Merge a45ab7029012da80547dc864fc7f6509c93afb67 into 453c6e38a51e9d1d5a2aa5fb7f1014a711913397

ultralytics/engine/results.py

@@ -115,6 +115,7 @@ class Results(SimpleClass):
         self.probs = Probs(probs) if probs is not None else None
         self.keypoints = Keypoints(keypoints, self.orig_shape) if keypoints is not None else None
         self.obb = OBB(obb, self.orig_shape) if obb is not None else None
+        self.centroids = Centroids(self.boxes.xyxy, self.orig_shape) if self.boxes is not None else None
         self.speed = {"preprocess": None, "inference": None, "postprocess": None}  # milliseconds per image
         self.names = names
         self.path = path
@@ -198,6 +199,7 @@ class Results(SimpleClass):
         boxes=True,
         masks=True,
         probs=True,
+        centroids=True,
         show=False,
         save=False,
         filename=None,
@@ -219,6 +221,7 @@ class Results(SimpleClass):
             boxes (bool): Whether to plot the bounding boxes.
             masks (bool): Whether to plot the masks.
             probs (bool): Whether to plot classification probability
+            centroids (bool): Whether to plot the centroids.
             show (bool): Whether to display the annotated image directly.
             save (bool): Whether to save the annotated image to `filename`.
             filename (str): Filename to save image to if save is True.
@@ -248,6 +251,7 @@ class Results(SimpleClass):
         pred_boxes, show_boxes = self.obb if is_obb else self.boxes, boxes
         pred_masks, show_masks = self.masks, masks
         pred_probs, show_probs = self.probs, probs
+        pred_centroids, show_centroids = self.centroids, centroids
         annotator = Annotator(
             deepcopy(self.orig_img if img is None else img),
             line_width,
@@ -291,6 +295,16 @@ class Results(SimpleClass):
             for k in reversed(self.keypoints.data):
                 annotator.kpts(k, self.orig_shape, radius=kpt_radius, kpt_line=kpt_line)
 
+        # Plot centroids
+        if pred_centroids is not None and show_centroids:
+            for i, centroid in enumerate(pred_centroids.xy):
+                if pred_boxes is not None:
+                    c = int(pred_boxes.cls[i])
+                    color = colors(c, True)
+                else:
+                    color = (0, 255, 0)  # Default to green if no class information
+                annotator.dot(centroid, color=color, radius=5)
+
         # Show results
         if show:
             annotator.show(self.path)
@@ -419,6 +433,13 @@ class Results(SimpleClass):
                     "y": (y / h).numpy().round(decimals).tolist(),
                     "visible": visible.numpy().round(decimals).tolist(),
                 }
+
+            if self.centroids is not None:
+                result["centroid"] = {
+                    "x": round(self.centroids.xy[i][0].item() / w, decimals),
+                    "y": round(self.centroids.xy[i][1].item() / h, decimals)
+                }
+
             results.append(result)
 
         return results
@@ -741,3 +762,47 @@ class OBB(BaseTensor):
         y2 = self.xyxyxyxy[..., 1].max(1).values
         xyxy = [x1, y1, x2, y2]
         return np.stack(xyxy, axis=-1) if isinstance(self.data, np.ndarray) else torch.stack(xyxy, dim=-1)
+
+
+class Centroids(BaseTensor):
+    """
+    A class for storing and manipulating detection centroids.
+
+    Attributes:
+        xy (torch.Tensor): A tensor containing x, y coordinates of centroids for each detection.
+        xyn (torch.Tensor): A normalized version of xy with coordinates in the range [0, 1].
+
+    Methods:
+        cpu(): Returns a copy of the centroids tensor on CPU memory.
+        numpy(): Returns a copy of the centroids tensor as a numpy array.
+        cuda(): Returns a copy of the centroids tensor on GPU memory.
+        to(device, dtype): Returns a copy of the centroids tensor with the specified device and dtype.
+    """
+
+    def __init__(self, boxes, orig_shape) -> None:
+        """
+        Initialize the Centroids object with bounding boxes and original image size.
+
+        Args:
+            boxes (torch.Tensor): A tensor of bounding boxes in xyxy format.
+            orig_shape (tuple): Original image size in (height, width) format.
+        """
+        centroids = torch.zeros(boxes.shape[0], 2, device=boxes.device)
+        centroids[:, 0] = (boxes[:, 0] + boxes[:, 2]) / 2  # x-coordinate
+        centroids[:, 1] = (boxes[:, 1] + boxes[:, 3]) / 2  # y-coordinate
+        super().__init__(centroids, orig_shape)
+
+    @property
+    @lru_cache(maxsize=1)
+    def xy(self):
+        """Returns x, y coordinates of centroids."""
+        return self.data
+
+    @property
+    @lru_cache(maxsize=1)
+    def xyn(self):
+        """Returns normalized x, y coordinates of centroids."""
+        xy = self.xy.clone() if isinstance(self.xy, torch.Tensor) else np.copy(self.xy)
+        xy[:, 0] /= self.orig_shape[1]
+        xy[:, 1] /= self.orig_shape[0]
+        return xy

ultralytics/utils/plotting.py

@@ -593,6 +593,23 @@ class Annotator:
         cv2.circle(self.im, center_bbox, pins_radius, color, -1)
         cv2.line(self.im, center_point, center_bbox, color, thickness)
 
+    def dot(self, xy, color=(255, 0, 0), radius=5):
+        """
+        Draw a dot on the image.
+
+        Args:
+            xy (tuple): The (x, y) coordinates of the dot's center.
+            color (tuple, optional): The color of the dot in BGR format. Defaults to (255, 0, 0) (blue).
+            radius (int, optional): The radius of the dot. Defaults to 5.
+        """
+        if self.pil:
+            # Convert to PIL ImageDraw
+            draw = ImageDraw.Draw(self.im)
+            draw.ellipse([xy[0]-radius, xy[1]-radius, xy[0]+radius, xy[1]+radius], fill=color, outline=color)
+        else:
+            cv2.circle(self.im, (int(xy[0]), int(xy[1])), radius, color, -1, lineType=cv2.LINE_AA)
+
+
 
 @TryExcept()  # known issue https://github.com/ultralytics/yolov5/issues/5395
 @plt_settings()