当前位置: X-MOL 学术Pattern Recogn. Lett. › 论文详情
Anomaly detection using improved deep SVDD model with data structure preservation
Pattern Recognition Letters ( IF 3.255 ) Pub Date : 2021-05-04 , DOI: 10.1016/j.patrec.2021.04.020
Zheng Zhang, Xiaogang Deng

Support vector data description (SVDD) is a classical anomaly detection algorithm. How to develop a deep version of SVDD is one valuable problem in the anomaly detection field. Aiming at this problem, an improved SVDD model called deep structure preserving SVDD (DSPSVDD) is proposed by integrating the deep feature extraction with the data structure preservation. Firstly, the typical SVDD methods are revisited in view of model depth profiles and the limitations of the present deep SVDD model are analyzed. Then in order to extract the deep data features more effectively, an enhanced comprehensive optimization objective is designed for the deep SVDD model by considering both the hypersphere volume minimization and the network reconstruction error minimization simultaneously. The experimental results on the MNIST, Fashion-MNIST, and MVTec AD image benchmark datasets show that the proposed DSPSVDD method achieves the better anomaly detection performance compared with the traditional deep SVDD method.



中文翻译:

使用改进的深度SVDD模型进行异常检测并保留数据结构

支持向量数据描述(SVDD)是一种经典的异常检测算法。在异常检测领域,如何开发深版本的SVDD是一个有价值的问题。针对这个问题,通过将深度特征提取与数据结构保存相结合,提出了一种改进的SVDD模型,称为深度结构保存SVDD(DSPSVDD)。首先,针对模型深度剖面,重新探讨了典型的SVDD方法,并分析了当前深SVDD模型的局限性。然后,为了更有效地提取深度数据特征,通过同时考虑超球面体积最小化和网络重构误差最小化,为深度SVDD模型设计了一个增强的综合优化目标。在MNIST,Fashion-MNIST,

更新日期:2021-05-04
全部期刊列表>>
欢迎新作者ACS
聚焦环境污染物
专攻离子通道生理学研究
中国作者高影响力研究精选
虚拟特刊
屿渡论文,编辑服务
浙大
上海中医药大学
苏州大学
江南大学
四川大学
灵长脑研究中心
毛凌玲
南开大学陈瑶
朱如意
中科院
南开大学
隐藏1h前已浏览文章
课题组网站
新版X-MOL期刊搜索和高级搜索功能介绍
ACS材料视界
华辉
天合科研
x-mol收录
试剂库存
down
wechat
bug