发表机构
WSC Sports（WSC体育）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究揭示时间笑声定位中标注者分歧的结构化特性，提出基于共形校准容差带的分歧校准评估方法，以更可靠地评测系统性能。
AI 中文摘要
标注者在笑声边界和细微轻笑上经常存在分歧，然而时间笑声定位通常仅针对单一参考标注进行评估。我们表明，这种分歧是结构化的而非随机噪声。我们使用每个视频3-5名标注者（alpha = 0.757）重新标注了SMILE-Temporal基准（672个视频，1,683个事件），发现了系统性模式：在偏移处（offset）的分歧是在起始处（onset）的1.73倍，轻笑（chuckle）比完整大笑（full laugh）更常产生分歧（77%对比20%），并且分歧可从事件属性预测（AUC = 0.831）。在这种结构下，针对单一标注者进行评估会失效：系统得分根据所选的真实标注（ground truth）变化0.246 F1，正确排序系统的概率仅为69.7%（对比针对所有标注者时的80%）。我们提出一种分歧校准的评估方法，使用共形校准的容差带（在偏移处更宽，0.727秒，起始处为0.5秒）针对完整标注者分布对预测进行评分。每个标注者的标注和分析代码可在以下网址获取：https://this https URL。
英文摘要
Annotators routinely disagree on laughter boundaries and subtle chuckles, yet temporal laughter localization typically evaluates against a single reference annotation. We show that this disagreement is structured rather than random noise. Re-annotating the SMILE-Temporal benchmark (672 videos, 1,683 events) with 3-5 annotators per video (alpha = 0.757), we find systematic patterns: disagreement is 1.73 times larger at offsets than onsets, far more common for chuckles than full laughs (77% vs. 20%), and predictable from event attributes (AUC = 0.831). Evaluating against a single annotator breaks down under this structure: system scores shift by 0.246 F1 depending on the chosen ground truth, correctly ranking systems only 69.7% of the time (vs. 80% against all annotators). We propose a disagreement-calibrated evaluation that scores predictions against the full annotator distribution using conformally calibrated tolerance bands (wider at offsets, 0.727s, than onsets, 0.5s). The per-annotator annotations and analysis code are available at https://github.com/WSCSports/MTLLFM-temporal-laughter-localization.
CommentsAccepted to the Workshop on Affective & Behavior Analysis in-the-wild, ECCV 2026
