基于球面正则化的支持向量描述视觉异常检测
DOI:
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

TP391. 4 TH701

基金项目:

国家自然科学基金(62202087)、广东省基础与应用基础研究基金(2024A1515010244,2021B1515120064)项目资助


Spherical regularized support vector description for visual anomaly detection
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    异常检测作为视觉领域中一项独特而关键的任务,在医疗、安保等领域具有广泛的前景。 异常检测目前受限于大规模 异常数据标注,因此现有方法集中在单类分类和弱监督学习,深度支持向量描述(Deep SVDD)是实现单类分类的常见方法。 然 而,传统 Deep SVDD 在开展异常检测时往往面临球体崩塌。 针对这一问题,提出了基于球面正则化的 SVDD 异常检测算法,通 过引入软间隔损失与支持向量的思想,优化模型学习流程。 进一步地,面向可标注样本,提出了基于 SVDD 的弱监督异常检测 方法。 在公开数据集 MNIST 和 CIFAR-10 上进行消融和对比实验,实验证明,相比于有监督算法,在 MNIST 数据集上,SRWSVDD 的性能提高了 3. 7% ,而在 CIFAR-10 数据集上则提高了 16. 7% 。 此外,与其他弱监督算法相比,SR-WSVDD 在 CIFAR- 10 数据集上提升了 1. 8% 。 所提出的 SR-SVDD 异常检测算法,弥补 Deep SVDD 容易发生球体崩塌的缺陷,使模型异常检测结 果更加准确。

    Abstract:

    Anomaly detection is an important task in the computer vision, such as medical, security. One of the challenges in anomaly detection is not easy to obtain large-scale annotated anomalous data. Existing methods focus on one-class classification and weakly supervised learning. Deep support vector data Description (Deep SVDD) is an important method to realize one-class anomaly detection. However, previous Deep SVDD often encounter the hypersphere collapse when constructing the model of the hypersphere. To solve this problem, support vector data description based on spherical regularization (SR-SVDD) is proposed in this paper. SR-SVDD applies the idea of support vectors to optimize the learning process by introducing slack terms. Furthermore, this paper proposes weakly supervised support vector data description based on spherical regularization (SR-WSVDD), which utilizes small amounts of labeled data. Ablation experiments and comparison experiments are carried out on MNIST and CIFAR-10. Experimental results show that, compared with supervised algorithms, the performance of SR-WSVDD is improved by 3. 7% on the MNIST, and 16. 7% on the CIFAR-10. In addition, compared with other weakly supervised algorithms, SR-WSVDD improves by 1. 8% on CIFAR-10 dataset. The proposed SR-SVDD solves the spherical collapse of previous Deep SVDD, and makes the anomaly detection results more accurate.

    参考文献
    相似文献
    引证文献
引用本文

邓诗卓,滕 达,李晓红,陈佳祺,陈东岳.基于球面正则化的支持向量描述视觉异常检测[J].仪器仪表学报,2024,45(3):315-325

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2024-05-31
  • 出版日期: