Fix facedet download (#15811)

* Support downloading face models

* Handle download and loading correctly

* Add face dir creation

* Fix error

* Fix

* Formatting

* Move upload to button

* Show number of faces in library for each name

* Add text color for score

* Cleanup

frigate/app.py

@@ -34,6 +34,7 @@ from frigate.const import (
     CLIPS_DIR,
     CONFIG_DIR,
     EXPORT_DIR,
+    FACE_DIR,
     MODEL_CACHE_DIR,
     RECORD_DIR,
     SHM_FRAMES_VAR,
@@ -96,14 +97,19 @@ class FrigateApp:
         self.config = config
 
     def ensure_dirs(self) -> None:
-        for d in [
+        dirs = [
             CONFIG_DIR,
             RECORD_DIR,
             f"{CLIPS_DIR}/cache",
             CACHE_DIR,
             MODEL_CACHE_DIR,
             EXPORT_DIR,
-        ]:
+        ]
+
+        if self.config.face_recognition.enabled:
+            dirs.append(FACE_DIR)
+
+        for d in dirs:
             if not os.path.exists(d) and not os.path.islink(d):
                 logger.info(f"Creating directory: {d}")
                 os.makedirs(d)

frigate/embeddings/embeddings.py

@@ -123,19 +123,6 @@ class Embeddings:
             device="GPU" if config.semantic_search.model_size == "large" else "CPU",
         )
 
-        if self.config.face_recognition.enabled:
-            self.face_embedding = GenericONNXEmbedding(
-                model_name="facedet",
-                model_file="facedet.onnx",
-                download_urls={
-                    "facedet.onnx": "https://github.com/NickM-27/facenet-onnx/releases/download/v1.0/facedet.onnx",
-                    "landmarkdet.yaml": "https://github.com/NickM-27/facenet-onnx/releases/download/v1.0/landmarkdet.yaml",
-                },
-                model_size="small",
-                model_type=ModelTypeEnum.face,
-                requestor=self.requestor,
-            )
-
         self.lpr_detection_model = None
         self.lpr_classification_model = None
         self.lpr_recognition_model = None

frigate/embeddings/maintainer.py

@@ -100,19 +100,6 @@ class EmbeddingMaintainer(threading.Thread):
                 self.lpr_config, self.requestor, self.embeddings
             )
 
-    @property
-    def face_detector(self) -> cv2.FaceDetectorYN:
-        # Lazily create the classifier.
-        if "face_detector" not in self.__dict__:
-            self.__dict__["face_detector"] = cv2.FaceDetectorYN.create(
-                "/config/model_cache/facedet/facedet.onnx",
-                config="",
-                input_size=(320, 320),
-                score_threshold=0.8,
-                nms_threshold=0.3,
-            )
-        return self.__dict__["face_detector"]
-
     def run(self) -> None:
         """Maintain a SQLite-vec database for semantic search."""
         while not self.stop_event.is_set():
@@ -395,10 +382,9 @@ class EmbeddingMaintainer(threading.Thread):
 
     def _detect_face(self, input: np.ndarray) -> tuple[int, int, int, int]:
         """Detect faces in input image."""
-        self.face_detector.setInputSize((input.shape[1], input.shape[0]))
+        faces = self.face_classifier.detect_faces(input)
-        faces = self.face_detector.detect(input)
 
-        if faces[1] is None:
+        if faces is None or faces[1] is None:
             return None
 
         face = None

frigate/util/downloader.py

@@ -51,12 +51,14 @@ class ModelDownloader:
         download_path: str,
         file_names: List[str],
         download_func: Callable[[str], None],
+        complete_func: Callable[[], None] | None = None,
         silent: bool = False,
     ):
         self.model_name = model_name
         self.download_path = download_path
         self.file_names = file_names
         self.download_func = download_func
+        self.complete_func = complete_func
         self.silent = silent
         self.requestor = InterProcessRequestor()
         self.download_thread = None
@@ -97,6 +99,9 @@ class ModelDownloader:
                 },
             )
 
+        if self.complete_func:
+            self.complete_func()
+
         self.requestor.stop()
         self.download_complete.set()
 

frigate/util/model.py

@@ -2,7 +2,7 @@
 
 import logging
 import os
-from typing import Any, Optional
+from typing import Any
 
 import cv2
 import numpy as np
@@ -10,6 +10,7 @@ import onnxruntime as ort
 from playhouse.sqliteq import SqliteQueueDatabase
 
 from frigate.config.semantic_search import FaceRecognitionConfig
+from frigate.const import MODEL_CACHE_DIR
 
 try:
     import openvino as ov
@@ -162,22 +163,62 @@ class FaceClassificationModel:
     def __init__(self, config: FaceRecognitionConfig, db: SqliteQueueDatabase):
         self.config = config
         self.db = db
-        self.landmark_detector = cv2.face.createFacemarkLBF()
+        self.face_detector: cv2.FaceDetectorYN = None
+        self.landmark_detector: cv2.face.FacemarkLBF = None
+        self.face_recognizer: cv2.face.LBPHFaceRecognizer = None
 
-        if os.path.isfile("/config/model_cache/facedet/landmarkdet.yaml"):
+        download_path = os.path.join(MODEL_CACHE_DIR, "facedet")
-            self.landmark_detector.loadModel(
+        self.model_files = {
-                "/config/model_cache/facedet/landmarkdet.yaml"
+            "facedet.onnx": "https://github.com/NickM-27/facenet-onnx/releases/download/v1.0/facedet.onnx",
-            )
+            "landmarkdet.yaml": "https://github.com/NickM-27/facenet-onnx/releases/download/v1.0/landmarkdet.yaml",
+        }
 
-        self.recognizer: cv2.face.LBPHFaceRecognizer = (
+        if not all(
-            cv2.face.LBPHFaceRecognizer_create(
+            os.path.exists(os.path.join(download_path, n))
-                radius=2, threshold=(1 - config.min_score) * 1000
+            for n in self.model_files.keys()
-            )
+        ):
+            # conditionally import ModelDownloader
+            from frigate.util.downloader import ModelDownloader
+
+            self.downloader = ModelDownloader(
+                model_name="facedet",
+                download_path=download_path,
+                file_names=self.model_files.keys(),
+                download_func=self.__download_models,
+                complete_func=self.__build_detector,
             )
+            self.downloader.ensure_model_files()
+        else:
+            self.__build_detector()
+
         self.label_map: dict[int, str] = {}
         self.__build_classifier()
 
+    def __download_models(self, path: str) -> None:
+        try:
+            file_name = os.path.basename(path)
+            # conditionally import ModelDownloader
+            from frigate.util.downloader import ModelDownloader
+
+            ModelDownloader.download_from_url(self.model_files[file_name], path)
+        except Exception as e:
+            logger.error(f"Failed to download {path}: {e}")
+
+    def __build_detector(self) -> None:
+        self.face_detector = cv2.FaceDetectorYN.create(
+            "/config/model_cache/facedet/facedet.onnx",
+            config="",
+            input_size=(320, 320),
+            score_threshold=0.8,
+            nms_threshold=0.3,
+        )
+        self.landmark_detector = cv2.face.createFacemarkLBF()
+        self.landmark_detector.loadModel("/config/model_cache/facedet/landmarkdet.yaml")
+
     def __build_classifier(self) -> None:
+        if not self.landmark_detector:
+            return None
+
         labels = []
         faces = []
 
@@ -203,6 +244,11 @@ class FaceClassificationModel:
                 faces.append(img)
                 labels.append(idx)
 
+        self.recognizer: cv2.face.LBPHFaceRecognizer = (
+            cv2.face.LBPHFaceRecognizer_create(
+                radius=2, threshold=(1 - self.config.min_score) * 1000
+            )
+        )
         self.recognizer.train(faces, np.array(labels))
 
     def __align_face(
@@ -267,7 +313,17 @@ class FaceClassificationModel:
         self.labeler = None
         self.label_map = {}
 
-    def classify_face(self, face_image: np.ndarray) -> Optional[tuple[str, float]]:
+    def detect_faces(self, input: np.ndarray) -> tuple[int, cv2.typing.MatLike] | None:
+        if not self.face_detector:
+            return None
+
+        self.face_detector.setInputSize((input.shape[1], input.shape[0]))
+        return self.face_detector.detect(input)
+
+    def classify_face(self, face_image: np.ndarray) -> tuple[str, float] | None:
+        if not self.landmark_detector:
+            return None
+
         if not self.label_map:
             self.__build_classifier()
 

web/src/pages/FaceLibrary.tsx

@@ -1,5 +1,6 @@
 import { baseUrl } from "@/api/baseUrl";
 import AddFaceIcon from "@/components/icons/AddFaceIcon";
+import ActivityIndicator from "@/components/indicators/activity-indicator";
 import UploadImageDialog from "@/components/overlay/dialog/UploadImageDialog";
 import { Button } from "@/components/ui/button";
 import {
@@ -18,6 +19,8 @@ import {
   TooltipTrigger,
 } from "@/components/ui/tooltip";
 import useOptimisticState from "@/hooks/use-optimistic-state";
+import { cn } from "@/lib/utils";
+import { FrigateConfig } from "@/types/frigateConfig";
 import axios from "axios";
 import { useCallback, useEffect, useMemo, useRef, useState } from "react";
 import { LuImagePlus, LuTrash2 } from "react-icons/lu";
@@ -25,6 +28,8 @@ import { toast } from "sonner";
 import useSWR from "swr";
 
 export default function FaceLibrary() {
+  const { data: config } = useSWR<FrigateConfig>("config");
+
   // title
 
   useEffect(() => {
@@ -108,6 +113,10 @@ export default function FaceLibrary() {
     [pageToggle, refreshFaces],
   );
 
+  if (!config) {
+    return <ActivityIndicator />;
+  }
+
   return (
     <div className="flex size-full flex-col p-2">
       <Toaster />
@@ -120,7 +129,7 @@ export default function FaceLibrary() {
         onSave={onUploadImage}
       />
 
-      <div className="relative flex h-11 w-full items-center justify-between">
+      <div className="relative mb-2 flex h-11 w-full items-center justify-between">
         <ScrollArea className="w-full whitespace-nowrap">
           <div ref={tabsRef} className="flex flex-row">
             <ToggleGroup
@@ -156,17 +165,24 @@ export default function FaceLibrary() {
                   data-nav-item={item}
                   aria-label={`Select ${item}`}
                 >
-                  <div className="capitalize">{item}</div>
+                  <div className="capitalize">
+                    {item} ({faceData[item].length})
+                  </div>
                 </ToggleGroupItem>
               ))}
             </ToggleGroup>
             <ScrollBar orientation="horizontal" className="h-0" />
           </div>
         </ScrollArea>
+        <Button className="flex gap-2" onClick={() => setUpload(true)}>
+          <LuImagePlus className="size-7 rounded-md p-1 text-secondary-foreground" />
+          Upload Image
+        </Button>
       </div>
       {pageToggle &&
         (pageToggle == "train" ? (
           <TrainingGrid
+            config={config}
             attemptImages={trainImages}
             faceNames={faces}
             onRefresh={refreshFaces}
@@ -175,7 +191,6 @@ export default function FaceLibrary() {
           <FaceGrid
             faceImages={faceImages}
             pageToggle={pageToggle}
-            setUpload={setUpload}
             onRefresh={refreshFaces}
           />
         ))}
@@ -184,11 +199,13 @@ export default function FaceLibrary() {
 }
 
 type TrainingGridProps = {
+  config: FrigateConfig;
   attemptImages: string[];
   faceNames: string[];
   onRefresh: () => void;
 };
 function TrainingGrid({
+  config,
   attemptImages,
   faceNames,
   onRefresh,
@@ -200,6 +217,7 @@ function TrainingGrid({
           key={image}
           image={image}
           faceNames={faceNames}
+          threshold={config.face_recognition.threshold}
           onRefresh={onRefresh}
         />
       ))}
@@ -210,9 +228,15 @@ function TrainingGrid({
 type FaceAttemptProps = {
   image: string;
   faceNames: string[];
+  threshold: number;
   onRefresh: () => void;
 };
-function FaceAttempt({ image, faceNames, onRefresh }: FaceAttemptProps) {
+function FaceAttempt({
+  image,
+  faceNames,
+  threshold,
+  onRefresh,
+}: FaceAttemptProps) {
   const data = useMemo(() => {
     const parts = image.split("-");
 
@@ -283,7 +307,15 @@ function FaceAttempt({ image, faceNames, onRefresh }: FaceAttemptProps) {
         <div className="flex w-full flex-row items-center justify-between gap-2">
           <div className="flex flex-col items-start text-xs text-primary-variant">
             <div className="capitalize">{data.name}</div>
-            <div>{Number.parseFloat(data.score) * 100}%</div>
+            <div
+              className={cn(
+                Number.parseFloat(data.score) > threshold
+                  ? "text-success"
+                  : "text-danger",
+              )}
+            >
+              {Number.parseFloat(data.score) * 100}%
+            </div>
           </div>
           <div className="flex flex-row items-start justify-end gap-5 md:gap-4">
             <Tooltip>
@@ -327,15 +359,9 @@ function FaceAttempt({ image, faceNames, onRefresh }: FaceAttemptProps) {
 type FaceGridProps = {
   faceImages: string[];
   pageToggle: string;
-  setUpload: (upload: boolean) => void;
   onRefresh: () => void;
 };
-function FaceGrid({
+function FaceGrid({ faceImages, pageToggle, onRefresh }: FaceGridProps) {
-  faceImages,
-  pageToggle,
-  setUpload,
-  onRefresh,
-}: FaceGridProps) {
   return (
     <div className="scrollbar-container flex flex-wrap gap-2 overflow-y-scroll">
       {faceImages.map((image: string) => (
@@ -346,9 +372,6 @@ function FaceGrid({
           onRefresh={onRefresh}
         />
       ))}
-      <Button key="upload" className="size-40" onClick={() => setUpload(true)}>
-        <LuImagePlus className="size-10" />
-      </Button>
     </div>
   );
 }

web/src/types/frigateConfig.ts

@@ -290,6 +290,7 @@ export interface FrigateConfig {
 
   face_recognition: {
     enabled: boolean;
+    threshold: number;
   };
 
   ffmpeg: {