Activity 14: Image Processing - Discovery Challenge

🎯 Learning Objectives

By completing this activity, you will:

• Image Resizing: Scale images to different sizes using OpenCV resize functions
• Color Space Conversion: Transform images between different color representations (BGR, Gray, HSV, RGB)
• Edge Detection: Apply Canny edge detection to identify object boundaries
• Image Filtering: Use Gaussian blur for image smoothing and noise reduction
• Practical Applications: Understand when to use each image processing technique

🚀 Getting Started (See Results in 30 Seconds!)

Explore the Complete Solution First

Before starting the challenge, explore the complete working solution to understand what you are building toward:

# Open the complete baseline notebook
jupyter notebook v1-baseline-100percent.ipynb

What to observe:

1. Question 1: How to resize large images to smaller dimensions
2. Question 2: How to convert images to different color spaces and apply filters
3. Bonus Challenge: How to upload and process your own images
4. Run all cells (Cell -> Run All) to see all transformations

Start Your Challenge

Once you have explored the baseline, open your challenge notebook:

# Open the student template
jupyter notebook activity-14-image-processing.ipynb

📋 What's Already Working

Your template comes with 60% working code including:

✅ Complete Examples Provided

• File downloads: Automated download of sample images (cats.jpg, outing_large.jpg)
• Library imports: OpenCV and Colab display helpers
• Image loading: Initial cv.imread() calls
• Bonus upload: Google Colab file upload functionality

✅ Supporting Code

• Display functions: Working cv2_imshow() examples
• Helper scripts: resize_image.py and convert_image.py for reference

📝 Your Tasks (2 TODOs to Complete)

Question One: Image Resizing (1 TODO - Easy)

Concepts: cv.resize(), dimension calculation, interpolation methods

What you'll do:

• Calculate new dimensions (20% of original)
• Use cv.resize() with INTER_AREA interpolation
• Display the resized image

Question 2: Image Conversions (1 TODO - Medium)

Concepts: cv.cvtColor(), cv.Canny(), cv.GaussianBlur(), color spaces

What you'll do:

• Convert to grayscale (for single-channel processing)
• Convert to HSV (for color-based segmentation)
• Convert to RGB (for matplotlib compatibility)
• Apply Canny edge detection (for boundary detection)
• Apply Gaussian blur (for smoothing)

✅ Success Criteria

After completing all TODOs, you should be able to:

• ✅ Resize outing_large.jpg to 20% of its original size
• ✅ Convert cats.jpg to grayscale, HSV, and RGB color spaces
• ✅ Apply Canny edge detection to find edges in the image
• ✅ Apply Gaussian blur with kernel size (7,7)
• ✅ Display all processed images in Google Colab

🚀 Extension Challenges

For advanced students who want to go beyond:

1. Custom Scaling: Try different scale factors (50%, 150%, 200%)
2. Multiple Interpolations: Compare INTER_AREA, INTER_LINEAR, INTER_CUBIC
3. Edge Parameters: Experiment with different Canny thresholds (e.g., 50/100, 100/200)
4. Blur Variations: Try different kernel sizes for Gaussian blur
5. Color Segmentation: Use HSV conversion to isolate specific colors
6. Upload Processing: Process your own image through all transformations

📁 Supporting Files

This activity includes:

💡 Tips

• Dimension Order: OpenCV uses (width, height) but image.shape returns (height, width, channels)
• Color Spaces: BGR is OpenCV default; RGB is for matplotlib; HSV is for color segmentation
• Edge Detection: Lower thresholds = more edges; higher thresholds = fewer edges
• Blur Kernel: Must be odd numbers (3,3), (5,5), (7,7), etc.
• Interpolation: Use INTER_AREA for shrinking, INTER_CUBIC for enlarging

🔍 Common Issues

📚 Key Concepts

• cv.resize(img, dim, interpolation): Resizes image to specified dimensions
• cv.cvtColor(img, conversion_code): Converts between color spaces
• cv.Canny(img, threshold1, threshold2): Detects edges using Canny algorithm
• cv.GaussianBlur(img, kernel_size, border_type): Applies Gaussian smoothing
• COLOR_BGR2GRAY: Converts BGR (color) to grayscale (single channel)
• COLOR_BGR2HSV: Converts BGR to Hue-Saturation-Value color space
• COLOR_BGR2RGB: Converts BGR (OpenCV) to RGB (standard)

🎓 Real-World Applications

• Resizing: Thumbnail generation, web optimization, memory management
• Grayscale: Simplified processing, text recognition, speed improvement
• HSV: Color-based object detection, skin detection, image segmentation
• Edge Detection: Object recognition, feature extraction, boundary detection
• Blur: Noise reduction, privacy masking, artistic effects

Estimated Completion Time: 20 minutes

Previous Activity: Activity 13 - OpenCV Basics

Next Activity: Activity 15 - Drawing with OpenCV