用距离回归与动态深度分配，让动作定位更准更快。
Temporal Zoom Networks: Distance Regression and Continuous Depth for Efficient Action Localization
- 用距离回归替代分类，降低边界检测方差3.3到16.7倍。
- 按难易程度动态分配计算量，在THUMOS14上达58.2%平均mAP。
- 适合处理短动作或边界模糊场景，推理效率提升29%。
时序动作定位需兼顾边界精度与计算效率。现有方法对所有时间位置使用统一计算，对简单边界浪费资源，对模糊边界又力不从心。本文提出两项互补创新：边界距离回归（BDR），将基于分类的边界检测改为有符号距离回归，使方差降低3.3–16.7倍；自适应时序精炼（ATR），以连续深度τ∈[0,1]动态分配Transformer层数，集中计算于困难边界附近。在THUMOS14上，本方法实现56.5% [email protected]和58.2%平均mAP@[0.3:0.7]，仅需151G FLOPs，比ActionFormer++少36%的FLOPs（55.7% [email protected]，235G）。相比均匀基线，精度提升2.9%（+1.8%平均mAP，5.4%相对），同时减少24% FLOPs与29%延迟，短动作性能提升尤其显著（+4.2%，8.6%相对）。训练耗时为基线1.29倍，但可通过知识蒸馏降至1.1倍，保持99.5%准确率。贡献包括：(i) 理论支撑的距离建模，信息论分析证明最优方差缩放；(ii) 连续深度分配机制，避免离散路由复杂度；(iii) 在四个数据集上持续提升，增益与边界异质性正相关。
原文摘要 · Abstract (English)
Temporal action localization requires both precise boundary detection and computational efficiency. Current methods apply uniform computation across all temporal positions, wasting resources on easy boundaries while struggling with ambiguous ones. We address this through two complementary innovations: Boundary Distance Regression (BDR), which replaces classification-based boundary detection with signed-distance regression achieving 3.3--16.7$\times$ lower variance; and Adaptive Temporal Refinement (ATR), which allocates transformer depth continuously ($τ\in[0,1]$) to concentrate computation near difficult boundaries. On THUMOS14, our method achieves 56.5\% [email protected] and 58.2\% average mAP@[0.3:0.7] with 151G FLOPs, using 36\% fewer FLOPs than ActionFormer++ (55.7\% [email protected] at 235G). Compared to uniform baselines, we achieve +2.9\% [email protected] (+1.8\% avg mAP, 5.4\% relative) with 24\% fewer FLOPs and 29\% lower latency, with particularly strong gains on short actions (+4.2\%, 8.6\% relative). Training requires 1.29$\times$ baseline FLOPs, but this one-time cost is amortized over many inference runs; knowledge distillation further reduces this to 1.1$\times$ while retaining 99.5\% accuracy. Our contributions include: (i) a theoretically-grounded distance formulation with information-theoretic analysis showing optimal variance scaling; (ii) a continuous depth allocation mechanism avoiding discrete routing complexity; and (iii) consistent improvements across four datasets with gains correlating with boundary heterogeneity.
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。