# US 12,182,989 B2 · Printed matter inspection device, printed matter inspection method, program, and printing apparatus > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,182,989 B2 - Document ID: US12182989B2 - Status: Granted - Kind: B2 - Application number: 17/691,505 - Filed: 2022-03-10 - Published: 2022-06-23 - Granted: 2024-12-31 - Earliest claimed: 2019-09-18 - Assignees: FUJIFILM Corporation - Applicants: FUJIFILM Corporation - Inventors: Masaki Seki - Web page: https://patentgrep.com/patent/US12182989B2 ## CPC - G06T 7/001: {using an image reference approach} - G06V 10/82: using neural networks - B41J 2/015: characterised by the jet generation process (B41J2/215 takes precedence) - B41J 2/145: Arrangement thereof - B41J 2/2142: {Detection of malfunctioning nozzles (generating single droplets or particles on demand by pressure, e.g. electromechanical transducers B41J2/045, B41J2/05; jet deflection sensors B41J2/125; for cleaning purposes B41J2/16579)} - B41J 2/2146: {for line print heads} - B41J 29/38: Drives, motors, controls or automatic cut-off devices for the entire printing mechanism - B41J 29/393: Devices for controlling or analysing the entire machine {; Controlling or analysing mechanical parameters involving printing of test patterns} - G01N 21/892: characterised by the flaw, defect or object feature examined - G06K 15/027: {Test patterns and calibration (arrangements for controlling or analysing printing mechanisms of typewriters or selective printing units B41J29/393 takes precedence; colour correction using test pattern analysis in general H04N1/6033)} - G06T 7/0004: {Industrial image inspection} - G06T 7/30: Determination of transform parameters for the alignment of images, i.e. image registration - G06V 10/774: Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting - G06V 10/7747: {Organisation of the process, e.g. bagging or boosting} - G06V 20/60: Type of objects - G06T 2207/10008: from scanner, fax or copier - G06T 2207/20081: Training; Learning - G06T 2207/20084: Artificial neural networks [ANN] - G06T 2207/30144: Printing quality - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - B41J: TYPEWRITERS; SELECTIVE PRINTING MECHANISMS, i.e. MECHANISMS PRINTING OTHERWISE THAN FROM A FORME; CORRECTION OF TYPOGRAPHICAL ERRORS - G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q) - G06K: GRAPHICAL DATA READING (image or video recognition or understanding G06V); PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS ## Abstract Provided are a printed matter inspection device, a printed matter inspection method, a program, and a printing apparatus capable of suppressing erroneous detection in detection of a defect occurring in one direction. The printed matter inspection device and the printing apparatus include a print data acquisition unit that acquires print data, an imaging data acquisition unit that acquires imaging data of a printed matter, a registration processing unit that executes registration between the print data and the imaging data by applying a specified registration accuracy, and an inspection processing unit that acquires defect information of the imaging data based on the print data. A learning model that has been learned by using a disturbance-imparted learning data set having, as an input, disturbance-imparted imaging data in which a disturbance that shifts a position of the imaging data in a direction intersecting a relative transport direction between the printing section and the medium in the printing apparatus is imparted, and having defect information as an output is applied to the inspection processing unit. ## Claims ### Claim 1 A printed matter inspection device comprising one or more processors configured to: acquire print data input to a printing apparatus in a case where a printed matter is generated; acquire imaging data of the printed matter; execute registration between the print data and the imaging data by applying a specified registration accuracy; and perform inspection processing to acquire defect information of the imaging data on the basis of the print data by using the print data and the imaging data on which the registration has been executed, wherein a learning model that has learned a relationship between a defect candidate detected from the imaging data and a defect in the printed matter is applied to the inspection processing, and the learning model executes the learning by using a disturbance-imparted learning data set having, as an input, disturbance-imparted imaging data in which a disturbance that shifts a position of the imaging data within a range of the registration accuracy in a direction intersecting a relative transport direction between a printer and a medium in the printing apparatus is imparted, and having defect information of the disturbance-imparted imaging data as an output. ### Claim 2 (depends on 1) The printed matter inspection device according to claim 1, wherein the learning model executes the learning by using the print data in combination as the disturbance-imparted learning data set. ### Claim 3 (depends on 1) The printed matter inspection device according to claim 1, wherein the learning model executes learning using a learning data set having, as an input, the print data and the imaging data in which the disturbance is not applied, and having defect information of the imaging data as an output. ### Claim 4 (depends on 1) The printed matter inspection device according to claim 1, wherein the one or more processors are configured to acquire the imaging data of the printed matter generated by using an ink jet printing apparatus. ### Claim 5 (depends on 4) The printed matter inspection device according to claim 4, wherein the one or more processors are configured to acquire the defect information representing the defect that occurs in a direction intersecting an arrangement direction of a plurality of nozzles provided in the ink jet printing apparatus, and the learning model executes learning using the disturbance-imparted imaging data in which the disturbance that shifts the position of the imaging data in the arrangement direction of the plurality of nozzles is imparted. ### Claim 6 (depends on 5) The printed matter inspection device according to claim 5, wherein the learning model further executes learning using the disturbance-imparted imaging data in which a change in an image structure of the imaging data with respect to the print data is imparted as a disturbance. ### Claim 7 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by correction processing in the printing apparatus is imparted as the disturbance. ### Claim 8 (depends on 7) The printed matter inspection device according to claim 7, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by jetting failure correction processing of the nozzles is imparted as the disturbance. ### Claim 9 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data according to a type of network applied to the printing apparatus is imparted as the disturbance. ### Claim 10 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data occurred depending on an individual difference of an imaging element provided in an imaging device for generating the imaging data is imparted as the disturbance. ### Claim 11 (depends on 10) The printed matter inspection device according to claim 10, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data according to a change in brightness of the imaging data is imparted as the disturbance. ### Claim 12 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data according to a change in brightness of the imaging data according to a difference in the medium is imparted as the disturbance. ### Claim 13 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by a noise according to a surrounding environment of the printer is imparted as the disturbance. ### Claim 14 (depends on 13) The printed matter inspection device according to claim 13, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by a noise according to an ambient temperature of the printer as the surrounding environment of the printer is imparted as the disturbance. ### Claim 15 (depends on 13) The printed matter inspection device according to claim 13, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by a noise according to an ambient humidity of the printer as the surrounding environment of the printer is imparted as the disturbance. ### Claim 16 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by a change in amount of ink applied to one pixel with respect to a specified amount is imparted as the disturbance. ### Claim 17 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by a stain on the medium is imparted as the disturbance. ### Claim 18 (depends on 6) The printed matter inspection device according to claim 6, wherein the learning model executes learning using the disturbance-imparted imaging data in which the change in the image structure of the imaging data with respect to the print data caused by an unevenness occurring in the printed matter is imparted as the disturbance. ### Claim 19 A printed matter inspection method comprising: acquiring print data input to a printing apparatus in a case where a printed matter is generated; acquiring imaging data of the printed matter; executing registration between the print data and the imaging data by applying a specified registration accuracy; and performing inspection processing to acquire defect information of the imaging data on the basis of the print data by using the print data and the imaging data on which the registration has been executed, wherein a learning model that has learned a relationship between a defect candidate detected from the imaging data and a defect in the printed matter is applied to the inspection processing, and the learning model executes the learning by using a disturbance-imparted learning data set having, as an input, disturbance-imparted imaging data in which a disturbance that shifts a position of the imaging data within a range of the registration accuracy in a direction intersecting a relative transport direction between a printer and a medium in the printing apparatus is imparted, and having defect information of the disturbance-imparted imaging data as an output. ### Claim 20 (depends on 19) A non-temporary and computer-readable recording medium for causing a computer to execute the printed matter inspection method according to claim 19, in a case where a command stored in the recording medium is read by the computer. ### Claim 21 A printing apparatus comprising: one or more processors configured to acquire print data applied in a case where a printed matter is generated; a transporter configured to transport a medium in a medium transport direction; and a printer configured to execute printing based on the print data on the medium transported using the transporter, wherein the one or more processors are further configured to: acquire imaging data of the printed matter; execute registration between the print data and the imaging data by applying a specified registration accuracy; and perform inspection processing to acquire defect information of the imaging data on the basis of the print data by using the print data and the imaging data on which the registration has been executed, wherein a learning model that has learned a relationship between a defect candidate detected from the imaging data and a defect in the printed matter is applied to the inspection processing, and the learning model executes the learning by using a disturbance-imparted learning data set having, as an input, disturbance-imparted imaging data in which a disturbance that shifts a position of the imaging data within a range of the registration accuracy in a direction intersecting a relative transport direction between the printer and the medium is imparted, and having defect information of the disturbance-imparted imaging data as an output.