CA3095023C - Methodes d'entrainement en continu pour systemes de determination d'anomalies dans une image d'un objet - Google Patents

Methodes d'entrainement en continu pour systemes de determination d'anomalies dans une image d'un objet

Info

Publication number
CA3095023C
CA3095023C CA3095023A CA3095023A CA3095023C CA 3095023 C CA3095023 C CA 3095023C CA 3095023 A CA3095023 A CA 3095023A CA 3095023 A CA3095023 A CA 3095023A CA 3095023 C CA3095023 C CA 3095023C
Authority
CA
Canada
Prior art keywords
anomaly
image
model
encoder
types
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CA3095023A
Other languages
English (en)
Other versions
CA3095023A1 (fr
Inventor
Negin SOKHANDAN ASL
Original Assignee
ServiceNow Inc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by ServiceNow Inc filed Critical ServiceNow Inc
Priority to CA3095023A priority Critical patent/CA3095023C/fr
Publication of CA3095023A1 publication Critical patent/CA3095023A1/fr
Application granted granted Critical
Publication of CA3095023C publication Critical patent/CA3095023C/fr
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G06N3/0455—Auto-encoder networks; Encoder-decoder networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0475—Generative networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0499—Feedforward networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/088—Non-supervised learning, e.g. competitive learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/0895—Weakly supervised learning, e.g. semi-supervised or self-supervised learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30164—Workpiece; Machine component

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Quality & Reliability (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)

Abstract

Il est décrit un système de détermination d’anomalies dans une image d’un objet qui est d’abord entraîné à l’aide d’un premier ensemble d’images correspondant à un premier type d’anomalies pour l’objet. Un modèle de l’objet est ainsi formé dans un espace latent. Une étiquette pour chaque image anormale est utilisée pour calculer des vecteurs contenant les moyennes et les écarts types pour chaque premier type d’anomalie. Ces moyennes et ces écarts types sont ensuite utilisés pour calculer une perte de log-vraisemblance pour chaque premier type d’anomalie. Le système est réentraîné à l’aide de deuxièmes ensembles d’images correspondant aux deuxièmes types d’anomalies pour l’objet. Les vecteurs sont complétés par les étiquettes pour chaque deuxième type d’anomalie. Un échantillon statistiquement suffisant de renseignements, sous forme de vecteurs de moyennes et d’écarts types, est fourni à l’espace latent. Une perte de log-vraisemblance pour chacun des premiers et deuxièmes types d’anomalies est calculée à partir de leur moyenne et de leur écart type respectifs.
CA3095023A 2020-10-02 2020-10-02 Methodes d'entrainement en continu pour systemes de determination d'anomalies dans une image d'un objet Active CA3095023C (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CA3095023A CA3095023C (fr) 2020-10-02 2020-10-02 Methodes d'entrainement en continu pour systemes de determination d'anomalies dans une image d'un objet

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CA3095023A CA3095023C (fr) 2020-10-02 2020-10-02 Methodes d'entrainement en continu pour systemes de determination d'anomalies dans une image d'un objet

Publications (2)

Publication Number Publication Date
CA3095023A1 CA3095023A1 (fr) 2022-04-02
CA3095023C true CA3095023C (fr) 2025-12-16

Family

ID=80929833

Family Applications (1)

Application Number Title Priority Date Filing Date
CA3095023A Active CA3095023C (fr) 2020-10-02 2020-10-02 Methodes d'entrainement en continu pour systemes de determination d'anomalies dans une image d'un objet

Country Status (1)

Country Link
CA (1) CA3095023C (fr)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116366790B (zh) * 2023-05-26 2023-10-27 深圳市神飞致远技术有限公司 一种网络视频存储方法及系统

Also Published As

Publication number Publication date
CA3095023A1 (fr) 2022-04-02

Similar Documents

Publication Publication Date Title
US11670072B2 (en) Systems and computer-implemented methods for identifying anomalies in an object and training methods therefor
US12223422B2 (en) Continuous training methods for systems identifying anomalies in an image of an object
US10853937B2 (en) Unsupervised image-based anomaly detection using multi-scale context-dependent deep autoencoding gaussian mixture model
US11314989B2 (en) Training a generative model and a discriminative model
JP2023549579A (ja) ビデオ行動認識のための時間ボトルネック・アテンション・アーキテクチャ
CN110537184A (zh) 利用生成式对抗网络的半监督回归
CN111950582B (zh) 为分类模型确定扰动掩模
Li et al. Fabric defect segmentation system based on a lightweight GAN for industrial Internet of Things
US12555365B2 (en) Systematic testing of AI image recognition
Zahid et al. Adversarial diffusion for few-shot scene adaptive video anomaly detection
CN113688844A (zh) 神经网络训练方法、装置、电子设备及存储介质
CA3095023C (fr) Methodes d'entrainement en continu pour systemes de determination d'anomalies dans une image d'un objet
EP3696771A1 (fr) Système de traitement d'une instance d'entrée, procédé et support
CN115471725A (zh) 用于异常检测的图像处理方法及设备和神经网络构建方法
WO2022070145A1 (fr) Systèmes et procédés informatisés destinés à identifier des anomalies dans un objet et procédés d'entraînement associés
Song et al. Loop closure detection of visual SLAM based on variational autoencoder
CA3095022C (fr) Systemes et methodes informatiques pour determiner des anomalies dans un objet et methodes d'entrainement connexes
Koshil et al. Anomalouspatchcore: exploring the use of anomalous samples in industrial anomaly detection
CN115115828A (zh) 数据处理方法、装置、程序产品、计算机设备和介质
US20250200422A1 (en) Post-hoc uncertainty quantification for machine learning systems
Li et al. A TG-AGD Anomaly Image Detection model based on residual bottleneck attention and time series prediction
Ribeiro Marnet et al. Uncertainty driven active learning for image segmentation in underwater inspection
US12430563B2 (en) Learning-based clean data selection
Hotta et al. Subspace‐Guided Feature Reconstruction for Unsupervised Anomaly Localization
Sharma et al. A semi-supervised multiscale generalized-VAE framework for one-class classification

Legal Events

Date Code Title Description
MFA Maintenance fee for application paid

Free format text: FEE DESCRIPTION TEXT: MF (APPLICATION, 4TH ANNIV.) - STANDARD

Year of fee payment: 4

U00 Fee paid

Free format text: ST27 STATUS EVENT CODE: A-2-2-U10-U00-U101 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: MAINTENANCE REQUEST RECEIVED

Effective date: 20240930

U11 Full renewal or maintenance fee paid

Free format text: ST27 STATUS EVENT CODE: A-2-2-U10-U11-U102 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: MAINTENANCE FEE PAYMENT DETERMINED COMPLIANT

Effective date: 20240930

Free format text: ST27 STATUS EVENT CODE: A-2-2-U10-U11-U102 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: MAINTENANCE FEE PAYMENT PAID IN FULL

Effective date: 20240930

D00 Search and/or examination requested or commenced

Free format text: ST27 STATUS EVENT CODE: A-2-2-D10-D00-D164 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: RESPONSE TO NOTICE OF ALLOWANCE

Effective date: 20241012

P11 Amendment of application requested

Free format text: ST27 STATUS EVENT CODE: A-2-2-P10-P11-P124 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: AMENDMENT AFTER ALLOWANCE (AAA) RECEIVED

Effective date: 20241012

R11 Change to the name of applicant or owner or transfer of ownership requested

Free format text: ST27 STATUS EVENT CODE: A-2-2-R10-R11-R127 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: TRANSFER RECORDAL REQUEST OR RESPONSE

Effective date: 20250423

W00 Other event occurred

Free format text: ST27 STATUS EVENT CODE: A-2-2-W10-W00-W111 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: CORRESPONDENT DETERMINED COMPLIANT

Effective date: 20250502

R14 Transfer of ownership recorded

Free format text: ST27 STATUS EVENT CODE: A-2-2-R10-R14-R129 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: TRANSFER REQUIREMENTS DETERMINED COMPLIANT

Effective date: 20250609

W00 Other event occurred

Free format text: ST27 STATUS EVENT CODE: A-2-2-W10-W00-W111 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: CORRESPONDENT DETERMINED COMPLIANT

Effective date: 20250620

P13 Application amended

Free format text: ST27 STATUS EVENT CODE: A-2-2-P10-P13-P126 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: AMENDMENT AFTER ALLOWANCE REQUIREMENTS DETERMINED COMPLIANT

Effective date: 20250627

D22 Grant of ip right intended

Free format text: ST27 STATUS EVENT CODE: A-2-4-D10-D22-D143 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: PRE-GRANT

Effective date: 20250820

W00 Other event occurred

Free format text: ST27 STATUS EVENT CODE: A-2-2-W10-W00-W111 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: CORRESPONDENT DETERMINED COMPLIANT

Effective date: 20250820

Free format text: ST27 STATUS EVENT CODE: A-2-2-W10-W00-W100 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: LETTER SENT

Effective date: 20250820

MFA Maintenance fee for application paid

Free format text: FEE DESCRIPTION TEXT: MF (APPLICATION, 5TH ANNIV.) - STANDARD

Year of fee payment: 5

U00 Fee paid

Free format text: ST27 STATUS EVENT CODE: A-4-4-U10-U00-U101 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: MAINTENANCE REQUEST RECEIVED

Effective date: 20250923

U11 Full renewal or maintenance fee paid

Free format text: ST27 STATUS EVENT CODE: A-4-4-U10-U11-U102 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: MAINTENANCE FEE PAYMENT PAID IN FULL

Effective date: 20250923

Q17 Modified document published

Free format text: ST27 STATUS EVENT CODE: A-4-4-Q10-Q17-Q103 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: DOCUMENT PUBLISHED

Effective date: 20251212

F11 Ip right granted following substantive examination

Free format text: ST27 STATUS EVENT CODE: A-4-4-F10-F11-X000 (AS PROVIDED BY THE NATIONAL OFFICE); EVENT TEXT: GRANT BY ISSUANCE

Effective date: 20251216