Learning deep feature correspondence for unsupervised anomaly detection and segmentation

人工智能 异常检测 模式识别(心理学) 分割 特征(语言学) 稳健性(进化) 水准点(测量) 计算机科学 特征提取 特征学习 背景(考古学) 无监督学习 深度学习 哲学 古生物学 基因 生物 化学 生物化学 地理 语言学 大地测量学
作者
Jie Yang,Yong Shi,Zhiquan Qi
出处
期刊:Pattern Recognition [Elsevier BV]
卷期号:132: 108874-108874 被引量:39
标识
DOI:10.1016/j.patcog.2022.108874
摘要

• A learnable deep feature correspondence (DFC) method is proposed for unsupervised anomaly detection and segmentation. • DFC achieves state of the art results on the benchmark unsupervised anomaly detection and segmentation task MVTec AD. • DFC is very effective for detecting and segmenting the anomalous structures and patterns that appear in confined local regions of images, especially the industrial anomalies. • The generality of DFC is demonstrated by applying it on a real industrial inspection scene. Developing machine learning models that can detect and localize the unexpected or anomalous structures within images is very important for numerous computer vision tasks, such as the defect inspection of manufactured products. However, it is challenging especially when there are few or even no anomalous image samples available. In this paper, we propose an unsupervised mechanism, i.e. deep feature correspondence (DFC), which can be effectively leveraged to detect and segment out the anomalies in images solely with the prior knowledge from anomaly-free samples. We develop our DFC in an asymmetric dual network framework that consists of a generic feature extraction network and an elaborated feature estimation network, and detect the possible anomalies within images by modeling and evaluating the associated deep feature correspondence between the two dual network branches. Furthermore, to improve the robustness of the DFC and further boost the detection performance, we specifically propose a self-feature enhancement (SFE) strategy and a multi-context residual learning (MCRL) network module. Extensive experiments have been carried out to validate the effectiveness of our DFC and the proposed SFE and MCRL. Our approach is very effective for detecting and segmenting the anomalies that appear in confined local regions of images, especially the industrial anomalies. It advances the state-of-the-art performances on the benchmark dataset – MVTec AD. Besides, when applied to a real industrial inspection scene, it outperforms the comparatives by a large margin.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
5709xx的应助被Dyying采纳,获得10
刚刚
Fipped发布了新的文献求助10
1秒前
乐观的素阴完成签到 ,获得积分10
2秒前
DW的应助被su777采纳,获得10
2秒前
fly圈圈呀完成签到,获得积分10
2秒前
456关注了科研通微信公众号
4秒前
xx完成签到 ,获得积分10
4秒前
5秒前
Nole的应助被开心德地采纳,获得10
7秒前
苹果芷荷完成签到,获得积分10
7秒前
十一完成签到 ,获得积分10
8秒前
Fipped完成签到,获得积分10
8秒前
9秒前
10秒前
阿申爱乐的应助被Dyying采纳,获得30
11秒前
领导范儿的应助被ttchen采纳,获得10
11秒前
11秒前
LMXS发布了新的文献求助10
11秒前
ldy完成签到,获得积分10
12秒前
Cakoibao的应助被小狮子采纳,获得10
13秒前
大个的应助被燕麦片采纳,获得10
13秒前
Nole的应助被开心德地采纳,获得10
13秒前
Gp1983发布了新的文献求助150
13秒前
学术猪八戒完成签到,获得积分10
15秒前
昼屿发布了新的文献求助10
15秒前
美罗培南完成签到 ,获得积分0
15秒前
Nuerrr完成签到 ,获得积分10
15秒前
Jiaqi0315完成签到,获得积分10
16秒前
搜集达人的应助被HaojunWang采纳,获得10
16秒前
6666发布了新的文献求助10
16秒前
_3完成签到,获得积分10
17秒前
深情安青的应助被开心德地采纳,获得10
19秒前
20秒前
21秒前
科研菜鸟完成签到,获得积分10
21秒前
22秒前
贾贡献的应助被ttchen采纳,获得10
22秒前
FashionBoy的应助被开心德地采纳,获得10
25秒前
8787发布了新的文献求助10
26秒前
勤奋仙人掌的应助被渡花采纳,获得10
28秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
拉伸高分子熔体中缠结行为的模拟和小角中子散射研究 1700
The World Atlas of Wine 9th edition 1000
Sex, Aging, and Death in a Medieval Medical Compendium: Trinity College Cambridge MS R.14.52, Its Texts, Language, and Scribe 500
高分子基功能界面水蒸发材料与应用研究 500
外场增强的水汽发电研究 500
基于结构调控的功能化水诱导发电材料与器件研究 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 工程类 有机化学 化学工程 生物化学 物理 复合材料 内科学 催化作用 光电子学 人工智能 细胞生物学 无机化学 物理化学 数学
热门帖子
关注 科研通微信公众号,转发送积分 8188380
求助须知:如何正确求助?哪些是违规求助? 9576202
关于积分的说明 21846057
捐赠科研通 7786438
什么是DOI,文献DOI怎么找? 3435307
关于科研通互助平台的介绍 2559604
邀请新用户注册赠送积分活动 2482312