This is a Facial-recognition project which will also Update Face record
As we all know that many Machine-Learnig based project reqiures altest 2-3 similar kind of images to train & of course a lot of time to train all the things BUT here it requires just one good* image of the person and thats all needed for recognition.
Language --> Python=3.8 System --> Nvidia Jetson-nano Operating System --> Ubuntu CSI-Camera (USB Camera is also Perfect)
-
face_recognition install using - pip install face_recognition import using - import face_recognition
-
cv2 install using - pip install opencv-python import using - import cv2
-
os install using - "Usually Preinstalled" import using - import os
-
pandas install using - pip install pandas import using - import pandas as pd / import cvlib
-
datetime install using - pip install datetime import using - from datetime import datetime
-
tkinter install using - pip install tinkter import using - from tkinter import *
-
tesorflow 2.0 (If GPU is used Verson of CUDA can create exceptions, so try to install the same version on tensorflow_gpu) install using - pip install tensorflow (For GPU users use - pip install tensorflow_gpu import using - import tensorflow as tf
-
numpy install using - pip install numpy import using - import numpy as np
-
Harcascade (for Frontal face) download using file give haarcascade_frontalface_default.xml
-
"update" : This function is used to update face records
-
"facee" : Use to recognize face of the person in the frame
-
"name" : this function provide security while updation new face in model by asking username and password from the user.