Exploring the Use of Different Feature Levels of CNN for Anomaly Detection

Isack Farady, Lakshay Bansal, Somchoke Ruengittinun, Chia Chen Kuo, Chih Yang Lin

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

Anomaly detection is the task of uncovering out-of-distribution samples from the majority of data. Typically, this is treated as a one-class classification problem where the only data available to analyze is the normal data. With regard to collecting features of normal data, the high-dimensional features from CNN can be used to learn the normality. The last layer of CNN with more semantic information is generally used to learn the normality. In contrast, this work proposes learning features from different levels of high-dimensional features instead of using only high-level features. With the assumption that the training data is normally distributed, we present an anomaly detection algorithm consisting of a deep feature extraction stage with ResNet18 followed by dimensionality reduction via PCA. The anomaly classification stage comprises two class-conditional transformation models implemented via Gaussian Mixture Model. Our proposal leverages feature-reconstruction error as anomaly scores between two high-dimensional feature vectors. In this study, we analyze and compare the effect of using different blocks of a pre-trained ResNet18 on a well-known industrial anomaly detection dataset. Results suggest that using the best output features of CNN can significantly improve the model's ability to predict anomalous samples.

Original languageEnglish
Title of host publicationICSEC 2022 - International Computer Science and Engineering Conference 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages146-150
Number of pages5
ISBN (Electronic)9781665491983
DOIs
StatePublished - 2022
Event26th International Computer Science and Engineering Conference, ICSEC 2022 - Sakon Nakhon, Thailand
Duration: 21 Dec 202223 Dec 2022

Publication series

NameICSEC 2022 - International Computer Science and Engineering Conference 2022

Conference

Conference26th International Computer Science and Engineering Conference, ICSEC 2022
Country/TerritoryThailand
CitySakon Nakhon
Period21/12/2223/12/22

Keywords

  • PCA
  • computer vision
  • dimensionality reduction
  • high-dimensional features
  • industrial anomaly detection

Fingerprint

Dive into the research topics of 'Exploring the Use of Different Feature Levels of CNN for Anomaly Detection'. Together they form a unique fingerprint.

Cite this