电子科技2025，Vol.38Issue(8)：27-32,6.DOI:10.16180/j.cnki.issn1007-7820.2025.08.004
基于Agglomerator特征提取的非迭代路由胶囊网络
A Non-Iterative Routing CapsNet Based on Agglomerator Feature Extraction
摘要
Abstract
In view of the problem of low interpretability of the feature extractor of capsule network,a new fea-ture extractor combining DenseCap(Dense Capsule)and Agglomerator is proposed in this study.By combining the densely connected low-level and high-level features with the local global features of Agglomerator,the adjacent two layers of features correspond to the local and the whole,which improves the interpretability.The parallel connection of DenseCap and Agglomerator makes the model structure more compact and reduces trainable parameters.Dense con-nection of absolute position coding with Agglomerator preserves the advantages of absolute value coding and relative position coding when calculating relative attention,and maintains translational isotropy.The experimental results show that compared with the Capsule Network and the original Agglomerator,the Agg-CapsNet(Agglomerator Cap-sule Network)has better accuracy in terms of CIFAR10,MNIST,Fashion MNIST and SmallNorb.In translation ex-periments of position coding,Agg-CapsNet is proved to have translational isotropy by visualization.关键词
特征提取/胶囊网络/Agglomerator/位置编码/等变性/注意力/局部整体特征/对比学习Key words
feature extraction/capsule network/Agglomerator/position coding/invariance/attention/part and whole feature/contrastive learning分类
信息技术与安全科学引用本文复制引用
倪庭轩,宋燕..基于Agglomerator特征提取的非迭代路由胶囊网络[J].电子科技,2025,38(8):27-32,6.基金项目
国家自然科学基金(62073223) （62073223）
上海市自然科学基金(22ZR1443400) National Natural Science Foundation of China(62073223) （22ZR1443400）
Shanghai Natural Science Foundation(22ZR1443400) （22ZR1443400）
