用统计约束优化动作分割，无需重训练就能提升精度。
Improving Temporal Action Segmentation via Constraint-Aware Decoding
- 引入转移置信度等结构先验，通过改进Viterbi算法进行推理时修正。
- 在多个数据集上提升准确率，半监督设置下最高增益达3.2%。
- 轻量级设计适合低资源场景，特别适合无标注数据少的领域应用。
时间动作分割（TAS）将未修剪视频划分为带标签的动作段。尽管全监督方法已取得进展，但动作变异性、边界模糊及高标注成本等问题仍存在，尤其在新或低资源领域。基于语法的方法虽利用结构先验改善分割，却依赖复杂解析，难以扩展。本文提出一种轻量级、基于约束的精炼框架，通过直接从标注数据中提取的统计结构先验（如转移置信度、动作边界集合、每类持续时间）增强预测。这些约束被整合至改进的Viterbi解码算法中，实现无需重新训练或增加模型复杂度的推理时优化。该方法有效提升了全监督与半监督模型的性能，纠正结构预测错误，同时保持高效性。代码已开源。
原文摘要 · Abstract (English)
Temporal action segmentation (TAS) divides untrimmed videos into labeled action segments. While fully supervised methods have advanced the field, challenges such as action variability, ambiguous boundaries, and high annotation costs remain, especially in new or low-resource domains. Grammar-based approaches improve segmentation with structural priors but rely on complex parsing limiting scalability. In this work, we propose a lightweight, constraint-based refinement framework that enhances TAS predictions by integrating statistical structural priors such as transition confidence, action boundary sets, and per-class duration, that can be directly extracted from annotated data. These constraints are integrated into a modified Viterbi decoding algorithm, allowing inference-time refinement without retraining or added model complexity. Our approach improves both fully and semi-supervised TAS models by correcting structural prediction errors while maintaining high efficiency. Code is available at https://github.com/LUNAProject22/CAD
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。