Rope-Opal-01

Reverted to previous model load parameters. Some people were having issues with the new settings.
Markers now clear propoerly when changing videos
Input Face loading errors are now handled gracefully
Added option to select Mean or Median when multi-selecting input faces

rope/Dicts.py

@@ -287,7 +287,7 @@
 # Text Selection
 'DetectTypeTextSelInfoText':        'FACE DETECTION MODEL:\nSelect the face detection model. Mostly only subtle differences, but can significant differences when the face is at extreme angles or covered.',
 'DetectTypeTextSelMode':            'Retinaface',
-'DetectTypeTextSelModes':           ['Retinaface', 'Yolov8'],
+'DetectTypeTextSelModes':           ['Retinaface', 'Yolov8', 'SCRDF'],
 'PreviewModeTextSelInfoText':       '',
 'PreviewModeTextSelMode':           'Video',
 'PreviewModeTextSelModes':          ['Video', 'Image'],
@@ -300,6 +300,9 @@
 'RestorerTypeTextSelInfoText':      'RESTORER TYPE:\nSelect the Restorer type.\nSpeed: GPEN256>GFPGAN>CF>GPEN512',
 'RestorerTypeTextSelMode':          'GFPGAN',
 'RestorerTypeTextSelModes':         ['GFPGAN', 'CF', 'GPEN256', 'GPEN512'],
+'MergeTextSelInfoText':      'INPUT FACES MERGE MATH:\nWhen shift-clicking face for merging, determines how the embedding vectors are combined.',
+'MergeTextSelMode':          'Mean',
+'MergeTextSelModes':         ['Mean', 'Median'],
 
 
 # Text Entry

rope/GUI.py

@@ -29,7 +29,7 @@ def __init__(self, models):
         super().__init__()
 
         self.models = models
-        self.title('Rope-Opal-00')
+        self.title('Rope-Opal-01')
         self.target_media = []
         self.target_video_file = []
         self.action_q = []
@@ -419,7 +419,7 @@ def create_gui(self):
         canvas = tk.Canvas(r_frame, style.canvas_frame_label_3, bd=0, width=width)
         canvas.grid(row=1, column=0, sticky='NEWS', pady=0, padx=0)
 
-        parameters_canvas = tk.Frame(canvas, style.canvas_frame_label_3, bd=0, width=width, height=970)
+        parameters_canvas = tk.Frame(canvas, style.canvas_frame_label_3, bd=0, width=width, height=1000)
         parameters_canvas.grid(row=0, column=0, sticky='NEWS', pady=0, padx=0)  
 
 
@@ -569,16 +569,18 @@ def create_gui(self):
         row += bottom_border_delta           
 
         # Cats and Dogs
-        self.widget['ThreadsSlider'] = GE.Slider2(parameters_canvas, 'ThreadsSlider', 'Threads', 3, self.update_data, 'parameter', 398, 20, 1, row, 0.62)
+        self.widget['ThreadsSlider'] = GE.Slider2(parameters_canvas, 'ThreadsSlider', 'Threads', 3, self.update_data, 'control', 398, 20, 1, row, 0.62)
 
         row += row_delta 
         self.widget['DetectTypeTextSel'] = GE.TextSelection(parameters_canvas, 'DetectTypeTextSel', 'Detection Type', 3, self.update_data, 'parameter', 'parameter', 398, 20, 1, row, 0.62)
         row += row_delta 
         self.widget['DetectScoreSlider'] = GE.Slider2(parameters_canvas, 'DetectScoreSlider', 'Detect Score', 3, self.update_data, 'parameter', 398, 20, 1, row, 0.62)
         row += row_delta 
-        self.widget['RecordTypeTextSel'] = GE.TextSelection(parameters_canvas, 'RecordTypeTextSel', 'Record Type', 3, self.update_data, 'parameter', 'parameter', 398, 20, 1, row, 0.62)
+        self.widget['RecordTypeTextSel'] = GE.TextSelection(parameters_canvas, 'RecordTypeTextSel', 'Record Type', 3, self.update_data, 'control', 'control', 398, 20, 1, row, 0.62)
         row += row_delta 
-        self.widget['VideoQualSlider'] = GE.Slider2(parameters_canvas, 'VideoQualSlider', 'FFMPEG Quality', 3, self.update_data, 'parameter', 398, 20, 1, row, 0.62)
+        self.widget['VideoQualSlider'] = GE.Slider2(parameters_canvas, 'VideoQualSlider', 'FFMPEG Quality', 3, self.update_data, 'control', 398, 20, 1, row, 0.62)
+        row += row_delta 
+        self.widget['MergeTextSel'] = GE.TextSelection(parameters_canvas, 'MergeTextSel', 'Merge Math', 3, self.update_data, 'control', 'control', 398, 20, 1, row, 0.62)            
 
     ### Other
         self.layer['tooltip_frame'] = tk.Frame(r_frame, style.canvas_frame_label_3, height=80)
@@ -964,23 +966,23 @@ def load_source_faces(self):
                     img = cv2.imread(file)                    
 
                     if img is not None:     
-                        img = torch.from_numpy(img).to('cuda')
+                        img = torch.from_numpy(img.astype('uint8')).to('cuda')
                         img = img.permute(2,0,1)
-                        kpss = self.models.run_detect(img, max_num=1)[0] # Just one face here
-
-                        ret = []
-
-                        if kpss is not None:
+                        try: 
+                            kpss = self.models.run_detect(img, max_num=1)[0] # Just one face here
+                        except IndexError:
+                            print('Image cropped too close:', file) 
+                        else:
                             face_emb, cropped_image = self.models.run_recognize(img, kpss)
-                            ret.append([kpss, face_emb, cropped_image])
-
-                        if ret:
-                            crop = cv2.cvtColor(ret[0][2].cpu().numpy(), cv2.COLOR_BGR2RGB)            
+                            crop = cv2.cvtColor(cropped_image.cpu().numpy(), cv2.COLOR_BGR2RGB)            
                             crop = cv2.resize(crop, (85, 85))
-                            faces.append([crop, ret[0][1]])
+                            faces.append([crop, face_emb])
+                            pass
+
                     else:
                         print('Bad file', file) 
 
+
         # Add faces[] images to buttons
         delx, dely = 100, 100
 
@@ -1160,7 +1162,10 @@ def toggle_source_faces_buttons_state_shift(self, event, button=-1):
                         temp_holder.append(self.source_faces[j]['Embedding'])
 
                 if temp_holder:
-                    tface['AssignedEmbedding'] = np.median(temp_holder,0)
+                    if self.widget['MergeTextSel'].get()=='Median':
+                        tface['AssignedEmbedding'] = np.median(temp_holder,0)
+                    elif self.widget['MergeTextSel'].get()=='Mean':
+                        tface['AssignedEmbedding'] = np.mean(temp_holder,0)    
 
                 break
 
@@ -1319,8 +1324,8 @@ def load_target(self, button, media_file, media_type):
 
 
         # delete all markers
-        for i in range(len(self.markers)):
-            self.video_slider_canvas.delete('all')
+
+        self.markers_canvas.delete('all')
 
         self.markers = []
         self.stop_marker = []
@@ -1547,16 +1552,18 @@ def save_selected_source_faces(self, text):
         text = text.get()
         # get embeddings from all highlightebuttons
         # iterate through the buttons
-        summed_embeddings = [0]*512
-        num_selected = 0
-        for i in range(len(self.source_faces)):
-            if self.source_faces[i]["ButtonState"]:
-                summed_embeddings += self.source_faces[i]["Embedding"]
-                num_selected += 1
-
-        # create the average embedding
-        if num_selected != 0:
-            ave_embedding = summed_embeddings/num_selected
+
+        temp_holder = []    
+
+        for button in self.source_faces:
+            if button["ButtonState"]:
+                temp_holder.append(button['Embedding'])
+
+        if temp_holder:
+            if self.widget['MergeTextSel'].get()=='Median':
+                ave_embedding = np.median(temp_holder,0)
+            elif self.widget['MergeTextSel'].get()=='Mean':
+                ave_embedding = np.mean(temp_holder,0)    
 
             if text != "":
                 with open("merged_embeddings.txt", "a") as embedfile: