Title:
液体食品のための包装容器における逸脱の検出
Document Type and Number:
Japanese Patent JP7469337
Kind Code:
B2
Abstract:
A monitoring system implements a method for versatile and efficient training of a machine learning-based model for subsequent detection and grading of deviations in packaging containers for liquid food in a manufacturing plant. The method comprises creating (31) a virtual model of a packaging container or of a starting material for use in producing the packaging container; obtaining (32) probability distributions for features that are characteristic of a deviation type; producing (33) reproductions of the virtual model with deviations included among the reproductions in correspondence with the probability distributions; associating (34) gradings with the reproductions; and inputting (35) the reproductions and the associated gradings for training of the machine learning-based model for subsequent detection and grading of an actual deviation in image data acquired in the manufacturing plant.
Inventors:
Peter Johannesson
Eric Bergvar
Eric Bergvar
Application Number:
JP2021575225A
Publication Date:
April 16, 2024
Filing Date:
June 15, 2020
Export Citation:
Assignee:
Tetra Laval Holdings and Finance SA
International Classes:
G06T7/00; B65B57/00; B65B57/02; G01N21/90
Domestic Patent References:
| JP2015176175A |
Other References:
Yuhuan Liu,外3名,Defect Inspection of Medicine Vials Using LBP Features and SVM Classifier,2017 2nd International Conference on Image, Vision and Computing (ICIVC),2017年06月02日,pp. 41-45
Attorney, Agent or Firm:
Yoshinobu Idogawa
Previous Patent: β-glucan composition and its uses
Next Patent: Detection chip and its manufacturing method, usage method, and reaction system
Next Patent: Detection chip and its manufacturing method, usage method, and reaction system