File size: 3,078 Bytes
0c87db7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
from chain_img_processor import ChainImgProcessor, ChainImgPlugin
from roop.face_helper import get_one_face, get_many_faces, swap_face
import os
from roop.utilities import compute_cosine_distance

modname = os.path.basename(__file__)[:-3] # calculating modname

# start function
def start(core:ChainImgProcessor):
    manifest = { # plugin settings
        "name": "Faceswap", # name
        "version": "1.0", # version

        "default_options": {
            "swap_mode": "selected",
            "max_distance": 0.65, # max distance to detect face similarity
        },
        "img_processor": {
            "faceswap": Faceswap
        }
    }
    return manifest

def start_with_options(core:ChainImgProcessor, manifest:dict):
    pass


class Faceswap(ChainImgPlugin):

    def init_plugin(self):
        pass


    def process(self, frame, params:dict):
        if not "input_face_datas" in params or len(params["input_face_datas"]) < 1:
            params["face_detected"] = False
            return  frame
        
        temp_frame = frame
        params["face_detected"] = True
        params["processed_faces"] = []

        if params["swap_mode"] == "first":
            face = get_one_face(frame)
            if face is None:
                params["face_detected"] = False
                return frame
            params["processed_faces"].append(face)
            frame = swap_face(params["input_face_datas"][0], face, frame) 
            return frame

        else:
            faces = get_many_faces(frame)
            if(len(faces) < 1):
                params["face_detected"] = False
                return frame
            
            dist_threshold = params["face_distance_threshold"]

            if params["swap_mode"] == "all":
                for sf in params["input_face_datas"]:
                    for face in faces:
                        params["processed_faces"].append(face)
                        temp_frame = swap_face(sf, face, temp_frame)
                return temp_frame
            
            elif params["swap_mode"] == "selected":
                for i,tf in enumerate(params["target_face_datas"]):
                    for face in faces:
                        if compute_cosine_distance(tf.embedding, face.embedding) <= dist_threshold:
                            temp_frame = swap_face(params["input_face_datas"][i], face, temp_frame)
                            params["processed_faces"].append(face)
                            break

            elif params["swap_mode"] == "all_female" or params["swap_mode"] == "all_male":
                gender = 'F' if params["swap_mode"] == "all_female" else 'M'
                face_found = False
                for face in faces:
                    if face.sex == gender:
                        face_found = True
                    if face_found:
                        params["processed_faces"].append(face)
                        temp_frame = swap_face(params["input_face_datas"][0], face, temp_frame)
                        face_found = False

        return temp_frame