Integration of optimization methods into image processing and machine learning applications

Optimization is one of the most important and most studied research topics today. It takes place in almost all fields such as engineering, science, energy, computers, etc. and can be found in basic processes for many purposes. It contributes to making the methods used in applications more effective and adaptive. However, due to the complexity of real-world scientific and engineering problems, existing simple optimization algorithms may be insufficient. Therefore, there is a need to develop methods that are more robust and can be successful in solving global optimization problems in many disciplines. In this thesis study, it is aimed to integrate metaheuristic algorithms into image segmentation and machine learning methods and improve the performance of these methods and their applications. In this context, the studies carried out with the developed methods are as follows: Multi-Level Image Thresholding Based on Golden Sine Algorithm II Global Optimization Problems and Chaotic Adversarial Learning Based Golden Sine Algorithm Developed for Breast Cancer Image Segmentation Adaptive Multi-Level Image Thresholding with Chaotically Augmented Rao Algorithm Multi-Level Thresholding Based on Particle Swarm Optimization Multi-level Thresholding Method with PSO Method Based on Visit Table and Multiple Search Strategies Emotion Classification from EEG Signals with Discrete PSO Method Based on Visit Table and Multiple Search Strategies. These developed methods focus on solving global problems in various disciplines. In this context; Experimental studies have been carried out on test functions, classical engineering problems, artificial images, skin and breast cancer images, images with different characteristics taken for various purposes and EEG signals, and the advantages of the proposed methods have been shown in terms of various measurement indices by comparing the methods and their results proposed in the literature.