检索结果-内蒙古大学图书馆

Causalseg: investigating causality modeling for semi-supervised video object segmentation

Multimedia Systems 2025年第3期31卷 1-16页

作者： Zhang, Zhengjin Li, Nannan Wang, Wenmin Guo, Huiwen School of Computer Science and Engineering Macau University of Science and Technology Taipa999078 China School of Computer and Artificial Intelligence Chaohu University Bantang Road Anhui Hefei238024 China Shenzhen Bochuang Robot Technology Xixiang Guangdong Shenzhen518172 China

In Semi-Supervised Video Object Segmentation (SVOS), existing methods are often hindered by the issue of spurious association due to the lack of adequate supervision information. This issue negatively impacts the segmentation performance. To address this, we introduce a novel approach that leverages causal intervention to enhance SVOS performance. Causal intervention is particularly effective in mitigating spurious associations, thereby improving segmentation performance. Our proposed causal module is a versatile component that can be seamlessly integrated into matching-based SVOS to enhance segmentation accuracy. Our approach involves three key steps. Firstly, we introduce the Causal CAM method to identify and eliminate spurious association locations within the video frame. Secondly, we adapt the IRNet to effectively generate class-specific instance map information during training. Lastly, we incorporate the AFB-URR, STCN and XMem models with our causal intervention method. Experimental results validate the effectiveness of our approach in improving object segmentation accuracy. On the DAVIS2017 dataset, our method achieved an overall J & F Mean scores of 76.3%, 85.3% and 87.5% respectively. Compared to the three benchmark methods, our approach demonstrated improvements of 1.7%, 3.4% and 1.3% respectively. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：