
+    height, width = image.shape[:2]
+    
+    # Create empty masks
+    masks = []
+    
+    for polygon in polygons:
+        if len(polygon) > 0:
+            points = np.array(polygon.reshape((-1, 1, 2)), dtype=np.int32)
+        
+        # Create an empty mask with the same size as the image
+        mask = np.zeros((height, width), dtype=np.uint8)
+        
+        # Draw the polygon on the mask
+        cv2.fillPoly(mask, [points], 1)
+        
+        # Add the mask to the list
+        masks.append(mask)
+    
+    return masks

setup.py

@@ -8,7 +8,7 @@
     long_description = "\n" + fh.read()
 
 
-VERSION = '1.2.8'
+VERSION = '1.2.9'
 DESCRIPTION = '''Patch-Based-Inference for detection/segmentation of small objects in images.'''
 
 setup(