JP2021022006A - Information processing device - Google Patents

Abstract

To provide a technology capable of efficiently and rationally deciding an appropriate tactile in designing of the tactile.SOLUTION: An information processing device (100) includes an obtaining unit (110) and a learning unit (120). The obtaining unit (110) obtains teacher data that contains either one or both of material information and surface shape information, and tactile phenomenon information corresponding thereto. The learning unit (120) refers to the teacher data, and learns a model (121) that outputs the tactile phenomenon information corresponding to either one of or both of the input material information and surface shape information.SELECTED DRAWING: Figure 1

Description

The present invention relates to an information processing device.

In recent years, in products that are used by human touch, it is required to realize a tactile sensation suitable for the product. Since the tactile sensation is a kind of human sensation, the design of the tactile sensation of a product is generally performed by trial and error in which a person makes a prototype and confirms the tactile sensation. Therefore, the design variables related to the tactile sensation of the product tend to be determined depending on the knowledge and sense of the person in charge of development (see, for example, Non-Patent Document 1).

"[Mazda's Takumi] Pursuing Human Sensation-Insatiable Quest to Support Texture Development (Part 2)" [online], May 17, 1991, [MAZDA] Mazda Official Blog Zoom-Zoom Blog, Internet < URL: https://blog.mazda.com/archive/20171220_01.html ＞

When designing the tactile sensation of a product as described above, rework may occur during material development, which may increase the cost and time required for development. In addition, design variables that have no causal relationship with the physical phenomenon that causes the tactile sensation may be set as management items, which may cost more than necessary. Furthermore, it may not be possible to obtain the understanding of the parties concerned such as the supplier because the target value is subjective or arbitrary. Furthermore, it is expected that it will take more time to change the tactile sensation based on the data acquired by others other than the developer. As described above, the above-mentioned method for designing a tactile sensation that involves trial and error leaves room for consideration from the viewpoint of efficiently and rationally determining an appropriate tactile sensation.

One aspect of the present invention is to provide a technique capable of efficiently and rationally determining an appropriate tactile sensation in the design of tactile sensation.

In order to solve the above problems, the information processing apparatus according to the first aspect of the present invention includes at least one of material information which is information about a material and surface shape information which is information about a surface shape. An acquisition unit that acquires teacher data including tactile phenomenon information that is information about a tactile phenomenon corresponding to at least one of the information, and at least one of the material information and the surface shape information is input to the input information. It is provided with a learning unit for learning a model that outputs corresponding tactile phenomenon information by referring to the teacher data.

According to this configuration, it is possible to generate a model that outputs the corresponding tactile phenomenon information from one or both of the material information and the surface shape information, and the tactile phenomenon information can be expressed quantitatively. The tactile sensation on the surface of the article can be determined efficiently and rationally.

In the first aspect, the teacher data may include the material information and the tactile phenomenon information corresponding to the material information, the model may include a regression analysis model, and the learning unit may include the learning unit. The model may be trained using the teacher data for each of the surface shape information.

According to this configuration, in the above model, it is more effective from the viewpoint of quantitatively expressing the tactile phenomenon information corresponding to the material information.

Further, in the first aspect, the teacher data may include the surface shape information and the tactile phenomenon information corresponding to the surface shape information, the model may include a neural network, and the learning unit may include the learning unit. , The model may be trained using the teacher data for each of the material information.

According to this configuration, in the above model, it is more effective from the viewpoint of quantitatively or uniquely expressing the tactile phenomenon information corresponding to the surface shape information.

Further, in the first aspect, the information processing apparatus may further include a teacher data generation unit that generates the teacher data by referring to a regression model showing the relationship between the material information and the tactile phenomenon information. ..

According to this configuration, it is possible to increase the teacher data having sufficient reliability by referring to the actually measured teacher data. Therefore, it is even more effective from the viewpoint of improving the accuracy of the estimation result of the tactile phenomenon information based on the model.

Further, the information processing apparatus according to the second aspect of the present invention includes a first material information which is information about a material, an acquisition unit which acquires a first surface shape information which is information about a surface shape, and the first material information. And the constraint condition using one of the second material information which is the information about the material and the second surface shape information which is the information about the surface shape while referring to the first surface shape information is satisfied. It includes an output unit that generates the other of the material information and the second surface shape information and outputs the generated information.

According to this configuration, the tactile sensation of the surface of the article having the first material and the surface shape can be reproduced by the second material and the surface shape by using one of the second material and the surface shape as a constraint condition. Is possible. Therefore, it is possible to efficiently and rationally determine an appropriate tactile sensation.

In the second aspect, the output unit inputs the first material information and the first surface shape information into the trained model, thereby causing a physical phenomenon corresponding to the first material information and the first surface shape information. The information may be specified, and the above-mentioned so as to satisfy the constraint condition using one of the second material information and the second surface shape information and to correspond to the specified physical phenomenon information. The other of the second material information and the second surface shape information may be generated.

According to this configuration, the physical phenomenon information can be quantitatively expressed, and the combination of the second material and the surface shape can be determined based on the quantitatively expressed physical phenomenon information. Therefore, it is even more effective from the viewpoint of rationally determining the combination in a short time.

In the second aspect, at least one of the first surface shape information and the second surface shape information may be two-dimensional image data or three-dimensional image data.

According to this configuration, it is possible to easily reproduce the surface of the article from the above image data. Therefore, it is even more effective from the viewpoint of easily embodying the surface state represented by the image data.

Further, the information processing apparatus according to the third aspect of the present invention has an acquisition unit that acquires at least one of material information that is information about the material and surface shape information that is information about the surface shape, and the acquisition unit acquires the information. By inputting the obtained material information and surface shape information into the trained model, an output unit that outputs physical phenomenon information corresponding to at least one of the material information and surface shape information acquired by the acquisition unit is provided.

According to this configuration, the tactile physical phenomenon can be uniquely determined from the quantitatively expressed surface shape information and material information. Therefore, it is possible to efficiently and rationally determine an appropriate tactile sensation.

Further, the information processing apparatus according to the fourth aspect of the present invention is physical phenomenon information which is information about a tactile phenomenon corresponding to at least one of material information which is information about a material and surface shape information which is information about a surface shape. By inputting the acquisition unit to acquire the information and the physical phenomenon information acquired by the acquisition unit into the trained model, at least one of the material information and the surface shape information corresponding to the physical phenomenon information acquired by the acquisition unit is output. It is provided with an output unit to be used.

According to this configuration, based on a desired tactile physical phenomenon that can be expressed quantitatively, the combination of the surface shape and the material that causes the tactile physical phenomenon can be determined in a short time and in a shorter time than the examination by trial and error by an engineer. It can be reasonably decided. Therefore, it is possible to efficiently and rationally determine an appropriate tactile sensation.

According to the above aspect of the present invention, it is possible to provide a technique capable of efficiently and rationally determining an appropriate tactile sensation in the design of tactile sensation.

It is a block diagram which shows an example of the functional structure of the information processing apparatus which concerns on Embodiment 1 of this invention. It is a block diagram which shows an example of the functional structure of the information processing apparatus which concerns on Embodiment 2 of this invention. It is a block diagram which shows an example of the functional structure of the information processing apparatus which concerns on Embodiment 3 of this invention. It is a figure for demonstrating from the input to the output of the information which concerns on Embodiment 3 of this invention.

[Embodiment 1]
Hereinafter, one embodiment of the present invention will be described in detail. FIG. 1 is a block diagram showing an example of a functional configuration of the information processing apparatus according to the first embodiment of the present invention. As shown in FIG. 1, the information processing apparatus 100 includes an acquisition unit 110 and a learning unit 120.

The acquisition unit 110 acquires teacher data. The teacher data includes at least one of the material information and the surface shape information, and the tactile phenomenon information corresponding to the at least one of the information.

Material information is information about the material. The material information can be appropriately selected within a range that specifies the material on the surface of the article, and examples thereof include a material name and a product name. The material information may be an analysis result obtained by quantitative analysis and qualitative analysis of the material on the surface of the article. Further, the material information may be information further classified according to the difference in the physical properties of the same type of material.

The surface shape information is information related to the surface shape. The surface shape information can be appropriately selected within a range for specifying the material on the surface of the article, and may be one kind or more. The surface shape information may be a physical property value representing the surface shape, and examples of the surface shape information include line roughness parameters such as arithmetic mean roughness, surface roughness parameters such as arithmetic mean height, and two of the article surface. Includes 3D image data and 3D image data on the surface of the article.

The surface shape information may be based on actual measurement or calculation (or estimation). The surface shape information based on the actual measurement can be acquired by using a known device capable of measuring or photographing the surface shape information. The surface shape information based on the calculation can be obtained by using a simulation using a known computer.

Of the material information and the surface shape information, the acquisition unit 110 may acquire only the material information, may acquire only the surface shape information, or may acquire both of these information. From the viewpoint of improving the accuracy of the tactile phenomenon information, it is preferable that the acquisition unit 110 acquires both the material information and the surface shape information.

The tactile phenomenon information corresponds to at least one or both of the material information and the surface shape information. The tactile phenomenon information is information on a tactile phenomenon, and may be represented by an amount related to a physical action occurring between an article and a user's finger touching the article (hereinafter, also referred to as “tactile physical quantity”). Examples of tactile physical quantities include the force applied to the surface of the article when it comes into contact, the temperature of the surface of the article at the time of contact, the vibration generated on the surface of the article due to the contact, the friction of the surface of the article, the hardness of the article, and the article at the time of contact. Includes shear forces generated in the article, the coefficient of friction between the user's fingers and the surface of the article, the trajectory of the fingers when the user touches the article, and data showing the contact time with the article. The tactile physical quantity can be detected by abutting a stylus equipped with various sensors capable of detecting the tactile physical quantity against the surface of an article and scanning the article.

As described above, the teacher data includes at least one of the material information and the surface shape information, and the tactile phenomenon information corresponding to the at least one of the information. The teacher data is data referred to in order to train the model 121 possessed by the learning unit 120, and can be appropriately determined according to the model.

The learning unit 120 trains the model 121 with reference to the teacher data. The model 121 may output predetermined tactile phenomenon information in response to input of at least one of material information and surface shape information. Model 121 may be one or more.

For example, as shown in FIG. 1, the acquisition unit 110 acquires material information, surface shape information, and tactile phenomenon information corresponding to these information as teacher data.

As an example, the teacher data includes material information and tactile phenomenon information corresponding to the material information. The model 121 may be, for example, a regression analysis model, and the learning unit 120 trains the model 121 using the teacher data for each surface shape information.

Learning the influence on the tactile physical phenomenon from such material information can be carried out as follows, for example. That is, a sufficient number of samples having substantially the same surface shape but different materials are prepared, and the tactile physical quantities (for example, the above-mentioned force, temperature, vibration, etc.) of the samples are measured. A regression model is created with the measured tactile physical quantity as the objective variable and the material information as the explanatory variable.

Further, as another example, the teacher data includes surface shape information and tactile phenomenon information corresponding to the surface shape information. The model 121 may be, for example, a neural network, and the learning unit 120 trains the model 121 using the teacher data for each material information.

Learning the influence on the tactile physical phenomenon from such surface shape information can be carried out as follows, for example. That is, a sufficient number of samples having substantially the same material information but different surface shapes are prepared, and the tactile physical quantity of the samples is measured. A sufficient number of data from the measurement are acquired, the neural network is trained, the weight of the path is determined for each material information, and the model is created in this way.

The model and the method for training it can be appropriately selected as long as the desired results can be obtained. Examples of models include neural networks and support vector machines. Examples of neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs) and fully coupled neural networks. Examples of algorithms for training the model include backpropagation and ID3.

[Summary]
As is clear from the above description, the information processing apparatus 100 includes at least one of the material information and the surface shape information, and the tactile phenomenon information which is the information about the tactile phenomenon corresponding to the at least one of the information. It includes an acquisition unit 110 for acquiring teacher data and a learning unit 120 for learning the model 121 with reference to the teacher data. Then, at least one of the material information and the surface shape information is input to the model 121, and the tactile phenomenon information corresponding to the input information is output. Since the tactile phenomenon information can be expressed quantitatively, according to the model 121, the tactile sensation on the surface of the article can be reasonably determined. Then, such a determination of the tactile sensation can be performed in an extremely short time as compared with trial and error by an engineer, and the efficiency is higher.

In the present embodiment, the teacher data includes material information and the corresponding tactile phenomenon information, the model 121 includes a regression analysis model, and the learning unit 120 trains the model 121 using the teacher data for each surface shape information. .. Since the material information can usually be uniquely determined, it can be appropriately predicted by a regression analysis model using the material information as an explanatory variable. Therefore, according to the present embodiment, the tactile phenomenon information corresponding to the material information can be quantitatively expressed.

Further, in the present embodiment, the teacher data includes surface shape information and tactile phenomenon information corresponding thereto, the model 121 includes a neural network, and the learning unit 120 uses the teacher data for each material information to provide the model 121. Let them learn. The surface shape information may be difficult to reproduce with a single parameter, but it can be represented by an image. From the viewpoint of handling the image as surface shape information, the model 121 using the surface shape information as an explanatory variable is preferably a neural network. Therefore, according to the present embodiment, the tactile phenomenon information corresponding to the surface shape information can be expressed quantitatively or uniquely.

[Embodiment 2]
Other embodiments of the present invention will be described below. For convenience of explanation, the same reference numerals will be added to the members having the same functions as the members described in the above embodiment, and the description will not be repeated.

FIG. 2 is a block diagram showing an example of a functional configuration of the information processing apparatus according to the second embodiment of the present invention. As shown in FIG. 2, the information processing apparatus 100a has the same configuration as the information processing apparatus 100 in the above-described first embodiment, except that the information processing apparatus 100a further includes a teacher data generation unit 130. The teacher data generation unit 130 generates teacher data with reference to the regression model. The regression model is a regression model that shows the relationship between material information and tactile phenomenon information.

The generation of such teacher data can be carried out, for example, by using the resample method in statistics. The regression model used for the reproduction of the teacher data may be the regression analysis model described above, or may be another regression model for the reproduction. Examples of teacher data playback techniques include the parametric bootstrap method.

In the parametric bootstrap method, a parametric model is usually applied to the data by the maximum likelihood method and resampled from the model, so that highly reliable data can be generated. Therefore, it is possible to generate data to be measured from a sample that cannot be actually measured due to problems such as trial cost, regardless of the sample. By generating such teacher data, the teacher data is further enriched, and as a result, it is possible to expand the learning of the regression model performed with reference to the teacher data.

In the information processing apparatus 100a, the acquisition unit 110 further acquires the teacher data generated by the teacher data generation unit 130. The learning unit 120 trains the model 121 with reference to the teacher data acquired by the acquisition unit 110, including the teacher data acquired by the acquisition unit 110 from the teacher data generation unit 130.

As is clear from the above description, in the present embodiment, the information processing apparatus 100a further includes a teacher data generation unit 130. Therefore, in addition to the effect of the first embodiment described above, the present embodiment further exerts the effect that the accuracy of the estimation result of the tactile phenomenon information based on the above regression model can be further improved.

[Embodiment 3]
Other embodiments of the present invention will be described below. For convenience of explanation, the same reference numerals will be added to the members having the same functions as the members described in the above embodiment, and the description will not be repeated.

FIG. 3 is a block diagram showing an example of the functional configuration of the information processing apparatus according to the third embodiment of the present invention. As shown in FIG. 3, the information processing apparatus 200 includes an acquisition unit 201, a physical phenomenon information generation unit 202, a learned model 203, a surface shape / material information generation unit 204, and a constraint condition identification unit 205. The acquisition unit 201 acquires information from the display unit 210 and the input unit 220. Further, the surface shape / material information generation unit 204 outputs information to the display unit 210.

The trained model 203 is a model trained in the above-described first or second embodiment, and includes, for example, at least one of the above-mentioned regression analysis model and the neural network.

The input unit 220 is a device for a user to input data, for example, a keyboard or a touch panel. The display unit 210 is a device capable of displaying an image, for example, a liquid crystal display device.

Hereinafter, generation of physical phenomenon information from image data and generation of image data from physical phenomenon information will be described. In the following description, in order to explain more clearly, the expressions "first" are used for the input data and "second" is used for the output data.

[Generation of physical phenomenon information from image data]
The acquisition unit 201 acquires the first material information which is information about the material and the first surface shape information which is information about the surface shape. For example, the user inputs the first material information of the first article to the input unit 220, and causes the display unit 210 to display a three-dimensional image as the first surface shape information of the first article. The acquisition unit 201 acquires the first material information from the input unit 220 and acquires the first surface shape information from the display unit 210.

The physical phenomenon information generation unit 202 generates the first physical phenomenon information corresponding to the first material information and the first surface shape information by referring to the first material information and the first surface shape information acquired by the acquisition unit 201. For example, the physical phenomenon information generation unit 202 specifies the first physical phenomenon information by inputting the first material information and the first surface shape information into the trained model 203.

Here, the "physical phenomenon information" is information representing a physical phenomenon brought about by the material information and the surface shape information, and may be the above-mentioned tactile phenomenon information. The tactile phenomenon information can be said to be information given to the human tactile sense by the material information and the surface shape information, and is an aspect of the physical phenomenon information. Further, the "first physical phenomenon information" is the physical phenomenon information corresponding to the first material information and the first surface shape information among the physical phenomenon information.

[Generation of image data from physical phenomenon information]
On the other hand, the constraint condition specifying unit 205 specifies one of the second material information and the second surface shape information. The second material information is material information different from the first material information, and the second surface shape information is surface shape information different from the first surface shape information. The constraint condition specifying unit 205 specifies the constraint condition by, for example, acquiring a signal of either the second material information or the second surface shape information input to the input unit 220 as a constraint condition by the user.

On the other hand, the surface shape / material information generation unit 204 acquires the first physical phenomenon information generated by the physical phenomenon information generation unit 202. Then, the surface shape / material information generation unit 204 uses the second material information and the second material information so that the constraint condition using one of the second material information and the second surface shape information acquired by the constraint condition specifying unit 205 is satisfied. Generate the other side of the surface shape information.

Here, the constraint condition using one of the second material information and the second surface shape information is, for example, to specify the material information corresponding to the first physical phenomenon information in the second material information, or the first. 1 It is to specify the surface shape information corresponding to the physical phenomenon information as the second surface shape information. Hereinafter, the process of generating the surface shape or material information by the surface shape / material information generation unit 204 will be further described.

(Search process)
As an example of the surface shape / material information generation process by the surface shape / material information generation unit 204, the search process described below can be mentioned.

The surface shape / material information generation unit 204 uses the second surface shape information and the second material information as a constraint condition with respect to a model similar to the learned model 203 used by the physical phenomenon information generation unit 202 in the search process. The second surface shape information or the second material information other than the specified one is input while being changed. Then, the surface shape / material information generation unit 204 specifies each region (physical phenomenon information) in the physical phenomenon space corresponding to each second surface shape and the second material information. Then, the surface shape / material information generation unit 204 specifies the physical phenomenon information that matches the first physical phenomenon information specified from the first surface shape information and the first material information, and the second surface shape information and the second material. Identify a combination of information.

When a combination of the second surface shape information and the second material information that specifies the physical phenomenon information that matches the first physical phenomenon information cannot be found, the surface shape / material information generation unit 204 uses the second surface shape information and the second material. Among the physical phenomenon information specified by the information, the physical phenomenon information having the highest degree of coincidence with the first physical phenomenon information may be specified as substantially matching the first physical phenomenon information. In this case, the surface shape information and the material information that specify the physical phenomenon information that substantially matches the first physical phenomenon information can be specified as the second surface shape information and the second material information.

The surface shape / material information generation unit 204 outputs the three-dimensional image data including the other of the generated second material information and the second surface shape information to the display unit 210. The three-dimensional image data includes one of the second material information and the second surface shape information, which are constraints, and the other of the second material information and the second surface shape information generated by the surface shape / material information generation unit 204. Image data including.

In this way, the information processing apparatus 200 refers to the first material information and the first surface shape information, and satisfies the constraint condition using one of the second material information and the second surface shape information. The other of the second material information and the second surface shape information is generated, and the generated information is output. The physical phenomenon information generation unit 202 to the constraint condition identification unit 205 correspond to an output unit that outputs the other information of the second material information and the second surface shape information generated as described above.

When the input to the output of the information regarding the surface state in the present embodiment is illustrated, it is schematically shown as shown in FIG. FIG. 4 is a diagram for explaining from input to output of information according to the third embodiment of the present invention. The process from input to output of information on the surface state in the present embodiment can be described as follows.

As shown in FIG. 4, first, the surface shape A is input as the first surface shape information. Then, the data of the surface shape A is converted into the first physical phenomenon information in the physical phenomenon space by using the weight in the trained model 203 (for example, CNN) of the data of the material information A input as the first material information. To do. Given the data of the surface shape A and the material composition A in this way, the predicted value of the physical phenomenon is determined by, for example, the trained model 203 of CNN.

Next, as a constraint condition, for example, the material information B as the second material information is specified. Then, the data of the surface shape B, which is another surface shape information (the above-mentioned second surface shape information) that is converted into the first physical phenomenon information while being weighted by the data of the material information B, is output. In this way, based on the above predicted value, the first physical phenomenon information in the physical phenomenon space is converted by using the weight of the material information B given as the constraint condition in the trained model 203 of the CNN. Then, the data of the surface shape B to be created is generated.

[Summary]
As is clear from the above description, the information processing apparatus 200 refers to the acquisition unit 201 that acquires the first material information and the first surface shape information, the first material information, and the first surface shape information, and the second It is provided with an output unit that generates the other of the second material information and the second surface shape information and outputs the generated information so that the constraint condition using one of the material information and the second surface shape information is satisfied. .. Therefore, the information processing apparatus 200 converts the first surface shape information (surface shape A) and the first material information (material configuration A) into the first physical phenomenon information, and the first physical phenomenon information is converted into the second surface shape information (the second surface shape information). It is possible to convert to the surface shape B) and the second material information (material composition B).

As a result, the management item on the material composition is uniquely obtained as the material information, and the management item on the surface shape is uniquely obtained from the objective data as the surface shape information. Furthermore, how these material information and surface shape information affect the tactile physical phenomenon (physical phenomenon information) that affects the quality of the article is clearly modeled. Therefore, when performing the tactile design, the effect of the material and the effect of the surface shape can be determined in a short time, quantitatively, and uniquely as compared with the conventional tactile design.

According to the tactile design, for example, the tactile sensation of an article made of a high-grade leather material can be reproduced with an article made of another resin material. Alternatively, the excellent tactile sensation of an article made of a highly functional material can be reproduced in a product made of a general purpose material. Alternatively, if the tactile sensation of an article of a predetermined excellent material becomes a problem, an article made of the material but having a desired tactile sensation can be manufactured.

In the present embodiment, the output unit inputs the first material information and the first surface shape information to the trained model 203, thereby causing a first physical phenomenon corresponding to the first material information and the first surface shape information. Identify the information. Then, the second material information and the second surface shape information are satisfied so that the constraint condition using one of the second material information and the second surface shape information is satisfied and corresponds to the specified first physical phenomenon information. Generate the other of. Since the model 121 described in the above-described first or second embodiment can be used as the trained model 203, the first physical phenomenon information can be quantitatively expressed. Therefore, the combination of another material and the surface shape (second material information and the second surface shape information) that can be converted into the first physical phenomenon information can be reasonably determined in a short time.

Further, in the present embodiment, either the first surface shape information or the second surface shape information is the three-dimensional image data. As described above, in the present embodiment, the surface shape information is provided as image data. Therefore, a technique for embodying an image, such as a transfer or a 3D printer, facilitates the realization of the surface of an article having surface shape information or the surface of an article having both surface shape information and material information.

Further, since the surface shape information is provided as image data, the surface shape information is uniquely determined. Therefore, the surface of the article having the second surface shape information and the second material information newly generated by the conversion can be uniquely determined.

[Embodiment 4]
Other embodiments of the present invention will be described below. For convenience of explanation, the same reference numerals will be added to the members having the same functions as the members described in the above embodiment, and the description will not be repeated.

The information processing apparatus of the present embodiment has an acquisition unit that acquires at least one of the material information and the surface shape information, and the acquisition unit by inputting the material information and the surface shape information acquired by the acquisition unit into the trained model. It is provided with an output unit that outputs physical phenomenon information corresponding to at least one of the material information and the surface shape information acquired by.

In the information processing device, for example, the information in FIG. 3 is shown in FIG. 3 except that the physical phenomenon information generation unit 202 outputs the physical phenomenon information to the display unit 210 and does not include the surface shape / material information generation unit 204 and the constraint condition identification unit 205. It can be configured in the same manner as the processing device 200. The physical phenomenon information generation unit 202 and the trained model 203 correspond to the above output unit.

According to this embodiment, the tactile physical phenomenon can be uniquely determined from the surface shape information and the material information expressed quantitatively. By converting the tactile physical phenomenon output here into a combination of different surface shape information and material information, for example, as described in the third embodiment, an appropriate tactile sensation can be efficiently and rationally determined. It is possible to do.

[Embodiment 5]
Other embodiments of the present invention will be described below. For convenience of explanation, the same reference numerals will be added to the members having the same functions as the members described in the above embodiment, and the description will not be repeated.

The information processing apparatus of the present embodiment inputs a physical phenomenon information corresponding to at least one of the material information and the surface shape information to the trained model, and the physical phenomenon information acquired by the acquisition unit. , It is provided with an output unit that outputs at least one of the material information and the surface shape information corresponding to the physical phenomenon information acquired by the acquisition unit.

The information processing device of FIG. 3 is, for example, the information processing device of FIG. 3 except that the input unit 220 outputs physical phenomenon information to the surface shape / material information generation unit 204 and does not include the acquisition unit 210 and the physical phenomenon information generation unit 202. It can be configured similarly to 200. In the present embodiment, the surface shape / material information generation unit 204 corresponds to the above acquisition unit, and the learned model 203, the surface shape / material information generation unit 204, and the constraint condition identification unit 205 correspond to the above output unit.

According to the present embodiment, "physical phenomenon information" as an objective variable and "material composition" as a constraint condition are given, and "surface shape information" which is image data is optimized. Can be provided. Therefore, based on a desired tactile physical phenomenon that can be expressed quantitatively, the combination of the surface shape and the material that causes the tactile physical phenomenon is determined rationally in a short time as compared with the trial and error examination by an engineer. can do. Therefore, it is possible to efficiently and rationally determine an appropriate tactile sensation.

[Example of realization by software]
The control block of the information processing device in the present embodiment may be realized by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or may be realized by software.

In the latter case, the information processing apparatus according to the embodiment of the present invention includes a computer that executes instructions of a program that is software that realizes each function. The computer includes, for example, one or more processors and a computer-readable recording medium that stores the program. Then, in the computer, the processor reads the program from the recording medium and executes it, thereby achieving the object of the present invention. As the processor, for example, a CPU (Central Processing Unit) can be used.

As the recording medium, a "non-temporary tangible medium", for example, a ROM (Read Only Memory) or the like, a tape, a disk, a card, a semiconductor memory, a programmable logic circuit, or the like can be used. Further, a RAM (Random Access Memory) for expanding the above program may be further provided.

Further, the program may be supplied to the computer via an arbitrary transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. It should be noted that one aspect of the present invention can also be realized in the form of a data signal embedded in a carrier wave, in which the above program is embodied by electronic transmission.

[Summary of Embodiment]
When a person touches an object, various physical phenomena occur at the interface due to the contact. When designing the tactile sensation of a product, design variables should be adjusted to reproduce the intended physical phenomenon when touched by a person. According to the embodiment of the present invention, "surface shape (also referred to as" grain ") information", which is image data that has been difficult in the past, is optimized and output in consideration of the influence of the "material information" to be selected. It is possible to do.

The present invention is not limited to the above-described embodiments, and various modifications can be made within the scope of the claims, and the present invention can be obtained by appropriately combining the technical means disclosed in the different embodiments. Embodiments are also included in the technical scope of the present invention.

[Modification example]
The above-described embodiment may be appropriately modified from the above-mentioned configuration, or further configuration may be added, as long as the effects of the present invention are exhibited.

For example, in the third embodiment, both the first surface shape information and the second surface shape information may be two-dimensional image data. Alternatively, one of the first surface shape information and the second surface shape information may be two-dimensional image data, and the other may be three-dimensional image data. The two-dimensional image data is data representing a plan view image drawn in a plane, and the three-dimensional image data is data representing a three-dimensional image drawn in a plane.

Further, in the third embodiment, the physical phenomenon information can be appropriately determined within a range of information representing the physical phenomenon brought about by the material information and the surface shape information.

Further, in the third embodiment, the physical phenomenon information generation unit 202 refers to the first material information and the first surface shape information acquired by the acquisition unit 201, but the first physical phenomenon is not input to the trained model 203. Information may be generated.

Further, in the third embodiment, the surface shape / material information generation unit 204 may further refer to information other than the second surface shape information and the second material information as the constraint condition, and the constraint condition specifying unit 205 may refer to the constraint condition specifying unit 205. , Such constraints may be further acquired. Examples of such additional constraints include machining accuracy and machining cost.

100, 100a, 200 Information processing device 110, 201 Acquisition unit 120 Learning unit 121 Model 130 Teacher data generation unit 202 Physical phenomenon information generation unit 203 Learned model 204 Surface shape / material information generation unit 205 Constraint condition identification unit 210 Display unit 220 Input section

Claims (9)

Teacher data including at least one of the material information which is the information about the material and the surface shape information which is the information about the surface shape, and the tactile phenomenon information which is the information about the tactile phenomenon corresponding to the at least one of the information. The acquisition department to acquire and
It is provided with a learning unit in which at least one of material information and surface shape information is input, and a model for outputting tactile phenomenon information corresponding to the input information is learned by referring to the teacher data. An information processing device that features it. The teacher data includes the material information and the tactile phenomenon information corresponding to the material information.
The model includes a regression analysis model.
The information processing apparatus according to claim 1, wherein the learning unit trains the model using the teacher data for each surface shape information. The teacher data includes the surface shape information and the tactile phenomenon information corresponding to the surface shape information.
The model includes a neural network
The information processing apparatus according to claim 1 or 2, wherein the learning unit trains the model using the teacher data for each material information. It further includes a teacher data generation unit that generates the teacher data by referring to a regression model showing the relationship between the material information and the tactile phenomenon information.
The information processing device according to claim 3. An acquisition unit that acquires first material information, which is information about the material, and first surface shape information, which is information about the surface shape.
In addition to referring to the first material information and the first surface shape information, the constraint condition using one of the second material information which is information about the material and the second surface shape information which is information about the surface shape is satisfied. In addition, an output unit that generates the other of the second material information and the second surface shape information and outputs the generated information.
Information processing device equipped with. The output unit
By inputting the first material information and the first surface shape information into the trained model, the physical phenomenon information corresponding to the first material information and the first surface shape information is specified.
The second material information and the second surface shape so as to satisfy the constraint condition using one of the second material information and the second surface shape information and to correspond to the specified physical phenomenon information. The information processing apparatus according to claim 5, which generates the other side of information. The information processing apparatus according to claim 5 or 6, wherein at least one of the first surface shape information and the second surface shape information is two-dimensional image data or three-dimensional image data. An acquisition unit that acquires at least one of material information, which is information about the material, and surface shape information, which is information about the surface shape.
By inputting the material information and the surface shape information acquired by the acquisition unit into the trained model, the output unit outputs the physical phenomenon information corresponding to at least one of the material information and the surface shape information acquired by the acquisition unit. ,
Information processing device equipped with. An acquisition unit that acquires physical phenomenon information that is information about a tactile phenomenon corresponding to at least one of material information that is information about a material and surface shape information that is information about a surface shape.
An output unit that outputs at least one of the material information and the surface shape information corresponding to the physical phenomenon information acquired by the acquisition unit by inputting the physical phenomenon information acquired by the acquisition unit into the trained model.
Information processing device equipped with.

Images