Why face analysis is a core CV application in Computer Vision - Why Metrics Matter

Metrics & Evaluation - Why face analysis is a core CV application

Which metric matters for this concept and WHY

In face analysis, accuracy is important to measure overall correct predictions. But more crucial are precision and recall because:

• Precision tells us how many detected faces are actually correct. This avoids false alarms.
• Recall tells us how many real faces were found. Missing faces can cause problems.

For example, in security, missing a face (low recall) is worse than a false alarm (low precision). So recall is often prioritized.

Confusion matrix or equivalent visualization (ASCII)

      | Predicted Face       | Predicted No Face    |
      |----------------------|----------------------|
      | True Positive (TP)  = 90           | False Negative (FN) = 10           |
      | False Positive (FP) = 5            | True Negative (TN)  = 895          |
    

Here, out of 100 actual faces, 90 were correctly detected (TP), 10 missed (FN). There were 5 false alarms (FP) among 900 non-face images.

Precision vs Recall tradeoff with concrete examples

Imagine a face unlock system on your phone:

• High precision, low recall: The system only unlocks when very sure. It rarely mistakes others for you (few false alarms), but sometimes it won't recognize you (misses).
• High recall, low precision: The system tries to recognize you all the time, but sometimes lets others in by mistake (false alarms).

For security, high precision is critical to avoid unauthorized access. For convenience, high recall helps avoid frustration.

What "good" vs "bad" metric values look like for this use case

• Good: Precision > 0.95, Recall > 0.90, Accuracy > 0.95. This means most faces are found and few false alarms.
• Bad: Precision < 0.70 or Recall < 0.60. Many faces missed or many false detections, making the system unreliable.

Metrics pitfalls (accuracy paradox, data leakage, overfitting indicators)

• Accuracy paradox: If most images have no faces, a model that always says "no face" can have high accuracy but is useless.
• Data leakage: If training and test images overlap, metrics look better but model won't work well in real life.
• Overfitting: Very high training accuracy but low test recall means the model memorized training faces but can't generalize.

Self-check question

Your face detection model has 98% accuracy but only 12% recall on faces. Is it good for production? Why not?

Answer: No, it is not good. The model misses most faces (low recall), which is critical in face analysis. High accuracy is misleading because most images have no faces, so the model just predicts "no face" often.

Key Result

In face analysis, high recall and precision are key to reliably detect faces and avoid false alarms, beyond just accuracy.

Practice

1. Why is face analysis considered a core application in computer vision?

easy

A. Because it is used only for artistic photo filters

B. Because it helps computers recognize people and their emotions

C. Because it replaces all other computer vision tasks

D. Because it only detects objects like cars and trees

Solution

1. Step 1: Understand the purpose of face analysis

   Face analysis helps computers identify who is in an image and what emotions they show.

2. Step 2: Compare options to this purpose

   Only Because it helps computers recognize people and their emotions matches this purpose; others describe unrelated or incorrect uses.

3. Final Answer:

   Because it helps computers recognize people and their emotions -> Option B

4. Quick Check:

   Face analysis = Recognize people and emotions [OK]

Hint: Face analysis = recognizing faces and feelings [OK]

Common Mistakes:

• Thinking face analysis detects all objects
• Believing it replaces all CV tasks
• Assuming it is only for fun filters

2. Which of the following code snippets correctly detects a face using OpenCV in Python?

easy

A. face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')

B. face_cascade = cv2.load('face_detector.xml')

C. face_cascade = cv2.detectFace('haarcascade.xml')

D. face_cascade = cv2.faceDetector('default.xml')

Solution

1. Step 1: Recall correct OpenCV face detection syntax

   OpenCV uses cv2.CascadeClassifier with a valid XML file for face detection.

2. Step 2: Check each option's correctness

   Only face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml') uses the correct function and file name format; others use invalid methods.

3. Final Answer:

   face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml') -> Option A

4. Quick Check:

   Correct OpenCV face detection code = face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml') [OK]

Hint: OpenCV face detection uses CascadeClassifier with XML file [OK]

Common Mistakes:

• Using non-existent cv2 functions
• Wrong method names like detectFace
• Incorrect file names or extensions

3. What will the output be when running this Python code snippet for face detection?

import cv2
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
img = cv2.imread('group_photo.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.1, 4)
print(len(faces))

medium

A. An error because detectMultiScale is misspelled

B. The image itself printed as text

C. Always prints zero regardless of image

D. Number of faces detected in the image

Solution

1. Step 1: Understand detectMultiScale output

   detectMultiScale returns a list of rectangles for detected faces; len(faces) counts them.

2. Step 2: Analyze the print statement

   Printing len(faces) outputs the number of faces found in the image.

3. Final Answer:

   Number of faces detected in the image -> Option D

4. Quick Check:

   len(faces) = number of faces [OK]

Hint: detectMultiScale returns face count with len() [OK]

Common Mistakes:

• Confusing detectMultiScale spelling
• Expecting image data printed
• Assuming zero faces always

4. Identify the error in this face detection code snippet:

import cv2
face_cascade = cv2.CascadeClassifier('haarcascade_frontalface_default.xml')
img = cv2.imread('photo.jpg')
gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.1, 4)
print(len(faces))

medium

A. Incorrect color conversion code; should use COLOR_BGR2GRAY

B. Missing import statement for numpy

C. detectMultiScale function is called with wrong parameters

D. The image file name is invalid

Solution

1. Step 1: Check color conversion code

   OpenCV reads images in BGR format, so conversion should be COLOR_BGR2GRAY, not COLOR_RGB2GRAY.

2. Step 2: Verify other parts

   Other parts like imports, function calls, and file name are correct or assumed valid.

3. Final Answer:

   Incorrect color conversion code; should use COLOR_BGR2GRAY -> Option A

4. Quick Check:

   Color conversion must match image format [OK]

Hint: Use COLOR_BGR2GRAY for OpenCV images [OK]

Common Mistakes:

• Using COLOR_RGB2GRAY instead of COLOR_BGR2GRAY
• Assuming missing numpy import causes error
• Thinking detectMultiScale parameters are wrong

5. You want to build a phone app that unlocks only when it recognizes your face. Which reason best explains why face analysis is essential for this app?

hard

A. It enables the app to detect all objects in the background

B. It helps the app to change the phone wallpaper automatically

C. It allows the app to detect and verify your unique facial features quickly

D. It makes the phone camera take selfies automatically

Solution

1. Step 1: Understand app requirement

   The app must recognize your face to unlock securely and quickly.

2. Step 2: Match face analysis role

   Face analysis identifies unique facial features to verify identity, enabling unlocking.

3. Final Answer:

   It allows the app to detect and verify your unique facial features quickly -> Option C

4. Quick Check:

   Face analysis = verify identity for unlocking [OK]

Hint: Face analysis verifies identity for secure unlocking [OK]

Common Mistakes:

• Confusing face analysis with background detection
• Thinking it changes wallpapers or takes selfies
• Ignoring the verification purpose