发现人体动作识别的图网络模型参数过多，提出稀疏化方法在少95%参数下性能几乎不变。
Are Spatial-Temporal Graph Convolution Networks for Human Action Recognition Over-Parameterized?
- 从密集网络中训练出稀疏结构，保留原性能
- 95%参数减少时准确率损失小于1%
- 多层级稀疏模型参数减至66%却提升超1%准确率
基于骨架的人体动作识别（HAR）中，时空图卷积网络（ST-GCNs）表现优异。然而，尽管模型不断演进，统一输入设置后性能差异不大。本文据此假设ST-GCNs对HAR任务存在过参数化，并通过彩票假设实验予以验证。为此提出一种新型稀疏ST-GCN生成器，可从随机初始化的稠密网络中训练出稀疏结构，性能与稠密版本相当。进一步通过融合不同稀疏度的结构，构建多层级稀疏模型，在四个数据集（NTU-RGB+D 60/120、Kinetics-400、FineGYM）上验证：稀疏模型仅用95%更少参数，顶1准确率下降<1%；多层级稀疏模型仅需66%参数，顶1准确率反而提升>1%。代码已开源。
原文摘要 · Abstract (English)
Spatial-temporal graph convolutional networks (ST-GCNs) showcase impressive performance in skeleton-based human action recognition (HAR). However, despite the development of numerous models, their recognition performance does not differ significantly after aligning the input settings. With this observation, we hypothesize that ST-GCNs are over-parameterized for HAR, a conjecture subsequently confirmed through experiments employing the lottery ticket hypothesis. Additionally, a novel sparse ST-GCNs generator is proposed, which trains a sparse architecture from a randomly initialized dense network while maintaining comparable performance levels to the dense components. Moreover, we generate multi-level sparsity ST-GCNs by integrating sparse structures at various sparsity levels and demonstrate that the assembled model yields a significant enhancement in HAR performance. Thorough experiments on four datasets, including NTU-RGB+D 60(120), Kinetics-400, and FineGYM, demonstrate that the proposed sparse ST-GCNs can achieve comparable performance to their dense components. Even with 95% fewer parameters, the sparse ST-GCNs exhibit a degradation of <1% in top-1 accuracy. Meanwhile, the multi-level sparsity ST-GCNs, which require only 66% of the parameters of the dense ST-GCNs, demonstrate an improvement of >1% in top-1 accuracy. The code is available at https://github.com/davelailai/Sparse-ST-GCN.
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。