Identification of effect pigments in target coatings

By employing digital image semantic segmentation and neural network methods, the challenge of identifying effect pigments in complex coatings was solved, achieving efficient and accurate color matching while reducing manual intervention and time consumption.

CN115668303BActive Publication Date: 2026-06-05BASF COATINGS GMBH
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202180037808.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-05-27
Filing Date
2021-05-20
Publication Date
2026-06-05
Estimated Expiration
2041-05-20

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively identify and match effect pigments in complex coatings, especially given the interactions between different types of effect pigments, resulting in time-consuming color matching processes and inconsistent results.

Method used

A semantic segmentation method for digital images is adopted. The coating image is annotated and classified pixel by pixel by training a neural network. Combined with image segmentation technology and annotation tools, the pigment type of each pixel is identified and labeled. The optimal matching formula is searched in the database using statistical data.

Benefits of technology

It improves the efficiency and accuracy of color matching, can automatically identify effect pigments in complex coatings, reduces human intervention, and provides consistent matching results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115668303B_ABST
    Figure CN115668303B_ABST
Patent Text Reader

Abstract

This invention relates to a computer-implemented method comprising at least the following steps: 1) providing a digital image and a corresponding formula for a coating composition having a known pigment and / or pigment class associated with the corresponding digital image; 2) classifying each pixel for each image by visually inspecting the corresponding image pixel-by-pixel using at least one image segmentation technique and annotating each pixel with pigment labels and / or pigment class labels consistent with the visual appearance and formula associated with the corresponding image using an image annotation tool; 3) providing an associated pixel-by-pixel annotated image for each image; 4) training a first neural network implemented and running on at least one computer processor, using the provided digital image as input and the associated pixel-by-pixel annotated image as output. 310), wherein a first neural network (310) is trained to classify each pixel in a corresponding input image using pigment labels and / or pigment class labels associated with corresponding associated annotated images, or to associate each pixel in a corresponding input image with pigment labels and / or pigment class labels associated with corresponding associated annotated images; 5) the trained first neural network (310) is made available in at least one computer processor to apply the trained first neural network to at least one unknown input image of the target coating and to assign pigment labels and / or pigment class labels to each pixel in the input image; 6) based on the assigned pigment labels and / or pigment class labels, statistical data of the corresponding identified pigments and / or pigment classes are determined and / or output for each input image. The invention further provides corresponding devices and computer-readable media.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Pigment identification of complex coating mixtures with sparkle color

    US20170200288A1

  • Pigment identification of complex coating mixtures with sparkle color

    US9905027B2

  • A target detection method and device for a parallel robot vision system

    CN109785333A

  • Methods, systems, and programming for annotating an image including scoring using a plurality of trained classifiers corresponding to a plurality of clustered image groups associated with a set of weighted labels

    US8909563B1