updated (removed realtime webcam

.vscode/settings.json

@@ -0,0 +1,3 @@
+{
+    "ros.distro": "noetic"
+}

app.py

@@ -72,17 +72,6 @@ class UI():
             self.distance_type = gr.State("cosine")
             gr.Markdown("## FaceAnalyzer face recognition test")
             with gr.Tabs():
-                with gr.TabItem('Realtime Recognize'):
-                    with gr.Blocks():
-                        with gr.Row():
-                            with gr.Column():
-                                self.rt_webcam = gr.Image(label="Input Image", source="webcam", streaming=True)
-                                self.start_streaming = gr.Button("Start webcam")
-                                self.start_streaming.click(self.start_webcam, [], [self.start_streaming])
-
-                            with gr.Column():
-                                self.rt_rec_img = gr.Image(label="Output Image")
-                                self.rt_webcam.change(self.process_webcam, inputs=self.rt_webcam, outputs=self.rt_rec_img, show_progress=False)
                 with gr.TabItem('Image Recognize'):
                     with gr.Blocks():
                         with gr.Row():
@@ -91,18 +80,6 @@ class UI():
                             with gr.Column():
                                 self.rt_rec_img = gr.Image(label="Output Image")
                                 self.rt_inp_img.change(self.process_image, inputs=self.rt_inp_img, outputs=self.rt_rec_img, show_progress=True)
-                with gr.TabItem('Add face from webcam'):
-                    with gr.Blocks():
-                        with gr.Row():
-                            with gr.Column():
-                                self.img = gr.Image(label="Input Image", source="webcam", streaming=True)
-                                self.txtFace_name = gr.Textbox(label="face_name")
-                                self.status = gr.Label(label="Status")
-                                self.txtFace_name.change(self.set_face_name, inputs=self.txtFace_name, outputs=self.status, show_progress=False)
-                                self.img.change(self.record_from_webcam, inputs=self.img, outputs=self.status, show_progress=False)
-                            with gr.Column():
-                                self.btn_start = gr.Button("Start Recording face")
-                                self.btn_start.click(self.start_stop)
                 with gr.TabItem('Add face from files'):
                     with gr.Blocks():
                         with gr.Row():
@@ -280,49 +257,6 @@ class UI():
             pickle.dump({"mean":embeddings_cloud_mean, "inv_cov":embeddings_cloud_inv_cov},f)
         print(f"Saved {name}")
 
-    def record_from_webcam(self, image):
-        if self.face_name is None or self.face_name=="":
-            self.embeddings_cloud=[]
-            self.is_recording=False
-            return "Please input a face name"
-        
-        if self.is_recording and image is not None:
-            if self.i < self.nb_images:
-                fa.image_size=(640, 480, 3)
-
-                # Process the image to extract faces and draw the masks on the face in the image
-                fa.process(image)
-                if fa.nb_faces>0:
-                    try:
-                        face = fa.faces[0]
-                        vertices = face.get_face_outer_vertices()
-                        image = face.getFaceBox(image, vertices, margins=(40,40,40,40))
-                        embedding = DeepFace.represent(image, enforce_detection=False)[0]["embedding"]
-                        self.embeddings_cloud.append(embedding)
-                        self.i+=1
-                    except Exception as ex:
-                        print(ex)
-                return f"Processing frame {self.i}/{self.nb_images}..."
-            else:
-                # Now let's find out where the face lives inside the latent space (128 dimensions space)
-
-                embeddings_cloud = np.array(self.embeddings_cloud)
-                embeddings_cloud_mean = embeddings_cloud.mean(axis=0)
-                embeddings_cloud_inv_cov = embeddings_cloud.std(axis=0)
-                # Now we save it.
-                # create a dialog box to ask for the subject name
-                name = self.face_name
-                with open(str(self.faces_path/f"{name}.pkl"),"wb") as f:
-                    pickle.dump({"mean":embeddings_cloud_mean, "inv_cov":embeddings_cloud_inv_cov},f)
-                print(f"Saved {name} embeddings")
-                self.i=0
-                self.embeddings_cloud=[]
-                self.is_recording=False
-                self.upgrade_faces()
-
-                return f"Saved {name} embeddings"
-        else:
-            return "Waiting"
         
     def record_from_files(self, images, face_name):
         if face_name is None or face_name=="":