通过编码时间、幅度与结构对称性,提升惯性传感器动作识别的鲁棒性。
Learning with Category-Equivariant Architectures for Human Activity Recognition
- 构建对称性范畴,统一处理周期时间偏移、增益缩放与传感器层级关系。
- 在UCI-HAR数据集上,对分布外扰动的鲁棒性显著优于循环填充CNN和普通CNN。
- 无需增加模型容量即可实现强不变性与泛化能力,适合高可靠性动作识别场景。
我们提出CatEquiv,一种用于惯性传感器人体动作识别(HAR)的类别等变神经网络,系统地编码了时间、幅度和结构对称性。引入一个联合表示循环时间平移、正增益缩放以及传感器层次偏序集的对称性范畴,捕捉数据的类别对称结构。CatEquiv实现了对该类别对称积的等变性。在UCI-HAR数据集上,面对分布外扰动时,其表现明显优于循环填充卷积神经网络和普通卷积神经网络。结果表明,强制执行类别对称性可实现强大的不变性和泛化能力,且无需额外模型容量。
原文摘要 · Abstract (English)
We propose CatEquiv, a category-equivariant neural network for Human Activity Recognition (HAR) from inertial sensors that systematically encodes temporal, amplitude, and structural symmetries. We introduce a symmetry category that jointly represents cyclic time shifts, positive gain scalings, and the sensor-hierarchy poset, capturing the categorical symmetry structure of the data. CatEquiv achieves equivariance with respect to the categorical symmetry product. On UCI-HAR under out-of-distribution perturbations, CatEquiv attains markedly higher robustness compared with circularly padded CNNs and plain CNNs. These results demonstrate that enforcing categorical symmetries yields strong invariance and generalization without additional model capacity.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。