Explicit Boundary Guided Semi-Push-Pull Contrastive Learning for Supervised Anomaly Detection

被引:24
|
作者
Yao, Xincheng [1 ]
Li, Ruoqi [1 ]
Zhang, Jing [3 ]
Sun, Jun [1 ]
Zhang, Chongyang [1 ,2 ]
机构
[1] Shanghai Jiao Tong Univ, Sch Elect Informat & Elect Engn, Shanghai, Peoples R China
[2] Shanghai Jiao Tong Univ, AI Inst, MoE Key Lab Artificial Intelligence, Shanghai, Peoples R China
[3] Acad Mil Sci, Res Inst Syst Engn, Beijing, Peoples R China
来源
2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) | 2023年
关键词
SUPPORT;
D O I
10.1109/CVPR52729.2023.02346
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Most anomaly detection (AD) models are learned using only normal samples in an unsupervised way, which may result in ambiguous decision boundary and insufficient discriminability. In fact, a few anomaly samples are often available in real-world applications, the valuable knowledge of known anomalies should also be effectively exploited. However, utilizing a few known anomalies during training may cause another issue that the model may be biased by those known anomalies and fail to generalize to unseen anomalies. In this paper, we tackle supervised anomaly detection, i.e., we learn AD models using a few available anomalies with the objective to detect both the seen and unseen anomalies. We propose a novel explicit boundary guided semi-push-pull contrastive learning mechanism, which can enhance model's discriminability while mitigating the bias issue. Our approach is based on two core designs: First, we find an explicit and compact separating boundary as the guidance for further feature learning. As the boundary only relies on the normal feature distribution, the bias problem caused by a few known anomalies can be alleviated. Second, a boundary guided semi-push-pull loss is developed to only pull the normal features together while pushing the abnormal features apart from the separating boundary beyond a certain margin region. In this way, our model can form a more explicit and discriminative decision boundary to distinguish known and also unseen anomalies from normal samples more effectively. Code will be available at https://github.com/xcyao00/BGAD.
引用
收藏
页码:24490 / 24499
页数:10
相关论文
共 50 条
  • [1] SSCL: Semi-supervised Contrastive Learning for Industrial Anomaly Detection
    Cai, Wei
    Gao, Jiechao
    PATTERN RECOGNITION AND COMPUTER VISION, PRCV 2023, PT IV, 2024, 14428 : 100 - 112
  • [2] Semi-Supervised Anomaly Detection with Contrastive Regularization
    Jezequel, Loic
    Vu, Ngoc-Son
    Beaudet, Jean
    Histace, Aymeric
    2022 26TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2022, : 2664 - 2671
  • [3] Hierarchical Semi-supervised Contrastive Learning for Contamination-Resistant Anomaly Detection
    Wang, Gaoang
    Zhan, Yibing
    Wang, Xinchao
    Song, Mingli
    Nahrstedt, Klara
    COMPUTER VISION, ECCV 2022, PT XXV, 2022, 13685 : 110 - 128
  • [4] Attention-Guided Supervised Contrastive Learning for Deepfake Detection
    Waseem, Saima
    Abu Bakar, Syed Abdul Rahman Bin Syed
    Ahmed, Bilal Ashfaq
    2024 IEEE 8TH INTERNATIONAL CONFERENCE ON SIGNAL AND IMAGE PROCESSING APPLICATIONS, ICSIPA, 2024,
  • [5] Semi-supervised vanishing point detection with contrastive learning
    Wang, Yukun
    Gu, Shuo
    Liu, Yinbo
    Kong, Hui
    PATTERN RECOGNITION, 2024, 153
  • [6] Semi-supervised Anomaly Detection with Reinforcement Learning
    Lee, Changheon
    Kim, JoonKyu
    Kang, Suk-Ju
    2022 37TH INTERNATIONAL TECHNICAL CONFERENCE ON CIRCUITS/SYSTEMS, COMPUTERS AND COMMUNICATIONS (ITC-CSCC 2022), 2022, : 933 - 936
  • [7] CADet: Fully Self-Supervised Anomaly Detection With Contrastive Learning
    Guille-Escuret, Charles
    Rodriguez, Pau
    Vazquez, David
    Mitliagkas, Ioannis
    Monteiro, Joao
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [8] Generative and Contrastive Self-Supervised Learning for Graph Anomaly Detection
    Zheng, Yu
    Jin, Ming
    Liu, Yixin
    Chi, Lianhua
    Phan, Khoa T.
    Chen, Yi-Ping Phoebe
    IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2023, 35 (12) : 12220 - 12233
  • [9] A NOVEL CONTRASTIVE LEARNING FRAMEWORK FOR SELF-SUPERVISED ANOMALY DETECTION
    Li, Jingze
    Lian, Zhichao
    Li, Min
    2022 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP, 2022, : 3366 - 3370
  • [10] SELF-SUPERVISED ACOUSTIC ANOMALY DETECTION VIA CONTRASTIVE LEARNING
    Hojjati, Hadi
    Armanfard, Narges
    2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 3253 - 3257