公开(公告)号：WO2021170796A1

公开(公告)日：2021-09-02

申请号：PCT/EP2021/054820

申请日：2021-02-26

Applicant: BASF COATINGS GMBH

Inventor： BAUGHMAN, Donald R ,  LEOPOLD, Matthew ,  BISCHOFF, Guido ,  SCOTT, Stuart K ,  MCGUCKIN, Jessica J

IPC: G06T7/90 ,  G01J3/46

Abstract: The present invention refers to a method and a device that can measure an unknown target coating, can search, based on the measured data of the target coating, a database for one or more best matching coating formulas, i.e. one or more preliminary matching formulas, within the database and that can refine the search using an image similarity metric between images of the one or more best matching coating formulas on the one side and images of the target coating on the other side, using deep learning techniques.

公开(公告)号：WO2021239593A1

公开(公告)日：2021-12-02

申请号：PCT/EP2021/063564

申请日：2021-05-20

Applicant: BASF COATINGS GMBH

Inventor： BAUGHMAN, Donald R ,  BISCHOFF, Guido ,  LEOPOLD, Matthew ,  MCGUCKIN, Jessica J ,  SCOTT, Stuart K

IPC: G06K9/46 ,  G06K9/62 ,  G06T7/90 ,  H04N1/60 ,  G01J3/46 ,  B44D3/00 ,  G01N21/25

Abstract: The present invention refers to a computer-implemented method, the method comprising at least the following steps: 1) providing digital images and respective formulas for coating compositions with known pigments and/or pigment classes associated with the respective digital images, 2) classifying, using an image annotation tool, for each image each pixel, by visually reviewing the respective image pixel-wise, using at least one image segmentation technique and annotating each pixel with a pigment label and/or a pigment class label in alignment with the visual appearance and the formula associated with the respective image, 3) providing, for each image, an associated pixel wise annotated image, 4) training a first neural network (310), implemented and running on at least one computer processor, with the provided digital images as input and the associated pixel-wise annotated images as output wherein the first neural network (310) is trained to classify/correlate every pixel in a respective input image with a pigment label and/or pigment class label of a respective associated annotated image, 5) making the trained first neural network (310) available in the at least one computer processor for applying the trained first neural network to at least one unknown input image of a target coating and for assigning a pigment label and/or a pigment class label to each pixel in the input image, 6) determining and/or outputting, for each input image, based on the assigned pigment labels and/or pigment class labels, a statistic of corresponding identified pigments and/or pigment classes, respectively. The present invention further provides a respective device and a computer-readable medium.

公开(公告)号：WO2021094496A1

公开(公告)日：2021-05-20

申请号：PCT/EP2020/081979

申请日：2020-11-12

Applicant: BASF COATINGS GMBH

Inventor： BISCHOFF, Guido ,  BAUGHMAN, Donald R ,  LEOPOLD, Matthew ,  SCOTT, Stuart K

IPC: G06T7/90 ,  G01J3/46 ,  H04N1/60 ,  B44D3/00 ,  G01N21/00 ,  G01N21/25 ,  G01N21/47

Abstract: The present invention refers to a computer-implemented method, the method comprising at least the following steps: - obtaining, using at least one measuring device, color values, texture values and digital images of a target coating, - retrieving from a database which comprises formulas for coating compositions and interrelated color values, interrelated texture values, and interrelated digital images, one or more preliminary matching formulas based on the color values and/or the texture values obtained for the target coating, - determining sparkle points within the respective obtained images and within the respective images associated with the one or more preliminary matching formulas, - creating subimages of each sparkle point from the respective images, - providing the created subimages to a convolutional neural network, the convolutional neural network being trained to correlate a respective subimage of a respective sparkle point with a pigment and/or pigment class, - determining, based on an output of the neural network, at least one of the one or more preliminary matching formulas as the formula(s) best matching with the target coating. The present invention further provides a respective device and a non- transitory computer readable medium.