Advanced Algorithms (COSC 520) - Project
Authors: Ladan Tazik, Laura Quiroga, Emon Sarker

Edge detection is a fundamental technique in image processing and computer vision. This usually corresponds to object boundaries. This project implements and compares numerically three widely used edge detection algorithms: Sobel, Roberts Cross, and Canny.

EdgeDetectionAlgorithms/ ├── src/ # Core algorithm implementations │ ├── sobel.py # Sobel edge detection algorithm │ ├── roberts.py # Roberts Cross edge detection algorithm │ ├── canny.py # Canny edge detection algorithm │ └── utils.py # Helper utilities for image processing ├── tests/ # Unit tests for algorithms │ ├── test_sobel.py │ ├── test_roberts.py │ └── test_canny.py ├── input_images/ # Directory for input images ├── bw_output_images/ # Output directory for processed images │ ├── sobel/ │ ├── roberts/ │ └── canny/ ├── mat_bin_ground_truth/ # Ground truth data for evaluation ├── synthetic_images/ # Synthetic test images ├── templates/ # Web interface templates │ └── index.html └── uploads/ # Temporary storage for web uploads 

• main.py: Demonstrates edge detection algorithms on a sample image, displaying results side by side.

• process_images.py: Batch processes images using all three edge detection algorithms:

  • Takes images from input_images/
  • Applies Sobel, Roberts, and Canny edge detection
  • Saves results to algorithm-specific directories in bw_output_images/

• evaluate_accuracy.py: Evaluates algorithm accuracy against ground truth data:

  • Compares edge detection results with ground truth from BSD500 dataset
  • Calculates precision, recall, F1 score, and Jaccard index
  • Outputs detailed results to accuracy_results.csv

• app.py: Flask web application for interactive edge detection:

  • Allows users to upload images
  • Processes images with all three algorithms
  • Displays original and edge-detected results

• src/sobel.py: Implements the Sobel operator using 3x3 convolution kernels for detecting horizontal and vertical edges.

• src/roberts.py: Implements the Roberts Cross operator using 2x2 convolution kernels for detecting diagonal edges.

• src/canny.py: Implements the Canny edge detection algorithm with five steps:

  1. Gaussian blur for noise reduction
  2. Intensity gradient computation
  3. Non-maximum suppression
  4. Double thresholding
  5. Edge tracking by hysteresis

• convert_to_bw.py: Converts input images to black and white format.

• generate_synthetic_images.py: Creates synthetic test images with known edge patterns.

• process_synthetic_images.py: Processes synthetic images with edge detection algorithms.

• plot_performance.py: Generates performance comparison visualizations.

Follow these steps to run the project:

1. Clone the project

   git clone https://github.com/Lauraquiroga/EdgeDetectionAlgorithms.git cd EdgeDetectionAlgorithms

2. Install dependencies
   Ensure that you have Python installed on your system.
   Create and activate a virtual environment (recommended version for the virtual environment: Python 3.12.7).

Install requirements mentioned in the requirements.txt file.

pip install -r requirements.txt

3. Run the project
   You can use the project in several ways:

• Process a single image:

python main.py

• Batch process multiple images:

python process_images.py

• Run the web interface:

python app.py

Then open http://localhost:5000 in your browser.

• Evaluate algorithm accuracy:

python evaluate_accuracy.py

4. Run the unit tests
   To run the test cases use the following command.

python -m unittest discover -s tests

The Sobel edge detection algorithm was first described by Irwin Sobel and Gary Feldman in 1968 as part of the Stanford Artificial Intelligence Laboratory (SAIL) project:
I. Sobel and G. Feldman, "A 3×3 isotropic gradient operator for image processing," Pattern Classification and Scene Analysis, pp. 271–272, 01 1973.

The Roberts Cross operator was first described in:
Lawrence G. Roberts, "Machine Perception of Three-Dimensional Solids," Ph.D. Dissertation, Massachusetts Institute of Technology, 1963.

The Canny algorithm was first described in:
J. F. Canny, "A Computational Approach to Edge Detection," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. PAMI-8, no. 6, pp. 679-698, Nov. 1986.

Code in Sobel and Roberts classes was adapted from:
adamiao. sobel-filter-tutorial. Accessed: 2025-02-25. 2019. url: https://github.com/adamiao/sobel-filter-tutorial/blob/master/sobel_from_scratch.py
Which is licensed under the GNU General Public License v3.0 (GPL-3.0)