改进狮子群优化算法在图像分割中的研究

打开文本图片集

中图分类号：TP391.41 文献标志码：B

Research on improving lion swarm optimization algorithm in image segmentation

WU Qian

（School of Information Engineering，Hefei Information Technology Vocational Colege，Hefei 23O601，China）

Abstract： To address the issue of the traditional lion swarm optimization（LSO） algorithm easily falling into local optima in multi-objective image segmentation scenarios，an improved LSO algorithm is proposed. The position update functions of lionesses and cubs are dynamically adjusted to enhance both global and local search capabilities.Furthermore，concepts from the artificial bee colony（ABC） algorithm is incorporated to further improve population diversity and robustness.In the improved LSO algorithm，the number of times each individual gets trapped in a local optimum is recorded，and a decay mechanism is applied in subsequent iterations to accelerate convergence toward the global optimum. The results show that the improved LSO algorithm is more effective at escaping local optima and locating the global optimum. The improved LSO algorithm is evaluated against multiple benchmark algorithms on several sets of test images，using both subjective visual assessment and objective quantitative metrics. The improved LSO algorithm consistently outperforms the traditional LSO and other comparative algorithms in terms of segmentation accuracy，convergence speed，and stability.

Key words： lion swarm optimization（LSO） algorithm；multi-objective scenario； image segmentation;artificial bee colony（ABC） algorithm；global optimal solution

0 引言

图像分割在数字图像处理领域发展中扮演着重要角色[1-2]。（剩余7324字）