# Ant Colony Optimization for Image Processing
## Thresholding Methods
- Single Thresholding
- Multilevel Thresholding
- Multi-threshold
## ACO Process
- Ant Initialization
- Pheromone Updates
- Probability Movement
## Proposed Method
- ACO + Thresholding
- Comparison with Otsu/Kittler
## Results
- DIBCO 2009 Benchmark
- Better PSNR Performance
USING_AN_ANT_COLONY_OPTIMIZATION_ALGORITHM
-
https://www.pirahansiah.com/contents/publications/Journals/USING_AN_ANT_COLONY_OPTIMIZATION_ALGORITHM/
-
Ant Colony Optimization for Image Edge Detection
1. Introduction
- Thresholding: Used in various computer vision applications like OCR, image segmentation, and object tracking.
- Ant Colony Optimization (ACO): Population-based metaheuristic for optimization.
- Objective: Combining ACO, edge detection, and thresholding for Optical Character Recognition (OCR) systems.
2. State of the Art
2.1 Thresholding Methods
- Categories: Single, Multilevel, Multi-thresholding
- Single Thresholding: Converts the image into binary (black and white).
- Pirahansiah’s Single Threshold Method: A custom single threshold method using PSNR.
- Multilevel Thresholding: Separates objects based on gray values using multiple thresholds.
- Multi-threshold: Uses multiple threshold values to identify objects in images.
2.2 Ant Colony Optimization (ACO)
- Introduction: Initially proposed by Marco Dorigo in 1992 for combinatorial optimization problems.
- Application: Used for image edge detection in this paper.
- Process:
- Initialize ants randomly.
- Move ants based on probability and pheromone updates.
- Update pheromone values for optimization.
3. Proposed Method
- Combining ACO and Thresholding: ACO is applied to enhance image thresholding in OCR systems.
- Comparison: The proposed method is compared with Otsu, Kittler, Illingworth, and Pirahansiah’s methods.
4. Results and Discussion
- Datasets: DIBCO 2009 benchmark, including printed and handwritten images.
- Performance: The proposed ACO-based method shows better PSNR results for thresholding compared to traditional methods.
- Comparison Results: The ACO method outperforms others in printed and handwritten datasets.
5. Conclusion
- Effective for OCR: The ACO-based thresholding method improves the edge detection and thresholding for OCR systems.
- Future Work: Optimizing ACO parameters for better performance in different types of images.