模式识别与人工智能
2025年4月5日 星期六   首 页     期刊简介     编委会     投稿指南     伦理声明     联系我们                                                                English
模式识别与人工智能  2021, Vol. 34 Issue (9): 809-823    DOI: 10.16451/j.cnki.issn1003-6059.202109004
“深度学习算法及在图像与视觉的应用”专题 最新目录| 下期目录| 过刊浏览| 高级检索 |
基于去冗余特征和语义关系约束的零样本属性识别
张桂梅1, 龙邦耀1, 曾接贤1, 黄军阳1
1.南昌航空大学 计算机视觉研究所 南昌 330063
Zero-Shot Attribute Recognition Based on De-redundancy Features and Semantic Relationship Constraint
ZHANG Guimei1, LONG Bangyao1, ZENG Jiexian1, HUANG Junyang1
1. Institute of Computer Vision, Nanchang Hangkong University, Nanchang 330063

全文: PDF (2937 KB)   HTML (1 KB) 
输出: BibTeX | EndNote (RIS)      
摘要 基于生成式的零样本识别方法在生成特征时受冗余信息和域偏移的影响,识别精度不佳.针对此问题,文中提出基于去冗余特征和语义关系约束的零样本属性识别方法.首先,将视觉特征映射到一个新的特征空间,通过互相关信息对视觉特征进行去冗余处理,在去除冗余视觉特征的同时保留类别的相关性,由于在识别过程中减少冗余信息的干扰,从而提高零样本识别的精度.然后,利用可见类和不可见类之间的语义关系建立知识迁移模型,并引入语义关系约束损失,约束知识迁移的过程,使生成器生成的视觉特征更能反映可见类和不可见类之间语义关系,缓解两者之间的域偏移问题.最后,引入循环一致性结构,使生成的伪特征更接近真实特征.在数据集上的实验证实文中方法提高零样本识别任务的精度,并具有较优的泛化性能.
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
张桂梅
龙邦耀
曾接贤
黄军阳
关键词 去冗余特征语义关系约束零样本识别域偏移属性识别    
Abstract:The generative zero shot recognition method is affected by redundant information and domain shifting while generating features, and thus its recognition accuracy is poor. To deal with the problem, a zero shot attribute recognition method based on de-redundant features and semantic relationship constraint is proposed. Firstly, the visual features are mapped to a new feature space, and the visual features are de-redundant via cross-correlation information. The redundant visual features are removed with the correlation of the categories preserved. The accuracy of zero shot recognition is improved due to the reduction of redundant information interference in the recognition process. Then, a knowledge transfer model is established using the semantic relationship between the seen and unseen classes, and the loss of semantic relationship is introduced to constrain the process of knowledge transfer. Consequently, the semantic relationship between the seen and unseen classes is reflected better by the visual features generated by the generator ,and the problem of domain shifting between them is alleviated as well. Finally, the cycle consistency structure is introduced to make the generated pseudo-features closer to the real features. Experiments on datasets show that the proposed method improves the accuracy of zero shot recognition tasks with better generalization performance.
Key wordsDe-redundancy Features    Semantic Relational Constraint    Zero-Shot Recognition    Domain Shifting    Attribute Recognition   
收稿日期: 2021-03-29     
ZTFLH: TP 319.4  
基金资助:国家自然科学基金项目(No.61462065,61763033)资助
通讯作者: 张桂梅,博士,教授,主要研究方向为计算机视觉、图像处理、模式识别.E-mail:guimei.zh@163.com.   
作者简介: 龙邦耀,硕士研究生,主要研究方向为计算机视觉、图像处理、模式识别.E-mail:609527215@qq.com.
曾接贤,硕士,教授,主要研究方向为计算机视觉、图像处理、模式识别.E-mail:zengjx58@163.com.
黄军阳,硕士研究生,主要研究方向为计算机视觉、图像处理、模式识别.E-mail:mike1160554026@163.com.
引用本文:   
张桂梅, 龙邦耀, 曾接贤, 黄军阳. 基于去冗余特征和语义关系约束的零样本属性识别[J]. 模式识别与人工智能, 2021, 34(9): 809-823. ZHANG Guimei, LONG Bangyao, ZENG Jiexian, HUANG Junyang. Zero-Shot Attribute Recognition Based on De-redundancy Features and Semantic Relationship Constraint. , 2021, 34(9): 809-823.
链接本文:  
http://manu46.magtech.com.cn/Jweb_prai/CN/10.16451/j.cnki.issn1003-6059.202109004      或     http://manu46.magtech.com.cn/Jweb_prai/CN/Y2021/V34/I9/809
版权所有 © 《模式识别与人工智能》编辑部
地址:安微省合肥市蜀山湖路350号 电话:0551-65591176 传真:0551-65591176 Email:bjb@iim.ac.cn
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn