/SpatialDomainFiltering-2

MATLAB implementation of spatial domain filtering for image restoration. Includes Low-Pass Filtering (LPF) for smoothing, High-Pass Filtering (HPF) for edge enhancement, and Weighted Average Filtering using a Gaussian filter. Demonstrates the effect of different kernel sizes (3x3 to 21x21).

Primary LanguageMATLAB

SpatialDomainFiltering-2

AIM

Perform various spatial domain filtering techniques on a given image. Apply filter operations with different kernel sizes: 3x3, 7x7, 11x11, 15x15, and 21x21.

Techniques Included

  1. Low-Pass Filtering
  2. High-Pass Filtering
  3. Weighted Average Filtering (Gaussian Filter)

Description

This project demonstrates the effects of spatial domain filtering on a grayscale image. By applying low-pass, high-pass, and weighted average filtering, the project visualizes how different kernel sizes influence image restoration and enhancement.

Filters Explained

  • Low-Pass Filtering: Smooths the image by averaging pixel values over a defined neighborhood.
  • High-Pass Filtering: Enhances edges by emphasizing high-frequency components and subtracting low-pass results.
  • Weighted Average Filtering: Smooths the image using a Gaussian filter, where pixel weights decrease with distance.

Directory Structure

project-directory/
├── cameraman.jpg         # Input grayscale image
├── lab06_spatial_filtering.m  # MATLAB script file
├── low_pass_filters.png  # Output for low-pass filtering
├── high_pass_filters.png # Output for high-pass filtering
├── weighted_avg_filters.png # Output for weighted average filtering
├── README.md             # Project documentation

Prerequisites

  1. MATLAB installed on your system.
  2. cameraman.jpg image in the project directory.
  3. Basic knowledge of MATLAB image processing commands.

How to Run

  1. Clone or download this repository.
  2. Place the cameraman.jpg file in the project directory.
  3. Open lab06_spatial_filtering.m in MATLAB.
  4. Run the script.
  5. Check the generated .png files for results.

Results

Generated Outputs

  • Low-Pass Filters:

Screenshot 2025-01-16 210838

  • High-Pass Filters:

Screenshot 2025-01-16 210852

  • Weighted Average Filters:

    Screenshot 2025-01-16 210913

Key Observations

  • Increasing kernel size for low-pass filters results in greater smoothing but blurs details.
  • High-pass filters highlight edges more distinctly with smaller kernel sizes.
  • Gaussian filters provide smooth transitions and are less sensitive to noise.

Author

Rahul Patel
Roll No: UI22EC58