基于双策略增强SHADE的自适应差分进化算法及其在多阈值图像分割中的应用
A Dual-Strategy Enhanced SHADE Algorithm with Adaptive Differential Evolution for Multi-Threshold Image Segmentation
摘要: 多阈值图像分割在医学影像分析中发挥着重要作用，元启发式算法为其提供了高效的阈值搜索途径。SHADE算法虽具备参数自适应能力，但仍受限于单一模式参数记忆、固定变异策略以及缺乏主动多样性维护机制等结构性约束。本文提出DSESHADE (Dual-Strategy Enhanced SHADE with Adaptive Diversity Maintenance)算法，引入三项改进策略：(1) 双模式自适应参数控制，依据适应度排名将种群划分为探索组与开采组，从两套独立参数中心值进行差异化采样并以EWMA方式更新记忆；(2) Softmax自适应双策略变异池，将DE/current-to-pbest/1与DE/rand/1两种互补策略以Softmax概率动态选择，配合S型精英比例调度与自适应块交叉；(3) 三级递进式多样性维护系统，按种群聚集程度依次施加协方差引导扰动、多样性注入和自适应部分重启。该算法与二维Rényi熵及非局部均值二维直方图相结合，构建面向医学病理图像的多阈值分割模型。在IEEE CEC 2017基准函数集(D = 30)上，DSESHADE以2.241的Friedman平均排名位列8种对比算法首位；消融实验验证了三项策略缺一不可。在乳腺癌病理图像分割实验中，DSESHADE在PSNR、SSIM、FSIM三项指标的全部6个评测组合中均取得最优排名。
Abstract: Multi-threshold image segmentation plays a vital role in medical image analysis, and metaheuristic algorithms offer efficient threshold search pathways for this task. Although SHADE features parameter self-adaptation through its success-history archive, it still suffers from a single-mode parameter memory that treats all individuals identically, reliance on a sole mutation strategy, and the absence of proactive diversity maintenance. This paper proposes the DSESHADE (Dual-Strategy Enhanced SHADE with Adaptive Diversity Maintenance) algorithm, which introduces three targeted improvements: (1) a dual-mode adaptive parameter control that partitions the population into exploration and exploitation groups based on fitness ranking and samples from two independent parameter centres updated via EWMA; (2) a Softmax adaptive dual-strategy mutation pool that probabilistically switches between DE/current-to-pbest/1 and DE/rand/1, combined with sigmoid-scheduled elite ratio and adaptive block crossover; and (3) a three-level progressive diversity maintenance system that applies covariance-guided perturbation, periodic diversity injection, and adaptive partial restart with escalating intensity. DSESHADE is integrated with 2D Rényi entropy and non-local means 2D histogram to form a multi-threshold segmentation model for medical pathological images. Experiments on the IEEE CEC 2017 benchmark suite (D = 30) show that DSESHADE ranks first among eight compared algorithms with a Friedman mean rank of 2.241. Ablation studies confirm the indispensability of all three strategies. In the breast cancer pathology image segmentation experiment, DSESHADE achieved the best ranking in all six evaluation combinations of the three indicators: PSNR, SSIM, and FSIM.
文章引用：程鑫, 陈虹羽, 陈雨涵. 基于双策略增强SHADE的自适应差分进化算法及其在多阈值图像分割中的应用[J]. 计算机科学与应用, 2026, 16(4): 511-522. https://doi.org/10.12677/csa.2026.164149

参考文献

[1] Otsu, N. (1979) A Threshold Selection Method from Gray-Level Histograms. IEEE Transactions on Systems, Man, and Cybernetics, 9, 62-66. [Google Scholar] [CrossRef] 
[2] Kapur, J.N., Sahoo, P.K. and Wong, A.K.C. (1985) A New Method for Gray-Level Picture Thresholding Using the Entropy of the Histogram. Computer Vision, Graphics, and Image Processing, 29, 273-285. [Google Scholar] [CrossRef] 
[3] Sahoo, P.K. and Arora, G. (2004) A Thresholding Method Based on Two-Dimensional Renyi’s Entropy. Pattern Recognition, 37, 1149-1161. [Google Scholar] [CrossRef] 
[4] Hammouche, K., Diaf, M. and Siarry, P. (2010) A Comparative Study of Various Meta-Heuristic Techniques Applied to the Multilevel Thresholding Problem. Engineering Applications of Artificial Intelligence, 23, 676-688. [Google Scholar] [CrossRef] 
[5] Houssein, E.H., Mohamed, G.M., Djenouri, Y., Wazery, Y.M. and Ibrahim, I.A. (2024) Nature Inspired Optimization Algorithms for Medical Image Segmentation: A Comprehensive Review. Cluster Computing, 27, 14745-14766. [Google Scholar] [CrossRef] 
[6] Storn, R. and Price, K. (1997) Differential Evolution—A Simple and Efficient Heuristic for Global Optimization over Continuous Spaces. Journal of Global Optimization, 11, 341-359. [Google Scholar] [CrossRef] 
[7] Bilal, Pant, M., Zaheer, H., Garcia-Hernandez, L. and Abraham, A. (2020) Differential Evolution: A Review of More than Two Decades of Research. Engineering Applications of Artificial Intelligence, 90, Article 103479. [Google Scholar] [CrossRef] 
[8] Tanabe, R. and Fukunaga, A. (2013) Success-History Based Parameter Adaptation for Differential Evolution. 2013 IEEE Congress on Evolutionary Computation, Cancun, 20-23 June 2013, 71-78. [Google Scholar] [CrossRef] 
[9] Tanabe, R. and Fukunaga, A.S. (2014) Improving the Search Performance of SHADE Using Linear Population Size Reduction. 2014 IEEE Congress on Evolutionary Computation (CEC), Beijing, 6-11 July 2014, 1658-1665. [Google Scholar] [CrossRef] 
[10] Aranguren, I., Valdivia, A., Morales-Castañeda, B., Oliva, D., Abd Elaziz, M. and Perez-Cisneros, M. (2021) Improving the Segmentation of Magnetic Resonance Brain Images Using the LSHADE Optimization Algorithm. Biomedical Signal Processing and Control, 64, Article 102259. [Google Scholar] [CrossRef] 
[11] Zhang, Y., Chen, G. and Cheng, L. (2023) Improving SHADE with a Linear Reduction P Value and a Random Jumping Strategy. In: Huang, DS., Premaratne, P., Jin, B., Qu, B., Jo, KH. and Hussain, A. Eds., Lecture Notes in Computer Science, Springer, 47-58. [Google Scholar] [CrossRef] 
[12] Lin, X. and Meng, Z. (2024) An Adaptative Differential Evolution with Enhanced Diversity and Restart Mechanism. Expert Systems with Applications, 249, Article 123634. [Google Scholar] [CrossRef] 
[13] Mittal, H. and Saraswat, M. (2018) An Optimum Multi-Level Image Thresholding Segmentation Using Non-Local Means 2D Histogram and Exponential Kbest Gravitational Search Algorithm. Engineering Applications of Artificial Intelligence, 71, 226-235. [Google Scholar] [CrossRef] 
[14] Hernandez del Rio, A.A., Cuevas, E. and Zaldivar, D. (2020) Multi-Level Image Thresholding Segmentation Using 2D Histogram Non-Local Means and Metaheuristics Algorithms. In: Oliva, D. and Hinojosa, S., Eds., Studies in Computational Intelligence, Springer International Publishing, 121-149. [Google Scholar] [CrossRef] 
[15] Zhang, J. and Sanderson, A.C. (2009) JADE: Adaptive Differential Evolution with Optional External Archive. IEEE Transactions on Evolutionary Computation, 13, 945-958. [Google Scholar] [CrossRef] 
[16] Jiang, C., Yang, W., Guo, Y., Wu, F. and Tang, Y. (2018) Nonlocal Means Two Dimensional Histogram-Based Image Segmentation via Minimizing Relative Entropy. Entropy, 20, Article 827. [Google Scholar] [CrossRef] [PubMed]
[17] Rényi, A. (1961) On Measures of Entropy and Information. Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics and Probability, 4, 547-562.
[18] Ghosh, A., Das, S., Das, A.K., Senkerik, R., Viktorin, A., Zelinka, I., et al. (2022) Using Spatial Neighborhoods for Parameter Adaptation: An Improved Success History Based Differential Evolution. Swarm and Evolutionary Computation, 71, Article 101057. [Google Scholar] [CrossRef] 
[19] Hansen, N. (2016) The CMA Evolution Strategy: A Tutorial. arXiv:1604.00772.
[20] Tizhoosh, H.R. (2005) Opposition-Based Learning: A New Scheme for Machine Intelligence. International Conference on Computational Intelligence for Modelling, Control and Automation and International Conference on Intelligent Agents, Web Technologies and Internet Commerce (CIMCA-IAWTIC’06), Vienna, 28-30 November 2005, 695-701. [Google Scholar] [CrossRef] 
[21] Awad, N.H., Ali, M.Z., Liang, J.J., Qu, B.Y. and Suganthan, P.N. (2016) Problem Definitions and Evaluation Criteria for the CEC 2017 Special Session and Competition on Single Objective Real-Parameter Numerical Optimization. Technical Report, Nanyang Technological University.
[22] Kennedy, J. and Eberhart, R. (1995) Particle Swarm Optimization. Proceedings of ICNN’95-International Conference on Neural Networks, Perth, 27 November 1995-1 December 1995, 1942-1948. [Google Scholar] [CrossRef] 
[23] Xia, X., Gui, L. and Zhan, Z. (2018) A Multi-Swarm Particle Swarm Optimization Algorithm Based on Dynamical Topology and Purposeful Detecting. Applied Soft Computing, 67, 126-140. [Google Scholar] [CrossRef] 
[24] Liang, J.J., Qin, A.K., Suganthan, P.N. and Baskar, S. (2006) Comprehensive Learning Particle Swarm Optimizer for Global Optimization of Multimodal Functions. IEEE Transactions on Evolutionary Computation, 10, 281-295. [Google Scholar] [CrossRef] 
[25] Brest, J., Zumer, V. and Maucec, M.S. (2006) Self-Adaptive Differential Evolution Algorithm in Constrained Real-Parameter Optimization. 2006 IEEE International Conference on Evolutionary Computation, Vancouver, 16-21 July 2006, 215-222. [Google Scholar] [CrossRef]