sface wrapper

deepface/basemodels/sface_opencv_wrapper.py

@@ -0,0 +1,55 @@
+import os
+
+import cv2 as cv
+import gdown
+
+from deepface.commons import functions
+
+_url = "https://github.com/opencv/opencv_zoo/raw/master/models/face_recognition_sface/face_recognition_sface_2021dec.onnx"
+
+
+class _Layer:
+    input_shape = (None, 112, 122, 3)
+
+
+class SFace:
+    def __init__(self, model_path, backend_id=0, target_id=0):
+        self._modelPath = model_path
+        self._backendId = backend_id
+        self._targetId = target_id
+        self._model = cv.FaceRecognizerSF.create(
+            model=self._modelPath,
+            config="",
+            backend_id=self._backendId,
+            target_id=self._targetId)
+
+        self.layers = [_Layer()]
+
+    def _preprocess(self, image, bbox):
+        if bbox is None:
+            return image
+        else:
+            return self._model.alignCrop(image, bbox)
+
+    def predict(self, image, bbox=None, **kwargs):
+        # Preprocess
+        # print(image.max())
+        # input_blob = self._preprocess(image, bbox)
+
+        # Forward
+        features = self._model.feature(image[0])
+        return features
+
+
+def load_model(*args, **kwargs):
+    home = functions.get_deepface_home()
+
+    file_name = home + '/.deepface/weights/face_recognition_sface_2021dec.onnx'
+    if not os.path.isfile(file_name):
+        print("sface weights will be downloaded...")
+
+        output = file_name
+        gdown.download(_url, output, quiet=False)
+
+    model = SFace(file_name, 0, 0)
+    return model