diff --git a/anylabeling/services/auto_labeling/segment_anything.py b/anylabeling/services/auto_labeling/segment_anything.py
index 9b40677..5ad9632 100644
--- a/anylabeling/services/auto_labeling/segment_anything.py
+++ b/anylabeling/services/auto_labeling/segment_anything.py
@@ -1,6 +1,5 @@
 import logging
 import os
-import sys
 from copy import deepcopy
 
 import cv2
@@ -11,11 +10,10 @@
 
 from anylabeling.utils import GenericWorker
 from anylabeling.views.labeling.shape import Shape
-from anylabeling.views.labeling.utils.opencv import qt_img_to_cv_img
-
+from anylabeling.views.labeling.utils.opencv import qt_img_to_cv_img, qt_img_to_rgb_cv_img
+from .lru_cache import LRUCache
 from .model import Model
 from .types import AutoLabelingResult
-from .lru_cache import LRUCache
 
 
 class SegmentAnything(Model):
@@ -178,7 +176,7 @@ def get_preprocess_shape(oldh: int, oldw: int, long_side_length: int):
         return (newh, neww)
 
     def apply_coords(
-        self, coords: np.ndarray, original_size, target_length
+            self, coords: np.ndarray, original_size, target_length
     ) -> np.ndarray:
         """
         Expects a numpy array of length 2 in the final dimension. Requires the
@@ -201,8 +199,8 @@ def run_decoder(self, image_embedding, resized_ratio):
             [input_points, np.array([[0.0, 0.0]])], axis=0
         )[None, :, :]
         onnx_label = np.concatenate([input_labels, np.array([-1])], axis=0)[
-            None, :
-        ].astype(np.float32)
+                     None, :
+                     ].astype(np.float32)
         onnx_coord = self.apply_coords(
             onnx_coord, self.size_after_apply_max_width_height, self.input_size
         ).astype(np.float32)
@@ -336,7 +334,7 @@ def predict_shapes(self, image, filename=None) -> AutoLabelingResult:
                     image_embedding,
                 ) = cached_data
             else:
-                cv_image = qt_img_to_cv_img(image)
+                cv_image = qt_img_to_rgb_cv_img(image, filename)
                 encoder_inputs, resized_ratio = self.pre_process(cv_image)
                 if self.stop_inference:
                     return AutoLabelingResult([], replace=False)
@@ -394,8 +392,8 @@ def on_next_files_changed(self, next_files):
         and run inference to save time for user.
         """
         if (
-            self.pre_inference_thread is None
-            or not self.pre_inference_thread.isRunning()
+                self.pre_inference_thread is None
+                or not self.pre_inference_thread.isRunning()
         ):
             self.pre_inference_thread = QThread()
             self.pre_inference_worker = GenericWorker(
diff --git a/anylabeling/services/auto_labeling/yolov5.py b/anylabeling/services/auto_labeling/yolov5.py
index 601956e..7db7ed2 100644
--- a/anylabeling/services/auto_labeling/yolov5.py
+++ b/anylabeling/services/auto_labeling/yolov5.py
@@ -6,8 +6,7 @@
 from PyQt5 import QtCore
 
 from anylabeling.views.labeling.shape import Shape
-from anylabeling.views.labeling.utils.opencv import qt_img_to_cv_img
-
+from anylabeling.views.labeling.utils.opencv import qt_img_to_rgb_cv_img
 from .model import Model
 from .types import AutoLabelingResult
 
@@ -157,7 +156,7 @@ def predict_shapes(self, image, image_path=None):
             return []
 
         try:
-            image = qt_img_to_cv_img(image)
+            image = qt_img_to_rgb_cv_img(image, image_path)
         except Exception as e:  # noqa
             logging.warning("Could not inference model")
             logging.warning(e)
diff --git a/anylabeling/services/auto_labeling/yolov8.py b/anylabeling/services/auto_labeling/yolov8.py
index c869a95..904153c 100644
--- a/anylabeling/services/auto_labeling/yolov8.py
+++ b/anylabeling/services/auto_labeling/yolov8.py
@@ -6,8 +6,7 @@
 from PyQt5 import QtCore
 
 from anylabeling.views.labeling.shape import Shape
-from anylabeling.views.labeling.utils.opencv import qt_img_to_cv_img
-
+from anylabeling.views.labeling.utils.opencv import qt_img_to_rgb_cv_img
 from .model import Model
 from .types import AutoLabelingResult
 
@@ -153,7 +152,7 @@ def predict_shapes(self, image, image_path=None):
             return []
 
         try:
-            image = qt_img_to_cv_img(image)
+            image = qt_img_to_rgb_cv_img(image, image_path)
         except Exception as e:  # noqa
             logging.warning("Could not inference model")
             logging.warning(e)
diff --git a/anylabeling/views/labeling/utils/opencv.py b/anylabeling/views/labeling/utils/opencv.py
index cb75506..99b0c4a 100644
--- a/anylabeling/views/labeling/utils/opencv.py
+++ b/anylabeling/views/labeling/utils/opencv.py
@@ -1,5 +1,33 @@
+import os.path
+
+import cv2
+import numpy as np
 import qimage2ndarray
 from PyQt5 import QtGui
+from PyQt5.QtGui import QImage
+
+
+def qt_img_to_rgb_cv_img(qt_img, img_path=None):
+    """
+    Convert 8bit/16bit RGB image or 8bit/16bit Gray image to 8bit RGB image
+    """
+    if img_path is not None and os.path.exists(img_path):
+        # Load Image From Path Directly
+        cv_image = cv2.imdecode(np.fromfile(img_path, dtype=np.uint8), -1)
+        cv_image = cv2.cvtColor(cv_image, cv2.COLOR_BGR2RGB)
+    else:
+        if qt_img.format() == QImage.Format_RGB32 or qt_img.format() == QImage.Format_ARGB32:
+            cv_image = qimage2ndarray.rgb_view(qt_img)
+        else:
+            cv_image = qimage2ndarray.raw_view(qt_img)
+    # To uint8
+    if cv_image.dtype != np.uint8:
+        cv2.normalize(cv_image, cv_image, 0, 255, cv2.NORM_MINMAX)
+        cv_image = np.array(cv_image, dtype=np.uint8)
+    # To RGB
+    if len(cv_image.shape) == 2 or cv_image.shape[2] == 1:
+        cv_image = cv2.merge([cv_image, cv_image, cv_image])
+    return cv_image
 
 
 def qt_img_to_cv_img(in_image):