Hybridized approach of artificial bee colony algorithm for detection of suspicious brain pattern using magnetic resonance images

摘要

Swarm intelligence based two optimization methods TABC and TABC-FPA has been discussed in this work for the segmentation and extraction of suspicious region from brain MR image. In these methods, TABC classifier uses texture features extracted from suspicious region as the source for classification. To improve the performance of TABC, TABC-FPA is proposed. In TABC-FPA hybridized approach the random local search operation of ABC is replaced by FPA technique of searching in neighborhood positions. Overall accuracy of TABC is 95.03% and that of TABC-FPA is 96.11%. In addition, the average execution time of TABC-FPA is reduced significantly when compared with TABC. The analysis indicates that the overall performance of TABC-FPA is appreciative.