Mohammad Nikzad - Islamic Azad University Science and Research branch, Tehran
Hamid Abrishami Moghaddam - K.N Toosi University of Technology, Tehran, Iran,

This paper introduces a new incremental evolutionary optimization method based on evolutionary group algorithm (EGA). The EGA was presented as an approach to overcome time-consuming drawbacks related to general evolutionary algorithms in large scale content-based image indexing retrieval (CBIR) optimization tasks. Here, we consider another challengeable limitation of usual evolutionary learning and optimization systems: learning in the scale-varying and dynamic environments. Hence, we present a new strategy based on EGA that is enhanced with the ability of incremental learning. Evaluation results on scale-varying and simulated dynamic CBIR systems show that the proposed method can continuously obtain good performance in the presence of environmental or scale changes.

Content-Based Image Indexing and Retrieval,Wavelet Correlogram, Evolutionary Algorithms (EAs),Incremental Learning