In this article, a fairly simple way is mentioned to implement facial recognition system using Python and OpenCV module along with the explanation of the code step by step in the comments.
Before starting we need to install some libraries in order to implement the code. Below you will see the usage of the library along with the code to install it:
-
OpenCV:
OpenCV (Open Source Computer Vision Library) is an open-source computer vision and machine learning software library. which is built to provide a common infrastructure for machine learning algorithms and computer vision. It has thousands of optimized algorithms which can be used different purposes like detecting and recognizing faces, identifying objects and many more. We need it to take pictures using our webcam and some manipulation needed to be done in the image.
To install the library you need to install pip in your system after that you can follow the steps in command prompt:
pip install opencv-contrib-python -
NumPy:
NumPy is the fundamental package for scientific computing in Python which provides a multidimensional array object other mathematical operations can be performed using this but simply speaking we just need it to convert our images into some form of an array so that we can store the model that has been trained.
To install the library you can type a simple line of code in your command shell:
pip install numpy -
Haar Cascade:
Haar Cascade is basically a classifier which is used to detect the objects for which it has been trained for, from the source. The result is an XML file which stores the trained result. If said simply the Haar Cascade is trained by superimposing the positive image over a set of negative images. The training requires a high spec system and a good internet connection and thousands of training images that is why it is carried out in the server. For increasing the efficiency of the results they use high-quality images and increase the number of stages for which the classifier is trained. We need haar cascade frontal face recognizer to detect the face from our webcam.
To download the haar cascade files of different objects you can go the below link:
GitHub: HaarCascades -
Python GUI (tkinter):
Tkinter is a simple GUI module used for implementing fairly simple GUI and helps us to interact with code in a simple way. Though for understanding the code its not important for you to know how it works.
If you wanna learn more about Tkinter click the link below
Python GUI – tkinter
Code: Python implementing to recognize face using GUI
Python3
# importing librariesimport tkinter as tk
from tkinter import Message, Text
import cv2
import os
import shutil
import csv
import numpy as np
from PIL import Image, ImageTk
import pandas as pd
import datetime
import time
import tkinter.ttk as ttk
import tkinter.font as font
from pathlib import Path
window = tk.Tk()
window.title("Face_Recogniser")
window.configure(background='white')
window.grid_rowconfigure(0, weight=1)
window.grid_columnconfigure(0, weight=1)
message = tk.Label(
window, text="Face-Recognition-System",
bg="green", fg="white", width=50,
height=3, font=('times', 30, 'bold'))
message.place(x=200, y=20)
lbl = tk.Label(window, text="No.",
width=20, height=2, fg="green",
bg="white", font=('times', 15, ' bold '))
lbl.place(x=400, y=200)
txt = tk.Entry(window,
width=20, bg="white",
fg="green", font=('times', 15, ' bold '))
txt.place(x=700, y=215)
lbl2 = tk.Label(window, text="Name",
width=20, fg="green", bg="white",
height=2, font=('times', 15, ' bold '))
lbl2.place(x=400, y=300)
txt2 = tk.Entry(window, width=20,
bg="white", fg="green",
font=('times', 15, ' bold '))
txt2.place(x=700, y=315)
# The function below is used for checking# whether the text below is number or not ?def is_number(s):
try:
float(s)
return True
except ValueError:
pass
try:
import unicodedata
unicodedata.numeric(s)
return True
except (TypeError, ValueError):
pass
return False
# Take Images is a function used for creating# the sample of the images which is used for# training the model. It takes 60 Images of# every new user.def TakeImages():
# Both ID and Name is used for recognising the Image
Id = (txt.get())
name = (txt2.get())
# Checking if the ID is numeric and name is Alphabetical
if(is_number(Id) and name.isalpha()):
# Opening the primary camera if you want to access
# the secondary camera you can mention the number
# as 1 inside the parenthesis
cam = cv2.VideoCapture(0)
# Specifying the path to haarcascade file
harcascadePath = "data\haarcascade_frontalface_default.xml"
# Creating the classier based on the haarcascade file.
detector = cv2.CascadeClassifier(harcascadePath)
# Initializing the sample number(No. of images) as 0
sampleNum = 0
while(True):
# Reading the video captures by camera frame by frame
ret, img = cam.read()
# Converting the image into grayscale as most of
# the processing is done in gray scale format
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# It converts the images in different sizes
# (decreases by 1.3 times) and 5 specifies the
# number of times scaling happens
faces = detector.detectMultiScale(gray, 1.3, 5)
# For creating a rectangle around the image
for (x, y, w, h) in faces:
# Specifying the coordinates of the image as well
# as color and thickness of the rectangle.
# incrementing sample number for each image
cv2.rectangle(img, (x, y), (
x + w, y + h), (255, 0, 0), 2)
sampleNum = sampleNum + 1
# saving the captured face in the dataset folder
# TrainingImage as the image needs to be trained
# are saved in this folder
cv2.imwrite(
"TrainingImage\ "+name + "."+Id + '.' + str(
sampleNum) + ".jpg", gray[y:y + h, x:x + w])
# display the frame that has been captured
# and drawn rectangle around it.
cv2.imshow('frame', img)
# wait for 100 milliseconds
if cv2.waitKey(100) & 0xFF == ord('q'):
break
# break if the sample number is more than 60
elif sampleNum > 60:
break
# releasing the resources
cam.release()
# closing all the windows
cv2.destroyAllWindows()
# Displaying message for the user
res = "Images Saved for ID : " + Id + " Name : " + name
# Creating the entry for the user in a csv file
row = [Id, name]
with open('UserDetails\UserDetails.csv', 'a+') as csvFile:
writer = csv.writer(csvFile)
# Entry of the row in csv file
writer.writerow(row)
csvFile.close()
message.configure(text=res)
else:
if(is_number(Id)):
res = "Enter Alphabetical Name"
message.configure(text=res)
if(name.isalpha()):
res = "Enter Numeric Id"
message.configure(text=res)
# Training the images saved in training image folderdef TrainImages():
# Local Binary Pattern Histogram is an Face Recognizer
# algorithm inside OpenCV module used for training the image dataset
recognizer = cv2.face.LBPHFaceRecognizer_create()
# Specifying the path for HaarCascade file
harcascadePath = "data\haarcascade_frontalface_default.xml"
# creating detector for faces
detector = cv2.CascadeClassifier(harcascadePath)
# Saving the detected faces in variables
faces, Id = getImagesAndLabels("TrainingImage")
# Saving the trained faces and their respective ID's
# in a model named as "trainer.yml".
recognizer.train(faces, np.array(Id))
recognizer.save("TrainingImageLabel\Trainer.yml")
# Displaying the message
res = "Image Trained"
message.configure(text=res)
def getImagesAndLabels(path):
# get the path of all the files in the folder
imagePaths = [os.path.join(path, f) for f in os.listdir(path)]
faces = []
# creating empty ID list
Ids = []
# now looping through all the image paths and loading the
# Ids and the images saved in the folder
for imagePath in imagePaths:
# loading the image and converting it to gray scale
pilImage = Image.open(imagePath).convert('L')
# Now we are converting the PIL image into numpy array
imageNp = np.array(pilImage, 'uint8')
# getting the Id from the image
Id = int(os.path.split(imagePath)[-1].split(".")[1])
# extract the face from the training image sample
faces.append(imageNp)
Ids.append(Id)
return faces, Ids
# For testing phasedef TrackImages():
recognizer = cv2.face.LBPHFaceRecognizer_create()
# Reading the trained model
recognizer.read("TrainingImageLabel\Trainer.yml")
harcascadePath = "data\haarcascade_frontalface_default.xml"
faceCascade = cv2.CascadeClassifier(harcascadePath)
# getting the name from "userdetails.csv"
df = pd.read_csv("UserDetails\UserDetails.csv")
cam = cv2.VideoCapture(0)
font = cv2.FONT_HERSHEY_SIMPLEX
while True:
ret, im = cam.read()
gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
faces = faceCascade.detectMultiScale(gray, 1.2, 5)
for(x, y, w, h) in faces:
cv2.rectangle(im, (x, y), (x + w, y + h), (225, 0, 0), 2)
Id, conf = recognizer.predict(gray[y:y + h, x:x + w])
if(conf < 50):
aa = df.loc[df['Id'] == Id]['Name'].values
tt = str(Id)+"-"+aa
else:
Id = 'Unknown'
tt = str(Id)
if(conf > 75):
noOfFile = len(os.listdir("ImagesUnknown"))+1
cv2.imwrite("ImagesUnknown\Image" +
str(noOfFile) + ".jpg", im[y:y + h, x:x + w])
cv2.putText(im, str(tt), (x, y + h),
font, 1, (255, 255, 255), 2)
cv2.imshow('im', im)
if (cv2.waitKey(1) == ord('q')):
break
cam.release()
cv2.destroyAllWindows()
takeImg = tk.Button(window, text="Sample",
command=TakeImages, fg="white", bg="green",
width=20, height=3, activebackground="Red",
font=('times', 15, ' bold '))
takeImg.place(x=200, y=500)
trainImg = tk.Button(window, text="Training",
command=TrainImages, fg="white", bg="green",
width=20, height=3, activebackground="Red",
font=('times', 15, ' bold '))
trainImg.place(x=500, y=500)
trackImg = tk.Button(window, text="Testing",
command=TrackImages, fg="white", bg="green",
width=20, height=3, activebackground="Red",
font=('times', 15, ' bold '))
trackImg.place(x=800, y=500)
quitWindow = tk.Button(window, text="Quit",
command=window.destroy, fg="white", bg="green",
width=20, height=3, activebackground="Red",
font=('times', 15, ' bold '))
quitWindow.place(x=1100, y=500)
window.mainloop() |
Sample Image of the GUI:
Folder Structure For reference :