MemoryLane/face_detector.py at main · Spiral-Memory/MemoryLane · GitHub

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
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
import cv2
from facenet_pytorch import MTCNN, InceptionResnetV1
import torch
import os
import numpy as np
from dotenv import load_dotenv
from pymongo import MongoClient
import datetime
import pytz
import text_speech

load_dotenv()
db_url = os.getenv("DB_URL", "mongodb://localhost:27017/")


workers = 0 if os.name == 'nt' else 2

device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
print('Running on device: {}'.format(device))

mtcnn = MTCNN(
    image_size=160, margin=0, min_face_size=20,
    thresholds=[0.6, 0.7, 0.7], factor=0.709, post_process=True,
    device=device
)

resnet = InceptionResnetV1(pretrained='vggface2').eval().to(device)


cam = cv2.VideoCapture(0)
ret, frame = cam.read()
if not ret:
        print("Failed to take frame")

name = ""

client = MongoClient(db_url)
db = client.memory_lane
relatives = db.relatives


try:
    loaded_dict = torch.load("face_embeddings.pt")
except:
    loaded_dict = {}

def updateEmbeddings():
    global loaded_dict
    try:
        loaded_dict = torch.load("face_embeddings.pt")
    except:
        loaded_dict = {}

def take_photo(intent):
    ret, frame = cam.read()
    if not ret:
        print("Failed to take frame")
        return

    frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)

    x_aligned, prob = mtcnn(frame, return_prob=True)
    if x_aligned is not None:
        # print('Face detected with probability: {:8f}'.format(prob))
        aligned = []
        aligned.append(x_aligned)

        # Generate embedding
        aligned = torch.stack(aligned).to(device)
        embeddings = resnet(aligned).detach().cpu()
        distances = []
        # Compare with existing embeddings (Calculate distance)
        for x in loaded_dict:
            for y in loaded_dict[x]:
                tempDistance = []
                tempDistance.append((embeddings - y).norm().item())
            tempDistance = np.array(tempDistance)
            distances.append(tempDistance.mean())

        distances = np.array(distances)

        # Find the closest match

        minIndex = np.argmin(distances)
        if distances[minIndex] > 0.9:
            print("Unknown face")
            text_speech.speak("Sorry, I don't know him")

        else:
            name = list(loaded_dict.keys())[minIndex]
            print("Face Recognized :", name)
            from bson.objectid import ObjectId
            result = relatives.find_one({"_id" : ObjectId(name)})

            if result:

                if result['gender'].lower() == 'male':
                    pronoun = "He"
                elif result['gender'].lower() == 'female':
                    pronoun = "She"
                else:
                    pronoun = "It"

                if intent == 'name':
                    text_speech.speak(pronoun + " is " + result[intent])
                    relatives.update_one({"name" : result['name']}, {"$set" : {"last_meet" : datetime.datetime.utcnow()}})

                elif intent == 'address':
                    text_speech.speak(pronoun + " lives in " + result[intent])
                    relatives.update_one({"name" : result['name']}, {"$set" : {"last_meet" : datetime.datetime.utcnow()}})

                elif intent == 'last_meet':
                    if 'last_meet' in result:
                        ist = pytz.timezone('Asia/Kolkata')
                        ist_now = result[intent].replace(tzinfo=pytz.utc).astimezone(ist)
                        # print("Last time you met was on", ist_now.strftime("%d %B %Y at %H:%M"))
                        text_speech.speak("Last time you met was on " + ist_now.strftime("%d %B %Y at %H:%M"))
                    else:
                        text_speech.speak("I don't remember when you met him the last time")

                elif intent == 'relationship':
                    text_speech.speak(pronoun + " is your " + result[intent])
                    relatives.update_one({"name" : result['name']}, {"$set" : {"last_meet" : datetime.datetime.utcnow()}})


    else:
        print("No face detected")
        text_speech.speak("Sorry, I didn't see any face")

def releaseCam():
    cam.release()