Name: Date:
Lesson 7: Computer Vision Basics
Directions: Complete the diagram below showing the four main steps in how AI recognizes images. Use the word bank to help you fill in the blanks.
Word Bank: pixels, pattern matching, classification, feature extraction, training data, confidence score, camera
┌─────────────────────────────────────────────────────────────────────────┐
│ │
│ STEP 1: IMAGE INPUT │
│ ┌─────────────────────────────────────────────────────────────────┐ │
│ │ │ │
│ │ A _________________ or phone captures an image │ │
│ │ │ │
│ │ The image is broken down into tiny colored squares called │ │
│ │ │ │
│ │ _________________ with RGB (Red, Green, Blue) values │ │
│ │ │ │
│ └─────────────────────────────────────────────────────────────────┘ │
│ ↓ │
│ STEP 2: FEATURE EXTRACTION │
│ ┌─────────────────────────────────────────────────────────────────┐ │
│ │ │ │
│ │ AI analyzes the pixel patterns to identify important features: │ │
│ │ │ │
│ │ • Edges and boundaries: _________________________________ │ │
│ │ │ │
│ │ • Colors and textures: __________________________________ │ │
│ │ │ │
│ │ • Shapes and proportions: _______________________________ │ │
│ │ │ │
│ │ This step is called ______________________ because the AI │ │
│ │ is pulling out the most important characteristics. │ │
│ │ │ │
│ └─────────────────────────────────────────────────────────────────┘ │
│ ↓ │
│ STEP 3: PATTERN MATCHING │
│ ┌─────────────────────────────────────────────────────────────────┐ │
│ │ │ │
│ │ AI compares the extracted features to patterns it learned │ │
│ │ from ______________________ (labeled example images) │ │
│ │ │ │
│ │ It looks for similarities between current image and stored │ │
│ │ patterns from training. │ │
│ │ │ │
│ │ The process of comparing patterns is called │ │
│ │ ______________________. │ │
│ │ │ │
│ └─────────────────────────────────────────────────────────────────┘ │
│ ↓ │
│ STEP 4: CLASSIFICATION AND OUTPUT │
│ ┌─────────────────────────────────────────────────────────────────┐ │
│ │ │ │
│ │ AI assigns a label to the image. This is called │ │
│ │ ______________________. │ │
│ │ │ │
│ │ Example output: │ │
│ │ Label: "Dog" │ │
│ │ ______________________ : 95% │ │
│ │ (This shows how certain the AI is about its prediction) │ │
│ │ │ │
│ └─────────────────────────────────────────────────────────────────┘ │
│ │
└─────────────────────────────────────────────────────────────────────────┘1. What is a pixel?
2. Every pixel has three values. What do they represent?
R: G: B:
3. Complete the table with RGB values:
| Color | R (0-255) | G (0-255) | B (0-255) |
|---|---|---|---|
| Pure Red | 255 | _____ | _____ |
| Pure Green | _____ | 255 | _____ |
| Pure Blue | _____ | _____ | 255 |
| White | 255 | _____ | _____ |
| Black | 0 | _____ | _____ |
4. A standard HD image is 1920 pixels wide and 1080 pixels tall. How many total pixels is that?
Calculation: 1920 × 1080 =
5. If each pixel needs 3 numbers (RGB), how many total numbers does the computer need to store one HD image?
Calculation: × 3 =
1. Explain in your own words: What is training data?
2. Why does AI need thousands of example images to learn?
3. Circle the best training dataset for teaching AI to recognize dogs:
Option A: 1,000 images of golden retrievers only
Option B: 50 images of different dog breeds in various settings
Option C: 5,000 images of 100+ dog breeds in different lighting, angles, and backgrounds
Explain your choice:
4. What could happen if AI is trained only on limited examples?
During the hands-on activity, you trained your own AI model. Answer these questions about your experience:
1. What objects did you train your AI to recognize?
Object 1: Object 2:
Object 3: (if applicable)
2. How many training images did you take for each object?
Object 1: ______ images Object 2: ______ images Object 3: ______ images
3. Did your AI recognize the objects correctly most of the time?
□ Yes, almost always □ Sometimes □ Not usually □ Rarely
4. What was the highest confidence score your AI achieved?
Object: Confidence Score: ________%
5. Describe one time your AI made a mistake or got confused:
6. Why do you think the AI made that mistake?
7. What did you do to help your AI perform better?
8. If you could train it again, what would you do differently?
Compare and contrast how humans and AI "see" images. Write at least three differences:
| Human Vision | Computer Vision |
|---|---|
Which is better at each task? Write "Human" or "AI":
List three real-world uses of image recognition AI that you learned about today:
1.
How it helps:
2.
How it helps:
3.
How it helps:
Brainstorm your own idea: Think of a problem in your school or community that image recognition AI could help solve.
Problem:
How AI image recognition could help:
1. Facial recognition can identify people in photos and videos. List two benefits and two concerns:
Benefits:
•
•
Concerns:
•
•
2. AI can be biased if training data isn't diverse. Explain what this means in your own words:
3. Your Opinion: Should schools use facial recognition for attendance? Why or why not?
□ Yes □ No □ Maybe/It depends
Because:
4. What's one rule you think should exist about how facial recognition technology is used?
Match each term with its correct definition:
Terms:
A. Pixel
B. Pattern Recognition
C. Training Data
D. Classification
E. Computer Vision
F. Feature Extraction
G. Confidence Score
H. Machine Learning
Definitions:
_____ 1. The field of AI that enables computers to interpret visual information
_____ 2. Tiny colored square that makes up a digital image with RGB values
_____ 3. Set of labeled example images used to teach AI what to look for
_____ 4. Process of finding recurring features or arrangements in data
_____ 5. Assigning a label or category to an image based on its features
_____ 6. Identifying important characteristics in an image like edges, colors, shapes
_____ 7. Percentage showing how certain AI is about its prediction
_____ 8. AI systems that learn patterns from examples rather than following rigid rules
The "Muffin or Chihuahua" Problem:
AI sometimes confuses photos of muffins (blueberry or chocolate chip) with photos of chihuahua dogs. They can look surprisingly similar!
1. Why do you think this happens? What features do muffins and chihuahuas have in common?
2. What could be done to help AI tell them apart more reliably?
3. Can you think of other pairs of objects that might confuse AI? List two:
•
•
Answer these final questions to reflect on your learning:
1. On a scale of 1-10, how well do you understand how AI processes images?
Before this lesson: ______ After this lesson: ______
2. What was the most interesting thing you learned today?
3. What's one question you still have about computer vision or AI?
4. How will what you learned today change how you think about AI?
5. Would you like to learn more about this topic?
□ Yes, definitely! □ Maybe □ Not really □ No thanks
If yes, what specific aspect interests you most?
Design Your Own Image Classifier:
If you could create any image recognition AI system, what would it do?
Name of your AI system:
What would it identify or classify?
Who would use it?
How would it help them?
What challenges might you face building it?
Draw a simple diagram or sketch of how your AI system would work:
[Large blank space for drawing]
Teacher Use Only:
□ Worksheet completed □ Diagram accurate □ Reflections thoughtful □ Vocabulary correct
Notes:
Overall Assessment: □ Exceeds Expectations □ Meets Expectations □ Approaching □ Needs Support
Evolve AI Institute • Lesson 7: How AI Sees Images
Page 1 of 6 (adjust page numbers based on actual printing)