Face recognition improvements (#16034)

frigate/api/classification.py

@@ -24,14 +24,18 @@ def get_faces():
     face_dict: dict[str, list[str]] = {}
 
     for name in os.listdir(FACE_DIR):
-        face_dict[name] = []
-
         face_dir = os.path.join(FACE_DIR, name)
 
         if not os.path.isdir(face_dir):
             continue
 
-        for file in os.listdir(face_dir):
+        face_dict[name] = []
+
+        for file in sorted(
+            os.listdir(face_dir),
+            key=lambda f: os.path.getctime(os.path.join(face_dir, f)),
+            reverse=True,
+        ):
             face_dict[name].append(file)
 
     return JSONResponse(status_code=200, content=face_dict)
@@ -81,6 +85,10 @@ def train_face(request: Request, name: str, body: dict = None):
     new_name = f"{name}-{rand_id}.webp"
     new_file = os.path.join(FACE_DIR, f"{name}/{new_name}")
     shutil.move(training_file, new_file)
+
+    context: EmbeddingsContext = request.app.embeddings
+    context.clear_face_classifier()
+
     return JSONResponse(
         content=(
             {

frigate/comms/embeddings_updater.py

@@ -9,6 +9,7 @@ SOCKET_REP_REQ = "ipc:///tmp/cache/embeddings"
 
 
 class EmbeddingsRequestEnum(Enum):
+    clear_face_classifier = "clear_face_classifier"
     embed_description = "embed_description"
     embed_thumbnail = "embed_thumbnail"
     generate_search = "generate_search"

frigate/data_processing/real_time/api.py

@@ -32,9 +32,12 @@ class RealTimeProcessorApi(ABC):
         pass
 
     @abstractmethod
-    def handle_request(self, request_data: dict[str, any]) -> dict[str, any] | None:
+    def handle_request(
+        self, topic: str, request_data: dict[str, any]
+    ) -> dict[str, any] | None:
         """Handle metadata requests.
         Args:
+            topic (str): topic that dictates what work is requested.
             request_data (dict): containing data about requested change to process.
 
         Returns:

frigate/data_processing/real_time/bird_processor.py

@@ -146,7 +146,7 @@ class BirdProcessor(RealTimeProcessorApi):
         if resp.status_code == 200:
             self.detected_birds[obj_data["id"]] = score
 
-    def handle_request(self, request_data):
+    def handle_request(self, topic, request_data):
         return None
 
     def expire_object(self, object_id):

frigate/data_processing/real_time/face_processor.py

@@ -12,6 +12,7 @@ import cv2
 import numpy as np
 import requests
 
+from frigate.comms.embeddings_updater import EmbeddingsRequestEnum
 from frigate.config import FrigateConfig
 from frigate.const import FACE_DIR, FRIGATE_LOCALHOST, MODEL_CACHE_DIR
 from frigate.util.image import area
@@ -353,45 +354,52 @@ class FaceProcessor(RealTimeProcessorApi):
 
         self.__update_metrics(datetime.datetime.now().timestamp() - start)
 
-    def handle_request(self, request_data) -> dict[str, any] | None:
-        rand_id = "".join(random.choices(string.ascii_lowercase + string.digits, k=6))
-        label = request_data["face_name"]
-        id = f"{label}-{rand_id}"
-
-        if request_data.get("cropped"):
-            thumbnail = request_data["image"]
-        else:
-            img = cv2.imdecode(
-                np.frombuffer(base64.b64decode(request_data["image"]), dtype=np.uint8),
-                cv2.IMREAD_COLOR,
+    def handle_request(self, topic, request_data) -> dict[str, any] | None:
+        if topic == EmbeddingsRequestEnum.clear_face_classifier.value:
+            self.__clear_classifier()
+        elif topic == EmbeddingsRequestEnum.register_face.value:
+            rand_id = "".join(
+                random.choices(string.ascii_lowercase + string.digits, k=6)
             )
-            face_box = self.__detect_face(img)
+            label = request_data["face_name"]
+            id = f"{label}-{rand_id}"
 
-            if not face_box:
-                return {
-                    "message": "No face was detected.",
-                    "success": False,
-                }
+            if request_data.get("cropped"):
+                thumbnail = request_data["image"]
+            else:
+                img = cv2.imdecode(
+                    np.frombuffer(
+                        base64.b64decode(request_data["image"]), dtype=np.uint8
+                    ),
+                    cv2.IMREAD_COLOR,
+                )
+                face_box = self.__detect_face(img)
 
-            face = img[face_box[1] : face_box[3], face_box[0] : face_box[2]]
-            ret, thumbnail = cv2.imencode(
-                ".webp", face, [int(cv2.IMWRITE_WEBP_QUALITY), 100]
-            )
+                if not face_box:
+                    return {
+                        "message": "No face was detected.",
+                        "success": False,
+                    }
 
-        # write face to library
-        folder = os.path.join(FACE_DIR, label)
-        file = os.path.join(folder, f"{id}.webp")
-        os.makedirs(folder, exist_ok=True)
+                face = img[face_box[1] : face_box[3], face_box[0] : face_box[2]]
+                _, thumbnail = cv2.imencode(
+                    ".webp", face, [int(cv2.IMWRITE_WEBP_QUALITY), 100]
+                )
 
-        # save face image
-        with open(file, "wb") as output:
-            output.write(thumbnail.tobytes())
+            # write face to library
+            folder = os.path.join(FACE_DIR, label)
+            file = os.path.join(folder, f"{id}.webp")
+            os.makedirs(folder, exist_ok=True)
 
-        self.__clear_classifier()
-        return {
-            "message": "Successfully registered face.",
-            "success": True,
-        }
+            # save face image
+            with open(file, "wb") as output:
+                output.write(thumbnail.tobytes())
+
+            self.__clear_classifier()
+            return {
+                "message": "Successfully registered face.",
+                "success": True,
+            }
 
     def expire_object(self, object_id: str):
         if object_id in self.detected_faces:

frigate/embeddings/__init__.py

@@ -211,6 +211,11 @@ class EmbeddingsContext:
 
         return self.db.execute_sql(sql_query).fetchall()
 
+    def clear_face_classifier(self) -> None:
+        self.requestor.send_data(
+            EmbeddingsRequestEnum.clear_face_classifier.value, None
+        )
+
     def delete_face_ids(self, face: str, ids: list[str]) -> None:
         folder = os.path.join(FACE_DIR, face)
         for id in ids:

frigate/embeddings/maintainer.py

@@ -140,7 +140,7 @@ class EmbeddingMaintainer(threading.Thread):
                     )
                 else:
                     for processor in self.processors:
-                        resp = processor.handle_request(data)
+                        resp = processor.handle_request(topic, data)
 
                         if resp is not None:
                             return resp