How to detect a face and draw a bounding box around it using OpenCV Python?

Face detection is a fundamental computer vision task that can be accomplished using OpenCV and Haar cascade classifiers. A Haar cascade classifier is an effective machine learning approach for object detection that uses pre-trained XML files to identify specific features like faces.

We will use haarcascade_frontalface_alt.xml as our Haar cascade file for detecting frontal faces in images.

Downloading Haar Cascade Files

OpenCV provides pre-trained Haar cascade files on GitHub ?

https://github.com/opencv/opencv/tree/master/data/haarcascades

To download the face detection cascade, click on haarcascade_frontalface_alt.xml, view it in raw format, then right-click and save to your project directory.

Note: Save all Haar cascade XML files in a haarcascades folder for better organization.

Face Detection Process

The face detection workflow involves these steps ?

• Import the OpenCV library

• Read the input image using cv2.imread()

• Convert the image to grayscale for processing

• Load the Haar cascade classifier with cv2.CascadeClassifier()

• Detect faces using detectMultiScale() which returns coordinates as (x, y, width, height)

• Draw bounding rectangles around detected faces using cv2.rectangle()

• Display the result

Example 1: Single Face Detection

This example demonstrates detecting and drawing a bounding box around one face ?

import cv2

# Read the input image
img = cv2.imread('people.jpg')

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

# Load the Haar cascade classifier
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')

# Detect faces in the image
faces = face_cascade.detectMultiScale(gray, 1.1, 2)
print('Number of detected faces:', len(faces))

# Draw rectangles around detected faces
for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x+w, y+h), (0, 255, 255), 2)

# Display the result
cv2.imshow('Face Detection', img)
cv2.waitKey(0)
cv2.destroyAllWindows()

Input Image:

The output shows the number of detected faces ?

Number of detected faces: 1

Output Image:

The face is highlighted with a yellow bounding box.

Example 2: Multiple Face Detection

This example shows detecting multiple faces in a group photo ?

import cv2

# Read the input image
img = cv2.imread('group_photo.jpg')

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

# Load the Haar cascade classifier
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')

# Detect faces with adjusted parameters for better detection
faces = face_cascade.detectMultiScale(gray, 1.1, 3)
print('Number of detected faces:', len(faces))

# Draw rectangles around all detected faces
for (x, y, w, h) in faces:
    cv2.rectangle(img, (x, y), (x+w, y+h), (0, 255, 255), 2)

# Display the result
cv2.imshow('Multiple Face Detection', img)
cv2.waitKey(0)
cv2.destroyAllWindows()

Input Image:

The output shows multiple detected faces ?

Number of detected faces: 15

Output Image:

All detected faces are enclosed in yellow bounding boxes.

Key Parameters

The detectMultiScale() method accepts important parameters ?

• scaleFactor (1.1): How much the image size is reduced at each scale

• minNeighbors (2-5): How many neighbors each face should have to be considered valid

• minSize: Minimum possible face size (optional)

• maxSize: Maximum possible face size (optional)

Conclusion

OpenCV's Haar cascade classifiers provide an efficient way to detect faces in images. By loading a pre-trained XML file and using detectMultiScale(), you can easily identify and draw bounding boxes around faces, making this technique valuable for various computer vision applications.