Freedom Sale
Independence Day Special — Unlock the AI Path 70% off our most popular AI course · Limited time offer
--Days
--Hrs
--Min
--Sec
Claim Your Discount
IT Expert Training Learn
Learn › Tutorials
✦ Beginner ⏱ 30 min

👁️ Build a Face Detection App with OpenCV

🎯 What You'll Build

Real-time face detection that draws bounding boxes around faces — works on webcam video, images, and saved files. No deep learning or GPU needed.

📋 What You'll Need

1

Detect faces in a photo

Load an image, convert to greyscale, and run the Haar Cascade detector in three lines.

import cv2

# Load the pre-trained face detector (built into OpenCV)
face_cascade = cv2.CascadeClassifier(
    cv2.data.haarcascades + 'haarcascade_frontalface_default.xml'
)

img  = cv2.imread('people.jpg')
grey = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

# Detect faces
faces = face_cascade.detectMultiScale(grey, scaleFactor=1.1, minNeighbors=5, minSize=(30,30))

print(f"Detected {len(faces)} face(s)")

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

cv2.imwrite('faces_detected.jpg', img)
print("Saved to faces_detected.jpg")
2

Live face detection from webcam

Capture video frames, detect faces in each frame, and display the result in a window.

import cv2

face_cascade = cv2.CascadeClassifier(
    cv2.data.haarcascades + 'haarcascade_frontalface_default.xml'
)

cap = cv2.VideoCapture(0)  # 0 = default webcam
print("Starting face detection — press Q to quit")

while True:
    ret, frame = cap.read()
    if not ret: break

    grey  = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(grey, 1.1, 5, minSize=(30,30))

    for (x, y, w, h) in faces:
        cv2.rectangle(frame, (x,y), (x+w, y+h), (0, 255, 0), 2)
        cv2.putText(frame, 'Face', (x, y-10),
                    cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0,255,0), 2)

    cv2.putText(frame, f'{len(faces)} face(s)', (10, 30),
                cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 255, 255), 2)

    cv2.imshow('Face Detection — press Q to quit', frame)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()
3

Blur detected faces for privacy

Replace each face region with a Gaussian blur to anonymise people in photos.

import cv2

def blur_faces(image_path, output_path, blur_strength=31):
    face_cascade = cv2.CascadeClassifier(
        cv2.data.haarcascades + 'haarcascade_frontalface_default.xml'
    )
    img  = cv2.imread(image_path)
    grey = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    faces = face_cascade.detectMultiScale(grey, 1.1, 5, minSize=(30,30))

    for (x, y, w, h) in faces:
        face_region = img[y:y+h, x:x+w]
        blurred     = cv2.GaussianBlur(face_region, (blur_strength, blur_strength), 0)
        img[y:y+h, x:x+w] = blurred

    cv2.imwrite(output_path, img)
    print(f"Blurred {len(faces)} face(s) — saved to {output_path}")

blur_faces('group_photo.jpg', 'group_photo_anonymised.jpg')
💡 Tip: scaleFactor=1.1 means the image is reduced by 10% at each scale step. Lower values find smaller faces but take longer. minNeighbors=5 means a face must be detected in 5 overlapping regions to count — higher values reduce false positives.

🎉 You Did It!

A three-step script that handles real-time video, static images, and privacy blurring. The Haar Cascade is a 2001 algorithm that still works well for frontal face detection. For better accuracy at angles, use cv2.dnn.readNetFromTensorflow() with a deep learning model.

Browse More Tutorials

Found something wrong?

Spotted a bug, broken code, or something that doesn't look right? Tell us what's off and we'll fix it.