arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.25015cs.CVcs.AI
收藏

IECLA:面向情感图像编辑的可解释情感贡献学习

IECLA: Interpretable Emotion Contribution Learning for Affective Image Editing

University of Science and Technology of China(中国科学技术大学);Harbin Institute of Technology(哈尔滨工业大学)

Yijie Guo, Dexiang Hong, Weidong Chen, Xinyan Liu, Zhendong Mao

AI总结:

本研究提出IECLA框架,将源情感归因与目标感知编辑分离,通过反事实干预构建监督,可更忠实地预测情感贡献,在六种编辑骨干上强化目标情感表达并保留图像内容结构。

摘要:

情感图像编辑(Affective Image Editing, AIE)旨在唤起目标情感,同时保留源图像的内容与结构。现有方法要么提升情感表达,要么提升视觉保真度,但仅靠目标情感无法明确应修改源图像的哪些区域、保留哪些区域,尤其在复杂场景中更为突出。本文提出面向情感图像编辑的可解释情感贡献学习框架(Interpretable Emotion Contribution Learning for Affective Image Editing, IECLA),该框架将源情感归因与目标感知编辑分离开来。IECLA通过反事实干预构建带符号的对象级监督,并训练上下文条件贡献预测器,以定位支持或反对源情感的证据。推理阶段,图像级评估器直接从输入估计源情感,预测器无需重复干预即可识别可编辑与需保留的证据;随后,目标感知编辑智能体将这些证据与期望情感转化为针对现有编辑器的、本地化且内容感知的操作。大量评估表明,IECLA在域内及跨数据集测试中均能更忠实地预测反事实贡献,外部视觉语言模型(VLM)与独立人类参考也验证了其本地化证据的质量;在六种编辑骨干网络上,IECLA均能持续强化目标情感表达,同时保留图像内容与结构,用户研究进一步显示其感知目标表达与自然度均有所提升。

英文摘要:

Affective image editing (AIE) aims to evoke a target emotion while preserving the content and structure of a source image. Existing methods improve emotional expression or visual fidelity, but the target emotion alone does not reveal which source regions should be changed and which should remain intact, especially in complex scenes. We introduce Interpretable Emotion Contribution Learning for Affective Image Editing (IECLA), a framework that separates source-emotion attribution from target-aware editing. IECLA constructs signed object-level supervision through counterfactual interventions and trains a context-conditioned contribution predictor to localize evidence that supports or opposes the source emotion. At inference, an image-level evaluator estimates the source emotion directly from the input, while the predictor identifies editable and preservation evidence without repeated interventions. A target-aware editing agent then translates this evidence and the desired emotion into localized, content-aware operations for existing editors. Extensive evaluations show that IECLA predicts counterfactual contributions more faithfully on both in-domain and cross-dataset tests, while external VLM and independent human references support the quality of its localized evidence. Across six editing backbones, IECLA consistently strengthens target-emotion expression while preserving image content and structure, and user studies further show improved perceived target expression and naturalness.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑