视觉传达设计中基于改进SFLA算法的动画人物生成研究OA

To enhance the intelligence and refinement of animation character generation and address issues such as low efficiency and limited diversity of traditional methods,this study proposes an animation character generation model based on an improved Shuffled Frog Leaping Algorithm(SFLA).By introducing adaptive mutation and hy-brid selection mechanisms and integrating Generative Adversarial Networks(GANs)to improve image realism and stylistic consistency,we constructed a multi-objective fitness function incorporating structural and semantic con-straints to drive population evolution.Experimental results demonstrate that in super-resolution tasks,the model achieves a generation quality of 32.15,structural fidelity of 0.812 5,and visual quality of 4.5.In online sketch col-oring tasks,the generation quality reaches 45.67 with a visual quality of 4.6.In practical applications,image genera-tion matching rate as high as 99%is achieved,with system stutter rates as low as 21%.These findings indicate that the proposed model enhances the efficiency,quality,and stability of animation character generation,providing reli-able automated support for pre-production animation and virtual character design,thereby advancing visual communi-cation design toward intelligent development.