JP7571524B2 - Method for identifying ink for inkjet printers and system for identifying ink for inkjet printers - Google Patents

Description

The present invention relates to an ink discrimination method and an ink discrimination system.

Patent document 1 describes an ink discrimination method that includes the steps of irradiating light onto the filled ink, measuring the amount of light that is transmitted through or reflected by the ink, whereby multiple amounts of light with different wavelengths are measured for one color of ink, and discriminating whether the filled ink is a specified ink based on the multiple measured amounts of light.

JP 2005-231356 A

However, in the ink discrimination method described in Patent Document 1, in order to be able to determine whether an ink is a specific ink or not based on the measured value of the amount of light transmitted through or reflected by the ink, it was necessary to set an appropriate threshold value for discrimination. In other words, if the set threshold value is not set appropriately, there was a problem in that the ink could not be accurately discriminated.

The ink discrimination method of the present invention generates a trained model by machine learning using training data that associates at least one ink characteristic data out of absorbance, transmittance, and reflectance observed by irradiating light onto the ink with the type of ink or attribute information of the ink including the type of ink, and discriminates the type of ink corresponding to the ink observation data based on output data of the trained model obtained by inputting at least one ink observation data out of absorbance, transmittance, and reflectance observed by irradiating light onto the ink.

The ink discrimination system of the present invention includes an ink determiner as a trained model obtained by machine learning using training data that associates at least one ink characteristic data out of absorbance, transmittance, and reflectance observed by irradiating light onto the ink with the type of ink or attribute information of the ink including the type of ink, and an ink discrimination unit that determines the type of ink corresponding to the ink observation data based on output data of the ink determiner obtained by inputting at least one ink observation data out of absorbance, transmittance, and reflectance observed by irradiating light onto the ink.

FIG. 1 is a block diagram showing a configuration of a machine learning device that generates a trained model according to an embodiment. 1 is a flowchart showing a process for generating a trained model through machine learning. FIG. 2 is a schematic diagram showing an example of a training model used in the embodiment. 1 is a block diagram showing a configuration of an ink determination system according to an embodiment. 5 is a flowchart illustrating an ink determination process according to the embodiment.

1. Embodiment The ink discrimination method of the present embodiment is characterized in that it discriminates ink using an ink determiner as a trained model generated by machine learning using training data that associates ink characteristic data observed when ink is irradiated with light with ink type or ink attribute information including ink type. When discriminating ink using the ink determiner, data observed when light is irradiated on the ink to be discriminated is obtained as ink observation data, the obtained ink observation data is input to the ink determiner, and the type of ink is discriminated based on output data of the ink determiner obtained in response to the input.

First, the process of generating an ink determiner will be specifically described with reference to FIGS.
Examples of the ink include, but are not limited to, inks used in inkjet printers. In addition to printing, the ink may be inks used for drawing, painting, writing, and the like.
Identifying the type of ink means, for example, identifying the ink name given by the ink manufacturer or the ink product number assigned by the ink manufacturer.

Ink attribute information refers to at least one of the following information: type of ink, i.e., ink name and ink product number, ink color, ink type, ingredients contained in the ink, ink manufacturer, ink manufacturing region, and ink manufacturing date.
The type of ink includes, for example, information such as water-based, oil-based, ultraviolet curing type, heat curing type, dye-based, pigment-based, and the like.
The ink characteristic data is at least one of the absorbance A, transmittance T %, and reflectance R %, which are observed when light is irradiated onto the ink.

The ink determiner 2 is generated by training a training model 5 using teacher data 4 in accordance with a learning program 3 in the machine learning device 1 shown in Fig. 1. The ink determiner 2 is a trained model generated using the training model 5 from the teacher data 4 obtained up to a certain point in time, and when it is provided to an ink discrimination system described below via a communication unit 50, it is treated as the ink determiner 2 in that ink discrimination system.
The machine learning device 1 is a computer system, and includes a machine learning unit 10, an input unit 20, a display unit 30, a memory unit 40, a communication unit 50, and the like.

The machine learning unit 10 includes a CPU, a RAM, and a ROM, and executes calculations necessary for machine learning according to the learning program 3 stored in the storage unit 40. The machine learning unit 10 may include a GPU or various processors designed for machine learning in order to perform machine learning.
CPU stands for Central Processing Unit, RAM stands for Random access memory, ROM stands for Read-Only Memory, and GPU stands for Graphics Processing Unit.

The input unit 20 is an information input means serving as a user interface, specifically, for example, a keyboard and a mouse pointer.
The display unit 30 is an information display means serving as a user interface, and displays, for example, information input from the input unit 20 and the results of calculations by the machine learning unit 10 under the control of the machine learning unit 10.

The memory unit 40 is a rewritable storage medium such as a hard disk drive or a memory card, and stores a learning program 3 on which the machine learning unit 10 operates, teacher data 4 for performing machine learning, a training model 5, an ink determiner 2 as a learned model generated as a result of machine learning, and the like.
The communication unit 50 includes a general-purpose interface such as a LAN interface or a USB interface, and transmits and receives information to and from external electronic devices.

In this embodiment, the machine learning device 1 performs machine learning using a training model 5 with ink property data of various inks and information on the ink type corresponding to the ink property data as teacher data 4. The information on the ink type is ink type or ink attribute information including the ink type.
The reason why data such as absorbance A, transmittance T%, and reflectance R% obtained by spectroscopic analysis of the ink are used as the ink characteristic data is that these characteristics differ depending on the type of ink.

The absorbance A, transmittance T%, and reflectance R% of the ink are obtained by using a spectrophotometer to evaluate the intensity of light absorbed by the ink, the intensity transmitted through the ink, and the intensity reflected by the ink relative to the intensity of light irradiated onto the sample ink.
When the intensity of the incident light is I0, the intensity of the transmitted light is I1, and the intensity of the reflected light is I2, they are calculated as follows.
Absorbance A=log(I0/I1)
Transmittance T%=I1/I0×100
Reflectance R%=I2/I0×100
In spectroscopic analysis, the wavelength of the irradiated light is divided into a predetermined wavelength range, for example, from the ultraviolet region to the infrared region, in 10 nm intervals, and a set of data is obtained for the absorbance A, transmittance T%, and reflectance R% for each wavelength range.

Figure 2 is a flowchart showing the process in which the machine learning unit 10 performs machine learning to generate an ink determiner 2. Before starting this process, ink characteristic data for multiple types of ink is collected as teacher data 4 associated with the ink type or ink attribute information including the ink type, and stored in the memory unit 40.

First, in step S<b>1 , the training model 5 and the teacher data 4 are obtained from the storage unit 40 .
Next, in steps S2 and S3, machine learning processing is performed using the training model 5 until generalization is completed. In step S3, the determination of whether generalization of the training model 5 is completed or not is made by threshold determination of the accuracy rate of the output obtained by inputting test data into the training model 5 up to that point.
When generalization is completed by providing the teacher data 4 to the training model 5 and performing machine learning, the ink determiner 2 is generated. In step S4, the ink determiner 2 is stored in the storage unit 40 as a learned model.

The training model 5 for machine learning can be defined in various ways. FIG. 3 is a diagram showing a schematic example of a training model used in this embodiment. In the figure, each of the total n layers of the CNN is indicated by L1 to Ln, and the nodes of a normal neural network are indicated by white circles. In this embodiment, a CNN is used, but other models such as various neural networks such as a capsule network type or a vector neural network type may also be used. CNN stands for Convolutional Neural Network.

The first layer L1 is provided with a plurality of nodes for inputting ink characteristic data for each fixed wavelength. In this embodiment, for example, the reflectance R% for each fixed wavelength indicated by the spectral reflectance data is input as input data to each node of the first layer L1, which is the input layer, and final output data corresponding to the reflectance R% is output from the final output layer Ln.
Instead of or in addition to the data on the reflectance R%, the transmittance T% or the absorbance A for each certain wavelength may be used. For example, when three pieces of data, absorbance A, transmittance T%, and reflectance R%, are used, three training models 5 shown in Fig. 3 are provided, and the results corresponding to the absorbance A, transmittance T%, and reflectance R% are output from the final layer of each training model 5, and a final output layer Ln is constructed to integrate and judge these results, and the final result is output.
Furthermore, among the ink attribute information, information that may affect the tendency of the ink characteristic data that is determined according to the type of ink may be input as input data from a newly provided node.

The output of each node in the first layer L1 is connected to a node in the next second layer L2 with a predetermined weighting. This is the same for the second layer L2 and onwards, up to the Ln-1th layer. By repeatedly correcting the weighting between each node in each layer using the teacher data 4, learning progresses and an ink determiner 2 is generated as a trained model.

Next, the ink discrimination system will be described with reference to FIG.
The ink discrimination system 1000 of this embodiment includes an ink discrimination device 100 and a spectroscopic analysis device 200 .

The ink discrimination device 100 is a computer system and includes an ink discrimination unit 110, an input unit 120, a display unit 130, a memory unit 140, a communication unit 150, etc.

The ink determining unit 110 includes a CPU, RAM, and ROM, and uses the ink determiner 2 stored in the storage unit 140 to execute calculations necessary to determine the type of ink.
The input unit 120 is an information input means serving as a user interface, specifically, for example, a keyboard and a mouse pointer.
The display unit 130 is an information display means that serves as a user interface, and under the control of the ink determining unit 110, displays, for example, information input from the input unit 120 and the results of calculations by the ink determining unit 110.

The storage unit 140 is a rewritable storage medium such as a hard disk drive or a memory card, and stores a program for operating the ink determining unit 110, an ink determiner 2 for determining the type of ink, and the like.
The communication unit 150 includes a general-purpose interface such as a LAN interface or a USB interface, and transmits and receives information to and from the spectroscopic analysis device 200 and external electronic devices including the machine learning device 1.

The spectroscopic analysis device 200 includes a spectroscopic analysis unit 210, a communication unit 250, and the like.
The spectroscopic analysis unit 210 includes a light source, a spectroscope, a detector, etc., and can obtain at least one ink observation data among absorbance A, transmittance T%, and reflectance R% that are observed by irradiating light onto the ink. Note that it is preferable that the spectroscopic analysis unit 210 can perform spectroscopic analysis with the same measurement specifications and analysis specifications as the spectrophotometer used when obtaining the above-mentioned ink characteristic data, that is, the ink characteristic data as the teacher data 4 in the machine learning when generating the ink determinator 2 as a learned model.

The communication unit 250 includes a general-purpose interface, such as a LAN interface or a USB interface, and transmits and receives information to and from the ink discrimination device 100 and external electronic devices. The communication unit 250 can transmit the ink observation data acquired by the spectroscopic analysis unit 210 to the ink discrimination device 100.

That is, the ink discrimination system 1000 includes an ink determiner 2 serving as a trained model obtained by machine learning using teacher data 4 that associates at least one ink characteristic data item among absorbance A, transmittance T%, and reflectance R% observed by irradiating light onto the ink with ink type or ink attribute information including ink type, and an ink discrimination unit 110 that determines the type of ink corresponding to the ink observation data based on the output data of the ink determiner 2 obtained by inputting at least one ink observation data item among absorbance A, transmittance T%, and reflectance R% observed by irradiating light onto the ink.

The ink discrimination device 100 may also have a machine learning function. Specifically, the ink discrimination device 100 may have a function of using the acquired ink determiner 2 as a training model, and further performing machine learning using the ink observation data newly obtained from the spectroscopic analysis device 200 and information on the correct ink type corresponding to the ink observation data as new teacher data 4N to refine the ink determiner 2.

Next, an ink discrimination flow using the ink discrimination device 100 will be described with reference to the flowchart shown in FIG.
To start the ink discrimination process, the ink discrimination device 100 has an ink determiner 2 as a trained model in the storage unit 140. That is, in the ink discrimination method of the present embodiment, a trained model is generated in advance by machine learning using teacher data 4 that associates at least one of ink characteristic data of absorbance A, transmittance T %, and reflectance R %, which are observed when light is irradiated onto the ink, with ink type or ink attribute information including ink type.

In the ink discrimination process, first, in step Sa1, a sample of the ink to be discriminated is prepared. Specifically, the ink sample is set in a state that allows evaluation by the spectroscopic analysis unit 210.
Next, in step Sa2, the spectroscopic analysis unit 210 performs spectroscopic analysis of the ink sample to obtain ink observation data. In the spectroscopic analysis, the wavelength of the irradiated light is divided into a predetermined wavelength range, for example, from the ultraviolet region to the infrared region, in 10 nm intervals, and data is obtained as a set of absorbance A, transmittance T%, and reflectance R% for that wavelength range.
Spectroscopic analysis unit 210 transmits the acquired ink observation data to ink discrimination device 100 via communication unit 250, and upon receiving the ink observation data, ink discrimination device 100 inputs the ink observation data to ink determiner 2 in ink discrimination unit 110 in step Sa3, and determines the type of ink corresponding to the ink observation data based on the output data of ink determiner 2. Ink discrimination unit 110 displays the determination result on display unit 130.

2. Effects of the embodiment According to the present embodiment, the following effects can be obtained.
The ink discrimination method of the present embodiment is characterized in that an ink determiner 2 is generated as a trained model by machine learning using teacher data 4 that associates at least one ink characteristic data out of absorbance A, transmittance T%, and reflectance R% observed by irradiating light on the ink with ink type or ink attribute information including ink type, and a type of ink corresponding to the ink observation data is discriminated based on output data of the ink determiner 2 obtained by inputting at least one ink observation data out of absorbance A, transmittance T%, and reflectance R% observed by irradiating light on the ink. In other words, since the type of ink is discriminated based on the result of judgment using the machine-learned ink determiner 2, there is no need to set a threshold value for discrimination, and a highly accurate discrimination can be performed without reviewing the threshold value to improve the accuracy of discrimination.

In addition, the ink discrimination method of this embodiment can include at least one of the following information in the ink attribute information used as training data 4 in machine learning: ink color, type, ingredients, manufacturer, production region, and production time. In this case, the accuracy of discrimination can be improved when discriminating the type of ink using an ink determiner 2 trained by machine learning using the training data 4.

The ink discrimination system 1000 of this embodiment includes an ink determiner 2 as a learned model machine-learned using teacher data 4 that associates at least one ink characteristic data of absorbance A, transmittance T%, and reflectance R% observed by irradiating light on the ink with ink type or ink attribute information including ink type, and an ink discrimination unit 110 that determines the type of ink corresponding to the ink observation data based on the output data of the ink determiner 2 obtained by inputting at least one ink observation data of absorbance A, transmittance T%, and reflectance R% observed by irradiating light on the ink. In other words, according to the ink discrimination system 1000 of this embodiment, the type of ink is determined based on the result of the determination using the ink determiner 2 trained by machine learning, so there is no need to set a threshold value for discrimination, and it is possible to perform highly accurate discrimination without reviewing the threshold value to improve the accuracy of discrimination.

1...machine learning device, 2...ink determiner, 3...learning program, 4...teacher data, 5...training model, 10...machine learning section, 20...input section, 30...display section, 40...storage section, 50...communication section, 100...ink discrimination device, 110...ink discrimination section, 120...input section, 130...display section, 140...storage section, 150...communication section, 200...spectroscopic analysis device, 210...spectroscopic analysis section, 250...communication section, 1000...ink discrimination system.

Claims (5)

generating a trained model by machine learning using training data that associates three pieces of ink-jet printer ink characteristic data, i.e., absorbance A, transmittance T%, and reflectance R%, which are observed when light is irradiated onto ink for ink-jet printers using a spectroscopic analysis device equipped with a spectroscopic analysis unit including a light source, a spectroscope, and a detector, with the type of ink for ink- jet printers or attribute information of the ink for ink-jet printers including the type of ink for ink-jet printers;
determining a type of ink for an ink jet printer corresponding to the ink observation data for the ink jet printer based on output data of the trained model obtained by inputting three pieces of ink jet printer ink observation data, the absorbance A, the transmittance T%, and the reflectance R% , which are observed by irradiating the ink for an ink jet printer used in the ink jet printer with light using the spectroscopic analysis device;
The attribute information is at least one of the color, type, ingredients, manufacturer, manufacturing region, and manufacturing date of the ink for the inkjet printer.
A method for identifying inks for inkjet printers . The type of the ink for the inkjet printer is at least one of water-based, oil-based, ultraviolet curing, heat curing, dye-based, and pigment-based components;
The method for identifying an ink for an ink-jet printer according to claim 1 . When the intensity of incident light incident on the inkjet printer ink is I0, the intensity of transmitted light from the inkjet printer ink is I1, and the intensity of reflected light from the inkjet printer ink is I2,

Absorbance A=log(I0/I1)
Transmittance T%=I1/I0×100
Reflectance R%=I2/I0×100
and

The I0, I1, and I2 are observed at wavelengths of 10 nm each from the ultraviolet region to the infrared region,
The absorbance A, the transmittance T%, and the reflectance R% are obtained for each wavelength of 10 nm . The ink-jet printer ink discrimination method according to claims 1 and 2. The training model used in the machine learning is a Convolutional Neural Network (CNN), with each of the n layers being denoted by L1 to Ln,
A training model A of the absorbance A, a training model T% of the transmittance T%, and a training model R% of the reflectance R% are set,
The training model A uses the absorbance A for each wavelength divided by 10 nm as input data to each node of a first layer L1 which is an input layer of the training model A for the absorbance A, and outputs final output data A corresponding to the absorbance A from a final output layer Ln;
The training model T% uses the transmittance T% for each wavelength divided by 10 nm as input data to each node of a first layer L1 which is an input layer of the training model T% of the transmittance T%, and outputs final output data T% corresponding to the transmittance T% from a final output layer Ln;
The training model R% uses the reflectance R% for each wavelength divided by 10 nm as input data to each node of a first layer L1 which is an input layer of the training model R% of the reflectance R%, and outputs final output data R% corresponding to the reflectance R% from a final output layer Ln;
The final output data A, the final output data T%, and the final output data R% are integrated to obtain the final result.
The method for identifying an ink for an ink jet printer according to claim 3. a spectroscopic analysis device including a spectroscopic analysis unit including a light source, a spectroscope, and a detector, which irradiates light onto an ink for an ink jet printer and observes three properties of the ink for an ink jet printer, namely , absorbance A, transmittance T %, and reflectance R %, of the ink for an ink jet printer;
an ink jet printer ink determiner as a trained model obtained by machine learning using training data that associates three pieces of ink jet printer ink characteristic data, i.e., absorbance A, transmittance T %, and reflectance R % , observed by the spectroscopic analyzer with the type of ink jet printer ink or attribute information of the ink jet printer ink including the type of ink jet printer ink;
an ink jet printer ink discrimination unit that discriminates the type of ink for the ink jet printer corresponding to the ink jet printer ink observation data based on output data of the ink jet printer ink determiner obtained by inputting three pieces of ink jet printer ink observation data, i.e., absorbance A, transmittance T %, and reflectance R %, observed by the spectroscopic analyzer ;
The attribute information is at least one of the color, type, ingredients, manufacturer, manufacturing region, and manufacturing date of the ink for the inkjet printer.
Ink discrimination system for inkjet printers .

Images