A Dynamic Self-Adaptive Music-Inspired Optimization Algorithm for the Hippocampus Localization in Histological Images: A Preliminary Study

The hippocampus is a structure in the medial temporal lobe of the brain that is involved in episodic memory function. The texture features of the hippocampus could give better differentiation between Alzheimer’s disease and normal controls. The localization of the hippocampus structure in MRI histological images is considered as a multimodal global continuous optimization problem, which is solved by means of soft computing techniques using stochastic global optimization methods. Recently, the harmony search (HS) algorithm, a music-inspired optimization method, was introduced as a new soft computing rival. However, the overall performance of this algorithm is quite sensitive to the proper settings of its parameters prior to starting the optimization process. Many have proposed HS-based variants that promote self-adaptive parameter settings. In this paper we propose a new HS-based algorithm with dynamic and self-adaptive features. Since this work represents an early step prior to considering a full implementation on actual biomedical images, the proposed algorithm is tested using a multimodal global continuous optimization benchmarking problems rather than actual hippocampus biomedical images. Results demonstrate the superiority of the proposed algorithm against many other HS-based competing methods.