Python | Haar Cascades for Object Detection

Object Detection is a computer technology related to computer vision, image processing and deep learning that deals with detecting instances of objects in images and videos. We will do object detection in this article using something known as haar cascades.

What are Haar Cascades?

Haar Cascade classifiers are an effective way for object detection. This method was proposed by Paul Viola and Michael Jones in their paper Rapid Object Detection using a Boosted Cascade of Simple Features .Haar Cascade is a machine learning-based approach where a lot of positive and negative images are used to train the classifier. 

• Positive images – These images contain the images which we want our classifier to identify.
• Negative Images – Images of everything else, which do not contain the object we want to detect.

Requirements:

• Make sure you have python, Matplotlib and OpenCV installed on your pc (all the latest versions).
• The haar cascade files can be downloaded from the OpenCV Github repository.

Implementation

Python3

# Importing all required packages
import cv2
import numpy as np
import matplotlib.pyplot as plt % matplotlib inline


# Read in the cascade classifiers for face and eyes
face_cascade = cv2.CascadeClassifier('../DATA / haarcascades / haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('../DATA / haarcascades / haarcascade_eye.xml')



# create a function to detect face
def adjusted_detect_face(img):

    face_img = img.copy()

    face_rect = face_cascade.detectMultiScale(face_img,
                                              scaleFactor = 1.2,
                                              minNeighbors = 5)

    for (x, y, w, h) in face_rect:
        cv2.rectangle(face_img, (x, y),
                      (x + w, y + h), (255, 255, 255), 10)\

    return face_img


# create a function to detect eyes
def detect_eyes(img):

    eye_img = img.copy()
    eye_rect = eye_cascade.detectMultiScale(eye_img,
                                            scaleFactor = 1.2,
                                            minNeighbors = 5)
    for (x, y, w, h) in eye_rect:
        cv2.rectangle(eye_img, (x, y),
                      (x + w, y + h), (255, 255, 255), 10)
    return eye_img

# Reading in the image and creating copies
img = cv2.imread('../sachin.jpg')
img_copy1 = img.copy()
img_copy2 = img.copy()
img_copy3 = img.copy()

# Detecting the face
face = adjusted_detect_face(img_copy)
plt.imshow(face)
# Saving the image
cv2.imwrite('face.jpg', face)

Code : Detecting the eyes

Python3

eyes = detect_eyes(img_copy2)
plt.imshow(eyes)
cv2.imwrite('face_eyes.jpg', eyes)

Code : Detecting face and eyes

Python3

eyes_face = adjusted_detect_face(img_copy3)
eyes_face = detect_eyes(eyes_face)
plt.imshow(eyes_face)
cv2.imwrite('face+eyes.jpg', eyes_face)

Haar Cascades can be used to detect any types of objects as long as you have the appropriate XML file for it. You can even create your own XML files from scratch to detect whatever type of object you want.