Face Detection using RaspberryPi and OpenCV

Siva Kumar | June 16, 2017 | DIY electronics, Raspberry Pi | 1 Comment

Face detection is getting important in many places. Knowing the importance and relevance,Almost all major industry vendors are doing something or other in this field.

Amazon – Amazon rekognition
Google – Google vision
Microsoft – Face API

Open source players
OpenCV
OpenFace

Most of these use deep neural network to detect faces.

I made an attempt to make use of OpenCV and Haarcascade calssifier

I split the content into two topics:

Required items and setting up
Running and results

Required items and setting up

Raspberry Pi with Camera connection kit
OpenCV libarary
Pre-trained classifier files for face and eyes
Putty for connecting Raspi to your laptop (or Monitor and keyboard to work on Raspi directly)

Camera connection kit:
Camera module need to connected to raspi using flex cable

Setting up the environment raspberry pi :

Required libraries need to be installed by executing below commands

	sudo apt-get update
	sudo apt-get upgrade
	sudo apt-get install build-essential
	sudo apt-get install cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev
	sudo apt-get install python-dev python-numpy libjpeg-dev libpng-dev libtiff-dev libjasper-dev libdc1394-22-dev
	sudo apt-get install python-opencv
	sudo apt-get install opencv
	sudo apt-get install python-matplotlib

view raw pi_opencv hosted with

by GitHub

Camera need to be enabled in raspi-config

Connection setup :

Confirm camera is working or not by executing this command

sudo raspistill -o filename.jpg

Setting the coding environment:

Python code is used to build the face detection which uses Haar cascade classifier (pre -trained)

	import io
	import picamera
	import cv2
	import numpy

	#Create a memory stream so photos doesn't need to be saved in a file
	stream = io.BytesIO()

	#Get the picture (low resolution, so it should be quite fast)
	#Here you can also specify other parameters (e.g.:rotate the image)
	with picamera.PiCamera() as camera:
	camera.resolution = (320, 240)
	camera.capture(stream, format='jpeg')

	#Convert the picture into a numpy array
	buff = numpy.fromstring(stream.getvalue(), dtype=numpy.uint8)

	#Now creates an OpenCV image
	img = cv2.imdecode(buff, 1)

view raw Part1 of code : Getting the image hosted with

by GitHub

Once we get the image, move on to classifier

	#Load a cascade file for detecting faces
	face_cascade = cv2.CascadeClassifier('/home/pi/Desktop/opencv/haarcascade_frontalface_default.xml')
	eye_cascade = cv2.CascadeClassifier('/home/pi/Desktop/opencv/haarcascade_eye.xml')

	#Convert to grayscale
	gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)

	#Look for faces in the image using the loaded cascade file
	faces = face_cascade.detectMultiScale(gray, 1.3, 5)

	print "Found "+str(len(faces))+" face(s)"

	#Draw a rectangle around every found face

	for (x,y,w,h) in faces:
	cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
	roi_gray = gray[y:y+h, x:x+w]
	roi_color = img[y:y+h, x:x+w]

	eyes = eye_cascade.detectMultiScale(roi_gray)
	for (ex,ey,ew,eh) in eyes:
	cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2)

	#Save the result2 image
	cv2.imwrite('result2.jpg',img)

view raw pi_face_cascade hosted with

by GitHub

Now run the python code, `sudo python facedetect.py`

Result image will be stored in the file.

After identifying /detecting the face, we can compare with known face DB.

Face Recognition can be used in numerous ways as we can provide contextual information to the user

1. Preferred treatment to customer (loyal or otherwise) depending on the customer profile
2. Warning to the user if he enters dangerZone or unauthorized zone
3. Can give specific treatment in case of senior citizens or kids

References
http://docs.opencv.org/…
https://www.hackster.io/…
http://www.bogotobogo.com/…

If one does not have camera , still he can run this program using the stock images from gallery ( in stead of live source)

Tweaked code for that case

	import io
	import numpy
	import cv2
	face_cascade = cv2.CascadeClassifier('/home/pi/Desktop/opencv/haarcascade_frontalface_default.xml')
	eye_cascade = cv2.CascadeClassifier('/home/pi/Desktop/opencv/haarcascade_eye.xml')
	img = cv2.imread('sachin.jpg')
	gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
	faces = face_cascade.detectMultiScale(gray, 1.3, 5)
	for (x,y,w,h) in faces:
	cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
	roi_gray = gray[y:y+h, x:x+w]
	roi_color = img[y:y+h, x:x+w]
	eyes = eye_cascade.detectMultiScale(roi_gray)
	for (ex,ey,ew,eh) in eyes:
	cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2)
	cv2.imwrite('result3.jpg',img)