JPH06309355A - Sensibility expression translation system - Google Patents

Abstract

PURPOSE:To objectively and correctly understand a feeling word or sensibility expression by translating the feeling word or sensibility expression into information having a specific meaning on the basis of translation information which is corrected by a correcting means. CONSTITUTION:The CPU of a system controller 7 as a translating means translates the feeling word or sensibility expression having different meanings according to environment into the information (material information, etc.) having the specific means. A basic dictionary/ user dictionary storage part 2 as a storage means stores translation information (translation dictionary table) prescribing the correspondence relation between the feeling word or sensitibity expression and the object (plain cloth) denoted with the information translated by the controller 7C. Further, the system controller 7C corrects a translation error generated in the translation information owing to the difference between the filling word or sensitibity expression and the object that the information translated by the translating means indicates if they are different. Consequently, the feeling word which has a vague meaning depending upon a person who uses it can be translated into the word having the specific meaning by using a dictionary dedicated to the person.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system for translating a feeling word or a sensory / intuitive emotional expression having a different meaning (a vague meaning) for each individual into information having a specific meaning. In particular, the present invention relates to a Kansei expression translation system provided with a translation dictionary that allows a user to learn / grow while appropriately translating a translation result according to the situation after once translating a certain feeling word or Kansei expression.

[0002]

2. Description of the Related Art For example, when an apparel designer wants to order his or her desired fabric from a fabric company (fabric producer), a conversation between the designer and a sales representative (or a fabric designer) of the fabric company. In (business talks), vague and sensual expressions (designer language) are often used. Designers typically do not provide accurate physical data (or specific dough code) for a dough when ordering a dough, "light, soft, warm and fluffy"
For example, we often order fabrics with expressions that are easy to understand but have no clue (designer language), such as "raw silk type wool that has little drape but is firm and firm."

[0003]

However, the designer language has a large individual difference, and there is virtually no common language established between the designer and the sales person (or between the designers). In other words, the fact is that there is no objective common language between the person who asks for the cloth and the person who supplies the cloth, and there is no smooth communication between the two.

Therefore, even if the designer uses a vague designer word (text word) regarding the texture to explain the contents of the cloth intended by the designer to the person in charge of the cloth sales (or the cloth designer), the designer's true intention is to execute the sales correctly. Unless the person in charge (or the fabric designer) is informed and the trial production / correction of the fabric is repeated many times (or many types), the fabric desired by the designer cannot be obtained in many cases. In such a situation, the fabric having the texture intended by the designer is not easily produced, and a great deal of waste is required in the negotiation time with the designer and the cost / time of the rebuilding due to the rebuilding (reproduction of the fabric).

This may occur not only when the apparel designer orders the dough, but also in any case where communication is required using an ambiguous emotional expression (for example, a business talk involving taste, smell, hearing, etc.). It's a problem.

An object of the present invention is to provide an emotional expression translation system for objectively and correctly catching a feeling word or an emotional expression that is difficult to express with numbers or physical data.

[0007]

The sensitivity expression translation system of the present invention uses a specific meaning of a feeling word or a sensitivity expression that can have different meanings depending on the environment (different users, different times even for the same user, etc.). A translation means for translating information (material information, etc.) that has the following; translation information (translation dictionary table) that defines a correspondence relationship between the texture word or the emotional expression and the object (material) indicated by the information translated by the translation means. ) Is stored, the translation word generated by the difference is corrected (learned) when the texture word or the emotional expression is different from the object indicated by the information translated by the translation means. And correction means.

[0008]

The translation means translates the texture words or the emotional expressions into the information having the specific meaning (texture information) based on the translation information corrected by the correction means.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. Here, it is assumed that the sensitivity expression translation system of the present invention is used in an information service system targeted for the apparel industry.

FIG. 1 shows a schematic configuration of an information service system to which a sensitivity expression translation system according to an embodiment of the present invention is applied. This information service system is operated by a value-added system company (value added system company; abbreviated as VAS), and a large number of apparel makers, retailers, cloth companies, and the like that make up the apparel industry join this VAS.

VAS has its own large-scale integrated database 100. This integrated database 100 is a transaction database (transaction DB) in which information on transaction status between each subscriber is registered in real time, and an apparel database (apparel DB in which various information of old and new products created by each apparel maker is registered. ), A production information database (production information DB) in which production-related information from product arrangements to sewing is registered, and a planning information database (planning information DB) in which information on planning of new products to be sold is registered. And a management support database (management support DB) in which information about management support of each subscriber is registered, and a textile database (textile DB) in which information about textiles (fabric / textile) used for the fabric of the product is registered. including. In principle, the contents of these databases included in the integrated database 100 can be a shared property of all VAS subscribers.

The VAS integrated database 100 can be connected to the terminals 20 of a plurality of apparel makers who join the VAS via the communication line 1. The terminal 20 has a system controller 7C that controls the overall operation of the terminal. The system controller 7C has a configuration similar to that of an ordinary small computer and includes a CPU, a memory and the like.

The terminal 20 further includes an input device 5A using a keyboard, a mouse, etc., and a cathode ray tube display (C
RT) or a liquid crystal display (LCD) or the like, and a user interface 7A for connecting the input device 5A and the display device 5B to the system controller 7C.
And a storage unit (storage device) 2 that stores a basic dictionary / user dictionary that is referred to when translating a Kansei expression, and a storage unit that stores data on various dough production methods selected corresponding to the Kansei expression. (Memory device) 3, a local database (textile DB) 6 that stores data of various fabrics that can be selected according to the emotional expression, and a system controller 7
The communication interface 4 is in charge of communication between the C and the VAS 100.

The terminal 2 is connected to the interface 4.
0 has a built-in apparel ID for identifying the apparel maker in which it is installed.
When communication is carried out between and, the apparel ID is automatically added to the communication data.

The storage unit 2 includes a basic dictionary set by the Kansei expression translation system by default, and a user dictionary created by a user (a designer of an apparel company, a sales representative of a cloth company, etc.) in the course of operating the system. It is a storage device.

This storage device stores RO for information that is not rewritten (basic dictionary information set as default, etc.).
RAM or E for information that can be configured in M and is rewritten (part of the basic dictionary and user dictionary information, etc.)
EPROM (Electrically erasable programmable RO
M). These dictionaries include a list of various fabrics, various texture expressions (combination of one or more texture words), table data indicating a correspondence relationship between the cloth list and the texture expressions, and the like. The production method data storage unit 3 also contains a list of various fabrics and specific data necessary for producing these fabrics.

The storage unit 2 may be provided in an IC card, a hard disk, or an optical disk (CDROM or MO disk) together with the databases 3, 6 and the like. FIG. 2 shows an example of an input template in the case of registering information on a fabric (textile product) in VAS by quoting a texture word. This template is a VAS integrated database 10 for information about textile products.
When registering 0, it is displayed on the display device 5B of FIG.
This template is for when (registration date) where (fabric company or apparel maker ID) who (person in charge or designer name) what (fabric code, textile code,
It is used to inform the VAS computer system of the expression (brand code, etc.) and the expression (characteristic expression of the texture, candidate selection of texture expression, etc.) specified.

FIG. 2 shows, for example, an apparel designer who took a denim cloth presented by a salesman of a cloth trading company and picked it up. It is supposed to be evaluated as "suit ground".

In order to notify the system controller 7C of the terminal device 20 of the texture evaluated by the designer in the texture words, the texture expression defaulted by the texture company or the like is appropriate for representing the cloth (denim). The selection menu is output in the center of the template in FIG. Here, texture words (words) having opposite meanings are displayed as a pair of upper and lower words. The texture expression set as the default is registered in the basic dictionary of the storage unit 2.

For example, if the designer feels that the fabric (denim) is rather "thin", put a check in the head of the word "thin" in the texture expression at the left end (or instead of the check). Write the number 1). Also, if the designer feels the fabric is rather soft, put a check in the beginning of the word "soft" in the next texture expression. If the designer feels the dough is moisturized, the word "moisturized" is checked in the head, and if the designer feels the material is "textured", the word "textured" is checked in the head. .

In the texture expression selection menu of this template, the display content can be replaced with another selection candidate by pressing the arrow key on the keyboard or by clicking the triangular scroll icon with the mouse pointer. After the selection candidates are replaced in this way, the word of the texture expression displayed in the same manner as described above is appropriately selected and checked.

When the input from the above template is completed, the fabric (the denim which the designer picks up and looks at the actual product) physically / objectively specified by the fabric code and the designer feels corresponding to the fabric. Correspondence with the emotional expression (text of the characteristic of the fabric) is attached. This correspondence is registered in the user dictionary of the storage unit 2 in the form of, for example, a table.

That is, a combination (logical product) of texture words (“light”, “soft”, “moderate slimy”, “adult”, “suit”) used by the designer is selected by the designer. One-to-one correspondence is associated with a combination of character codes of texture expression (“thin”, “soft”, “moist”, “rough”). Then, the combination of these texture expression codes is added to the cloth code of the cloth (denim) that is the source of the texture words, etc.
They are associated with each other and registered in the user dictionary of the storage unit 2. The data of the cloth (denim) itself is the local database 6 and the integrated database 100 of VAS.
Be registered on both sides (or at least one side).

After the registration with the template of FIG. 2, the same designers are "light", "soft", "moderate slimy", "adult oriented", "suitable" designer words or their synonyms ( For example, software is synonymous with soft, and adult is synonymous with adult) and the material is expressed using this expression (this expression varies from designer to designer). Then, by referring to the user dictionary for the designer, the character codes of the words “thin”, “soft”, “moist”, and “rough” (the character code itself does not differ from person to person). ) Combination.

The one or more character codes obtained by translating the designer language can be used as a keyword when searching the local database 6 or the VAS integrated database 100. (That is, the designer language is a combination of character codes that does not have individual differences at the place where the communication interface 4 of FIG. 1 exits.) Alternatively, the designer can use the combination of translated character codes without searching. The intended fabric (denim) can be accurately specified.

It should be noted that the fabric company has developed its own new product fabric as shown in FIG.
In the case of newly registering in the VAS using this template, the person (for example, the sales person or the cloth designer) of the cloth company, not the apparel designer, inputs the characteristics (text) of the cloth and selects the texture expression. In this case, the apparel ID in FIG. 2 is the ID of the cloth company.

FIG. 3 shows an example of an input template in the case of registering the correspondence between a certain material and a texture word. In this example, it is assumed that a translation dictionary for texture words is created from the top down.

Now, a salesman of a cloth company is
Designer of apparel maker with D = A001 (Hanako Isho)
, Textile code 040200000030, brand code 20000, color code 30, pattern code 29
87, fabric code 025, and consider a fabric sample named "Wool Poplin". The salesman has explained that the material of this fabric is 100% wool, which is suitable for autumn and winter. In addition, the production method of the fabric named “wool poplin” specified by the above code is also determined, and the production method is the production method database 3 and / or VA.
It is stored in the production information DB of the S integrated database 100.

Designer (Hanako Isho), who took the sample of "Wool Poplin" and looked at it, said, "This doesn't have much drape and is firm and firm. It has a silky feel." Said his impressions. Then,
The salesman added to the template of FIG. 3 displayed on the apparel terminal 20 of FIG. 1 or an electronic notebook / notebook computer having the same function as this terminal, as a feature of the cloth, “there is little drape but elasticity. Type silk type wool with. Next, the salesman selects a texture expression related to the cloth from the texture expression selection menu while asking the designer a question. (The designer may input the features of the fabric and select the texture expression selection menu.) As a result, as shown in FIG. 4, the basic fabric name (wool poplin) and the features of the fabric are targeted. Correspondence between the designer words (silk-type wool with little drape but firmness and elasticity) and the corresponding texture expression (small drape; strong, silk rubbed; woolen fabric) and their codes (xxxxxx) The table is completed (in Fig. 4, the production method and other indications are omitted due to space limitations). This correspondence table becomes a part of Hanako Isho's dedicated translation dictionary. That is, referring to this translation dictionary, if Hanako Isho says "silk type wool with little drape but firmness and firmness", it is converted to a code representing small drape; firmness; silk rub; woolen fabric ( If you search for a textile database based on these codes, you will find Wool Poplin.

The input / selection operation of the texture expression for the specific fabric (wool poplin) as described above is also performed for other fabrics (flannel etc.). Then,
As shown in FIG. 5, for various fabrics, a translation dictionary table showing the correspondence between the designer words registered for each individual (Hanako Isho), the corresponding fabric names, and the production methods thereof is created.

The table in FIG. 5 functions as a dedicated translation dictionary for a designer, Hanako Isho, who belongs to an apparel maker with ID = A001. The contents of this translation dictionary (correspondence between fabric names and designer words) will differ if the fabrics are the same and the designers are different. Even if the designer is the same,
If the registration time is different, the content may be different. Further, it is possible that multiple fabrics are registered in duplicate for the same designer language.

For example, as illustrated in FIG. 5, at the time of initial registration on March 11, 1992, the only fabric that corresponds to the designer's words, "silk type wool with little drape but firmness and elasticity" is "wool poplin". It wasn't, but it was 1992 1
At the time of the 5th registration on the 1st of October (5th edition), it is possible for the same designer to additionally register the fabric "Tropical" with the designer's words "Silk type wool with less drape but firmness and elasticity". Is.

In this case, after October 1, 1992, when the textile DB is searched based on the designer word "silk type wool with little drape but firmness and elasticity", "tropical" in addition to "wool poplin" Can be found, but there is no problem. What the designer (Hanako Isho) wants is "wool poplin" or "tropical"
This is because it is sufficient for the designer to identify those samples (images and other data displayed on the terminal 20 and, if necessary, the actual product).

The designer (Hanako Isho), to whom the personalized translation dictionary having the structure shown in FIG. 4 and FIG. 5 is assigned, vaguely desires the fabric with the designer's words, "silk type wool with little drape but firmness and elasticity". As a result, after “Wool Poplin” and “Tropical” are found in the textile database 6, when the designer selects “Wool Poplin”, the salesman of the cloth company displays “Wool Poplin” of the cloth code 025 in FIG. Is ordered from the company's stock, and ID = A001
You can immediately deliver to Hanako Isho, an apparel maker.

If the company does not have the stock, the production method database 3 is searched based on the production method code (U4589) corresponding to the "wool poplin" selected by the designer, and the dough production data as shown in FIG. 6 is obtained. Comes out.
If the fabric is produced based on this production data, the fabric (designed as "Wool Poplin" of the fabric code 025) required by the designer (Hanako Isho) can be finished in one shot without repeating the trial production of the fabric.

FIG. 7 shows how the contents of the texture words used by the VAS user Hanako Isho (designer of an apparel maker with ID = A001) are decomposed / analyzed in the fabric search. At the time of October 1, 1992, Hanako Isho, who described "wool poplin" as "silk-type wool with little drape but firmness and resilience", responded to the detailed characteristics of the fabric and the corresponding characteristics after a while. It is quite possible to forget the expression used in the designer language.

Now, Hanako Isho has forgotten the fabric "wool poplin", but remembers the image of the fabric,
Suppose that an order is placed from a salesman of a fabric company with the designer's words, "It has a little drape, but it has a sharp feel and is dull and silky." Then, the salesman uses his own electronic notebook or the terminal 20 of the apparel maker to display a template such as that shown in FIG. 3 and says, "A drape is small, but it has a sharp feeling and has a glare and silk rub. Enter the characteristics of the fabric called "mono".

There is no fabric registered in Hanako Isho's translation dictionary with a word that is 100% consistent with the expression "there is a drape, but it has a sharp feel and is dull and silky-wool-like". Therefore, even if the textile database 6 is searched with this expression, "wool poplin" does not appear. However, Hanako Isho argues to the salesman that there is a certain memory of "a wool-like thing with a few drapes, but with a sharp feeling and a light and silky rubbing".

Therefore, the salesman of the cloth company decides to change the search method. The texture expressed in designer language can be decomposed into visual content, tactile content, auditory content, olfactory content and so on. So, let's break down the designer words used by Hanako Isho, "a wool-like thing with a few drapes, but with a sharp feel and light and silk rub." Then, since "low drapability" is first mentioned as the visual (appearance) content, this is the first keyword of the search. Next, as the tactile content, it is mentioned that "there is a feeling of sharpness (synonymous with strong stiffness)", and this is the second keyword of the search. Further, it is mentioned that the content of the auditory sense is "rubbing of silk", and this is the third keyword of the search. In addition, regarding the fiber structure of the fabric, "wool-like"
Since there is a word such as, the fourth keyword of the search is that it is a wool (wool) type fabric, and because there is a word of "wow light", it is the fifth keyword of the search that it is a glossy fabric. it can.

Even if the textile database 6 is searched by using all of the first to fifth keywords (under the AND condition), "wool poplin" is not always found. Therefore, reduce the keywords (that is, relax the search conditions), and first
~ Causes the CPU of the system controller 7C to search the textile database 6 for all combinations of any four of the fifth keywords. Then, "wool poplin" and "tropical" can be found from the first to fourth keywords (small drape; sharpness / strong elasticity; silk rubbing / silk type; wool-like / wool / wool fabric). Therefore, if both the "wool poplin" and "tropical" fabric samples (or fabric data including the image) are shown to Hanako Isho, "wool poplin" will be selected.

If the designer's words "a little drape, but with a dull feel and a dull and silky feel like wool" is entered using the template in FIG. 3, it will be entered into Hanako Isho's dedicated translation dictionary. It is additionally registered.
Then, after that, if you order fabric from Hanako Isho, you can find "wool poplin" in one shot with the word "wool-like thing with little drape but sharpness and glare and silk rubbing" . Also, if you order a fabric with the word "silk type wool with a few drapes, but with firmness and elasticity," you will find "tropical" in addition to "wool poplin."

FIG. 8 shows another example of the cloth search translation dictionary showing the relationship between the texture words, the cloth names corresponding to the words, and the production method thereof. In this example, the correspondence relationship between the designer word and the cloth is not registered in the dictionary in advance as in the case of FIGS. 4 and 5. In this example, it is assumed that a translation dictionary for texture words is created from the bottom up.

For example, suppose designer Hanako Isho, an apparel maker with ID = A001, is currently designing a certain coat. Hanako Isho is the design of the coat /
The atmosphere is clearly recognizable in my head, but I am not able to decide the best fabric for it.

In this case, Hanako Isho first calls a template for inputting the cloth search data as shown in FIG. 9 on the display device 5B of the apparel terminal 20 of the company. And
Entering "soft and warm feeling" as a characteristic of the fabric that has only a vague image, and using the keyboard or mouse from the texture expression selection menu as the character code corresponding to this designer word, "soft"
And "warm" is selected.

Then, the system controller 7C shown in FIG.
The CPU searches the textile DB of its own local database 6 and / or the integrated database 100 of VAS with the character code of "soft" as the first keyword. For example, it is assumed that 100 kinds of fabrics are found in the search in the first stage.

Next, the CPU of the system controller 7C
Searches for a fabric that meets the conditions, using the character code of "warm" as the second keyword from the 100 types of fabrics found. For example, it is assumed that ten kinds of fabrics are found in the second-stage search.

When the above search is completed, a window including a table as shown in the upper part of FIG. 8 pops up on the display device 5B of the terminal 20. This window displays the first two types (Melton and Flannel) of the ten types of fabrics found based on the combination of the two keywords "soft" and "warm", and at the same time, the first candidate. The material (wool / wool) of the material (melton) and the code (M8642) of the production method are displayed. The remaining eight kinds of cloths appear on the screen of the display device 5B by scrolling the candidate fields of the basic cloth name list with the arrow keys of the keyboard or the like.

If the cursor is moved to the second to tenth candidate cloth names (for example, the second candidate flannel) appearing on the screen and the return key or enter key of the keyboard of the terminal 20 is hit there. , The cloth (flannel) becomes the first candidate, and the cloth (melton) of the first candidate is carried down to the second candidate (candidate change).

This candidate change is registered in the dictionary file dedicated to Hanako Isho provided in the storage unit 2. For example, 1993 5
When the candidate change is registered for the first time at 3:30 pm on the 12th of the month, as a result, the screen of the display device 5B changes as shown in the lower part of FIG. 7.

After that, when the designer Hanako Iza of the apparel maker with ID = A001 inputs "Soft and warm feeling", "Flannel" comes to be searched first. After that, Hanako Isho searches for another fabric with the keywords "soft" and "warm" and selects and registers it. The other fabric is searched first according to Hanako Isho's designer word "soft and warm". Come on. In this way, the dictionary file (for example, the table including the contents of FIG. 8) of the specific individual (ID = A001, apparel maker Hanako Isho) can learn and grow about the relationship between the designer language and the corresponding fabric. .

When Hanako Isho finds flannel as a result of searching the textile database with the designer word "soft and warm", she can order the sample of flannel cloth found at the customer's cloth company. (If the customer's cloth company is a VAS subscriber, this order can be made by e-mail via communication line 1. In the case of this electronic order, the flannel cloth code to be ordered, The production method code and other related data can also be automatically transferred to the ordering fabric company.) The fabric company receiving the order can immediately deliver the flannel fabric sample to Hanako Isho. Unfortunately, even if there is no stock, the production method code for that fabric (F57
92) makes it possible to immediately produce flannel dough samples without trial and error.

FIG. 10 is a flow chart for explaining a procedure in which a designer searches for a desired fabric from a texture word by referring to a translation dictionary dedicated to himself at the fabric planning stage and produces it. This procedure is executed by the CPU of the system controller 7C of FIG. Here, ID = A00
It is assumed that the dedicated dictionary (Figs. 4 and 5) of Hanako Isho, an apparel maker, is already registered in the user dictionary of the storage unit 2, and she plans a fabric made of wool for autumn and winter ( Step ST10).

First, Hanako Isho uses his own apparel terminal 2
In step 0, a cloth search data input sheet having a screen configuration as shown in FIG. 3 is called. Hanako Isho describes the plan content of the fabric (image / feeling of the fabric) with his own words "silk-type wool with little drape but firmness and elasticity", and this is indicated in the "Characteristics of the fabric" section of the input sheet. (Step ST12). This input sheet also has other fields, such as the fabric code, but leave the fields for items unknown at that time blank.

The CPU of the system controller 7C decomposes the designer word "silk type wool with little drape but firmness and elasticity" having individual differences and referring to the dictionary of FIG. Expression (small drape;
Strong and strong; silk rub / silk type; woolen fabric / wool)
Translate to the combination of the character codes (step ST1
4).

Next, the CPU of the system controller 7C refers to the translation dictionary shown in FIG. 5 with the keyword of the translated character code (small drape; strong stiffness; silk rub / silk type; woolen fabric / wool) as a keyword. Production method U4
589 Wool Poplin and Tropical of Production Method T9876 Textile Database 6 and / or VA
The textile database of the S integrated database 100 is searched / discovered (step ST16).

Hanako Isho who operates the terminal 20 checks the information of one or more fabrics (for example, wool poplin and tropical) found by this search (step ST).
18). If none of the found fabrics is desired (NO in step ST20), Hanako Isho changes the search condition by changing the texture expression code selected in the template of FIG. 3 (step ST22). This change of the search condition is to change / learn the translation dictionary of Hanako Isho.

When the search conditions are changed, the combination of texture expressions after the change may not be registered in the personal dictionary (user dictionary of the storage unit 2) of Hanako Isho at that time. In this case, the general dictionary (the basic dictionary of the storage unit 2; the dictionary size is much larger than the user dictionary) in which the code corresponding to the modified texture expression is registered by default is referred to, and the combination of the modified texture expression is added to it. It is translated into a combination of corresponding expression codes (step ST14). After this translation, the textile database is searched based on the combination of the translated expression codes (step ST16), and some fabrics are found (step ST18).

When one or more fabrics (for example, Angola and Flannel) found by the above search are likely to satisfy the wishes of Hanako Isho (step ST20, Yes), Hanako Isho is found for example as the first candidate fabric. You can order (Angola) from the handling fabric company. When Angola ordered from the fabric company is not in stock,
This first candidate fabric (Angola sample) will be manufactured (step ST24).

Hanako Isho picks up an actual sample of the first candidate fabric (Angola) from the fabric company and checks whether it meets his or her wishes (step ST2).
6). If you feel that it is a little different from your wishes (Step S
T28, No), Hanako Isho checks whether there is a second candidate. If the second candidate (flannel) has been searched (step ST30, Yes), Hanako Isho selects the second candidate and upgrades it to the first candidate (step ST3).
2). This upgrade (change of candidate order) is to change / learn the translation dictionary of Hanako Isho.

Hanako Isho can order the dough (flannel) upgraded from the second candidate to the first candidate from the handling dough company. If the flannel ordered from the fabric company is not in stock, this new first candidate fabric (sample of flannel) will be produced (step ST2).
4).

Hanako Isho picks up an actual sample of a new first candidate fabric (flannel) from the fabric company and checks whether it meets his wishes (step ST26). If it still feels different from your wishes (step ST28, No), Hanako Isho checks whether there is a third candidate. If there is no third candidate (step ST3
(0, no), Hanako Isho starts over from step ST12.

If the sampled fabric (for example, flannel) meets the wishes of Hanako Isho (step ST28, YES), step ST22 or ST3
At 2 it is checked if the translation dictionary has changed.
If there is a change (step ST34, Yes), the translation dictionary (user dictionary) dedicated to Hanako Isho is updated by the change (step ST38), and the process of FIG. 10 ends. If there is no change (NO in step ST34), Hanako Isho's personalized translation dictionary (user dictionary) is left as it is, and the process of FIG. 10 ends.

FIG. 11 shows an example of a procedure for creating a translation dictionary exclusively for each designer in advance. Here, it is assumed that the material (fabric) is first, and the corresponding texture words are related later.

First, a salesman of a certain cloth company shows a sample (actual) of 20 kinds of standard cloth to the customer's apparel designer (Hanako Isho) (step ST40). Next, the salesman calls the template of FIG. 3 in the terminal device 20 or his / her electronic notebook. Then, the salesman asks how the designer feels the material for each material they are showing. That is, from the thesaurus of texture words prepared in advance by the textile company for the standard 20 kinds of textiles, the texture words that fit the feeling of the designer (Hanako Isho) are selected for each textile sample (step ST42). For example, in the example of FIG. 3, the designer Hanako Isho of the actual sample of wool poplin is asked to select a texture word such as “small drape; firmness”.

In the selection of the texture words, when there is no suitable word in the texture word thesaurus prepared in advance by the fabric company in order to represent the texture actually felt by Hanako Isho on the wool poplin at hand ( Step ST44,
No), ask Hanako Isho to give you the appropriate words.
Register it in the thesaurus of texture words (step S
T46). This thesaurus is stored in the storage unit 2 in FIG.

When an appropriate texture is additionally registered in the thesaurus in this way (step ST44, Yes), the salesman inquires Hanako Isho whether or not a material other than the standard 20 kinds is desired. If a fabric other than the standard 20 types is requested (step ST48, Yes), another fabric sample is shown to Hanako Isho (step ST50), and an appropriate texture word is selected for the fabric (steps ST42 to ST50). ).

When the texture words for all the desired textures have been selected (step ST48, No), a table (see FIG. 4) showing a correspondence relationship between each texture sample and the texture words selected from the thesaurus for the texture samples is created. The CPU of the system controller 7C creates it exclusively for Hanako Isho (step ST52).

FIG. 12 is a flow chart for explaining a procedure when a designer searches for a desired fabric from a texture word by referring to his own translation dictionary in the fabric planning stage and produces it. This procedure is a development of FIG. 10, and is executed by the CPU of the system controller 7C of FIG. Here, it is assumed that the dedicated dictionary (FIGS. 4 and 5) of Hanako Isho of the apparel maker with ID = A001 is already registered in the user dictionary of the storage unit 2.

First, a fabric having Hanako Isho is planned (step ST60). Next, she calls the cloth search data input sheet having the screen configuration shown in FIG. 3 on her apparel terminal 20. Hanako Isho describes the plan content of the fabric (image / feeling of the fabric) in her own words (for example, warm and soft wool), and enters it in the "characteristics of the fabric" field of the input sheet. Further, as described with reference to FIG. 11, a texture expression prepared in advance for the image group of the standard fabric sample or its characteristic (warm or soft) is selected from the thesaurus of texture words (step ST62).

Next, the system controller 7C of FIG.
Refers to the translation dictionary (user dictionary of the storage unit 2) and / or the basic dictionary (storage unit 2) dedicated to Hanako Isho created in the procedure of FIG. 11, for example. Translate the texture expression (designer language) (step ST64). The texture expression after translation is coded (see FIG. 4), and the textile database 6 (or VAS textile DB) is searched based on this code (step ST66).

As a result of this search, for example, Angola, Wool Georgette and Flannel shown in FIG. 5 are found. Then, the CPU of the system controller 7C
Characteristic parts that are objectively common to the three found fabrics (soft) and characteristic parts of the texture expression of those fabrics (light / soft / warm /
Check the correspondence with fluffy; hanging / soft / stiff; light / soft / warm (step ST
68).

If the characteristic part (soft) common to the three fabrics found as a result of the search does not match the characteristic part of the texture expression of those fabrics registered in the translation dictionary dedicated to Hanako Isho, step ST70, No. ), It is determined that the retrieved three fabrics are not the fabrics that Hanako Isho desires. In this case, the words used in the search (warm,
Replace "soft" with the synonyms (warm, fluffy, etc.) in the thesaurus of storage unit 2 to change the search condition, and step ST
Repeat 62 to ST70. This change of the search condition is to change / learn the translation dictionary of Hanako Isho.

If there is a place (soft) where the characteristic part (soft) common to the three fabrics found as a result of the search matches the characteristic part of the texture expression of those fabrics registered in the translation dictionary dedicated to Hanako Isho ( (Step ST70, Yes), it is determined that the three retrieved fabrics may be fabrics desired by Hanako Isho. However, this determination is based on the CPU of the system controller 7C and is not a determination by Hanako Isho.

Hanako Isho decides whether to try to make any one of the three fabrics (Angola, Wool Georgette, Flannel) found as a result of the search (if the stock is available, then order it). If it is not manufactured (or ordered) (step ST74, No), Fig. 1
The process of 2 ends.

If the first candidate (for example, Angola) of the three fabrics (Angola, Wool Georgette, Flannel) found as a result of the search is produced (or ordered) (step ST74, Yes), Hanako Isho Communicate your requirements to the producer (step ST7
6) Have the first candidate fabric (Angola with production method code A0123) manufactured (step ST78).

The manufactured (or ordered from the stock) Angora is checked by Hanako Isho (step ST80). If Hanako Isho picks up an actual sample of Angola and feels that it is different from his or her wish (step ST82, No), it checks whether there is a second candidate. If the second candidate (wool georgette) has been searched (step ST84, Yes), Hanako Isho selects the second candidate and upgrades it to the first candidate (step ST).
86). This upgrade (change of candidate order) is to change / learn the translation dictionary of Hanako Isho.

Hanako Isho can order the fabric (wool georgette) upgraded from the second candidate to the first candidate to the handling fabric company. If the wool georgette ordered from the cloth company is not in stock, the new first candidate cloth (sample of wool georgette) will be manufactured (step ST78).

Hanako Isho picks up an actual sample of a new first candidate fabric (wool georgette) from the fabric company and checks whether it meets his wishes (step ST80). If it still feels different from your wishes (step ST82, No), Hanako Isho checks whether there is a third candidate. If there is a third candidate (flannel) (YES in step ST84), steps ST74 to S74
T82 is repeated. Third candidate fabric (Flannel)
However, it is still useless (step ST82, No), and if there are no other candidate fabrics (step ST84,
No), Hanako Isho starts over from step ST62.

If the sampled fabric (for example, flannel) meets the wishes of Hanako Isho (step ST82, YES), step ST72 or ST8
At 6, it is checked whether the translation dictionary has been changed.
If there is a change (YES in step ST88), the translation dictionary (user dictionary) dedicated to Hanako Isho is updated due to the change (step ST90), and the process of FIG. 12 ends. If there is no change (No in step ST88), Hanako Isho's personalized translation dictionary (user dictionary) is left as it is, and the process of FIG. 12 ends.

After that, if the translation dictionary for Hanako Isho after the updating (learning) is referred to, steps ST74 to ST
The desired dough (for example, flannel) can be found immediately without repeating the 82 dough trial.

FIG. 13 shows another example of the procedure for creating a translation dictionary for each designer in advance. This procedure is shown in FIG.
This is a developed form of, and here, it is assumed that an object (fabric) is first associated with a texture word corresponding to it, and another word (fabric) is further associated with the word.

First, a salesman of a cloth company shows a sample of 20 kinds of standard cloth to the customer's apparel designer (Hanako Isho) (step ST100). Next, the salesman calls the template of FIG. 3 on the terminal 20 or his / her own electronic notebook, and from the texture thesaurus prepared in advance by the cloth company for the standard 20 kinds of cloth, the feeling of Hanako Isho for each cloth sample Ask the person to choose a texture word that fits in (step ST102). This thesaurus is stored in the storage unit 2 in FIG.

In selecting the texture words, Hanako Isho has an appropriate word (for example, “transparent”) to describe the texture actually felt by the sample fabric (for example, the second lace from the bottom in FIG. 5) at hand. ) Is not in the thesaurus of texture words prepared in advance by the cloth company (step ST1)
04, No), suitable texture words (transparent) for Hanako Isho
Is added, and that (transparent) is additionally registered in the thesaurus of texture words (step ST106).

In this way, when the appropriate texture words are additionally registered in the thesaurus (step ST104, Yes), the designer words (feeling fluffy with a feeling of sheerness) corresponding to Hanako Isho's feeling are associated with the fabric sample (lace). The correspondence is registered later in the translation dictionary (step ST116).

Next, the salesman inquires to Hanako Isho whether or not a material other than the standard 20 kinds is desired. Standard 20
If a material other than seed is requested (step ST108,
Yes), show another fabric sample to Hanako Isho (step ST110), and ask him to select an appropriate texture word for the fabric (steps ST102 to ST11).
0).

When the texture words for all the desired textures have been selected (step ST108, No), a translation table showing the correspondence between each texture sample and the texture words selected from the thesaurus (FIG. 4, FIG. 4). 5) is created exclusively for Hanako Isho (step ST11).
2). Up to this point, the process is similar to the example of FIG.

After the translation table is created, the combination of texture words that Hanako Isho selected for wool poplin in the past (for example, the designer words registered in the translation tables of FIGS. 4 and 5) also applies to another fabric. If you think so,
Ask her to select another fabric (for example, tropical) (step ST114). After that, the correspondence between the combination of texture words that Hanako Isho selected for wool poplin in the past and the tropical that was selected to correspond to this was additionally registered in the translation dictionary table for Hanako Isho ( Step ST116).

FIG. 14 shows an example of a learning procedure for a translation dictionary once created. In this example, it is assumed that the sensitivity of the designer (Hanako Isho) to which this translation dictionary is assigned changes between when the dictionary is registered and when the search is performed thereafter.

Designer Hanako Isho, an apparel maker with ID = A001, has registered a texture word in the texture word thesaurus stored in the storage unit 2 in order to retrieve the textile database 6 from his / her terminal 20. It is assumed that 1 or more (for example, warm with software) is input (step ST120). Then, the CPU of the system controller 7C refers to the translation dictionary having the contents shown in FIG.
Translate the texture words (soft and warm) used by Hanako Isho into the corresponding texture expression codes (step ST12).
2).

The CPU of the system controller 7C searches the textile database 6 based on the translated code (step ST124). If the flannel found as a result does not match the feeling of Hanako Isho (step ST126, No), Hanako Isho changes the search range (search condition) by changing the texture word to, for example, "soft and fluffy". Try (Step ST12
8). Then, the changed texture word is translated into another code (step ST122), and the textile database is searched by the translated texture word code (step ST124).

As a result, for example, camel hair is found. If this fabric (camel hair) matches the feeling of Hanako Isho (step ST126, Yes), the words (soft and warm) that Hanako Isho first used in the search correspond to the fabric (camel hair) that was searched. The contents of the dictionary table in FIG. 4 can be updated. (The opportunity for Hanako Isho to decide whether or not to update this is step ST1.
You can give it at 30).

FIG. 15 shows another example of the learning procedure of the translation dictionary once created. Now, designer Hanako Isho, an apparel maker with ID = A001,
In order to search the textile database 6, it is assumed that one or more registered texture words are combined (for example, soft and hot) and input into the thesaurus of the texture words stored in the storage unit 2 (step ST140). Then,
The CPU of the system controller 7C refers to the translation dictionary as shown in FIG. 8 and translates the texture words (soft and hot) used by Hanako Isho into the corresponding texture expression codes (step ST142).

The CPU of the system controller 7C searches the textile database 6 based on the translated code (step ST144). If nothing is found as a result (NO in step ST146), Hanako Isho changes the texture word to, for example, "soft and warm" (step ST140). Then, the changed texture word is translated into another code (step ST142), and the textile database is searched by the translated texture word code (step ST144).

It is assumed that Melton and Flannel are found as a result of this search (step ST146, Yes). Then, all the found fabrics are listed in the display device 5B of the terminal device 20 with a priority order (candidate order) (step ST148). First of the listed fabrics
If the candidate (melton) is intended by Hanako Isho (step ST150, Yes), the process of FIG. 15 ends.

When the first candidate (melton) out of the listed materials is not intended by Hanako Isho (step ST150, No), the second and subsequent candidate materials are checked. If there is no desired one below the second candidate (step ST152, No), step ST140-
ST150 is repeated.

If there is a desired material below the second candidate (step ST152, Yes), the material (flannel) desired by Hanako Isho among the materials below the second candidate is upgraded to the first candidate (step ST154). Then, the contents of the dictionary are changed so that the flannel becomes the first candidate for Hanako Isho's texture expression "soft and warm" (step ST156). In this way, Hanako Isho's private dictionary is learned.

FIG. 16 shows still another example of the procedure for creating and learning the translation dictionary. Now, designer Hanako Isho, an apparel maker with ID = A001, thinks about the fabric he or she imagines from the terminal 20 of his company.
It is assumed that the above-mentioned texture words have been input (step ST16).
0). If the private dictionary file for Hanako Isho has already been registered in the storage unit 2 (step ST162, Yes), the inputted texture word is translated into the corresponding texture expression code by referring to this dictionary file (step ST164). ).

If Hanako Isho's private dictionary file is not registered yet (step ST162, No), by copying the system default dictionary in the basic dictionary of storage unit 2, Hanako Isho's special dictionary file is created. It is created for the time being (step ST166). after this,
The CPU of the system controller 7C refers to the created dictionary file and translates the input texture word into a corresponding texture expression code (step ST168).

The CPU of the system controller 7C searches the textile database 6 on the basis of the translated codes of the texture words (step ST170). If the dough is not found by this search (step ST17
2, No), the processes of steps ST160 to ST170 are repeated.

When one or more doughs are found by the search at step ST170 (step ST172, Yes), all found doughs are given priority (candidate order) and listed on the display device 5B of the terminal 20 ( Step ST174). If the first candidate of the listed fabrics is the one intended by Hanako Isho (step ST17)
(6, Yes), the process of FIG. 16 ends.

The first candidate among the listed materials)
Is not what Hanako Isho intended (step ST
176, No), fabrics below the second candidate are checked. If there is no desired one below the second candidate (step S
(T178, No), steps ST160 to ST176 are repeated.

If there is a desired material below the second candidate (step ST178, Yes), the material desired by Hanako Isho among the materials below the second candidate is upgraded to the first candidate (step ST180). Then, the contents of the dictionary are changed so that the upgraded material becomes the first candidate for the texture expression of Hanako Isho (step ST182). In this way, Hanako Isho's private dictionary is learned.

FIG. 17 shows a configuration to which a sensitivity expression translation system according to another embodiment of the present invention is applied. In this embodiment, in addition to the apparel terminal 20 of FIG.
An electronic notebook (or a small and lightweight notebook personal computer) 20X having a function equivalent to that of the notebook is used.
This electronic notebook 20X is usually carried by a salesman of a cloth company.

The basic structure is the same as that of the terminal device 20 of FIG. 1, but this electronic notebook 20X has a pen input device 5X, and the databases 3 and 6 are IC cards. The translation dictionary data of each designer created by the electronic organizer 20X is transmitted to the terminal device 20 via the communication cable 1X or the like.
It can be collected at. If the electronic notebook 20X has a floppy disk drive, the translation dictionary data can be delivered and received by using a floppy disk.

According to the present invention, even if an ambiguous hand-written word is used as an expression means, there is a translation means for the word, so that between one party (apparel designer) and the other party (textile sales / textile designer). Communication will be surely carried out. Although the apparel industry is assumed in the above description, the present invention can be applied not only to the apparel industry but also to other industries in which texture expression is involved, for example, the cosmetics industry.

[0106]

According to the emotional expression translation system of the present invention, an ambiguous feeling word (and a combination of a plurality of feeling words) having different meanings depending on the person using the word (feeling) having a specific meaning (texture) Expression code). Even if the pre-translation word has no objectivity, the post-translation word (the code that indicates the object / information that corresponds to the word in a one-to-one correspondence) has its meaning specified and is objective. There is no individual difference in the latter words. Therefore, the translated word can be subjected to a uniform processing by a computer, and the content intended by the original texture word can be accurately transmitted to a third party.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration to which a sensitivity expression translation system according to an embodiment of the present invention is applied.

FIG. 2 is a diagram showing an example of an input template in the case of registering information on a fabric (textile product) by quoting a texture word.

FIG. 3 is a diagram showing another example of the input template in the case of registering the information on the cloth by quoting the texture words.

FIG. 4 is a diagram showing an example of a translation dictionary showing correspondence relationships between texture words (designer words) registered for each individual, texture expression codes corresponding to the texture words, and cloth names corresponding to the texture expression codes.

FIG. 5 is a diagram showing an example of a translation dictionary showing a correspondence relation between a texture word registered for each individual, a corresponding fabric name, and a production method thereof.

FIG. 6 is a diagram showing an example of production data of fabrics listed in the translation dictionary of FIG. 4 or 5;

FIG. 7 is a diagram for explaining how the contents of a feeling word used by a designer (user) are decomposed / analyzed.

FIG. 8 is a diagram showing another example of a translation dictionary showing a relationship between a texture word, a corresponding fabric name, and a production method thereof.

FIG. 9 is a diagram showing an example of a data input sheet used when searching for fabrics in a designer language.

FIG. 10 is a flowchart illustrating an example of a procedure for searching for a desired material from a texture word with reference to a translation dictionary and manufacturing the material.

FIG. 11 is a flowchart illustrating an example of a procedure for creating a translation dictionary.

FIG. 12 is a flowchart illustrating another example of a procedure for searching for a desired material from a texture word with reference to a translation dictionary and producing the material.

FIG. 13 is a flowchart illustrating another example of a procedure for creating a translation dictionary.

FIG. 14 is a flowchart illustrating an example of a learning procedure of a translation dictionary.

FIG. 15 is a flowchart illustrating another example of a translation dictionary learning procedure.

FIG. 16 is a flowchart illustrating yet another example of a procedure for creating and learning a translation dictionary.

FIG. 17 is a block diagram showing a configuration to which a sensitivity expression translation system according to another embodiment of the present invention is applied.

[Explanation of symbols]

1, 1X ... communication line, 2 ... basic dictionary / user dictionary storage, 3 ... production method database storage, 4 ...
..Communication interfaces, 5A ..., Input devices, 5B
..Display device, 5X ... Pen input device, 6 ... Local database (textile DB), 7A ... User interface circuit, 7C ... System controller, 20 ... Apparel terminal, 20X ... Electronic notebook / notebook computer, 100 ... Integrated database (information service company).

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Shinichi Yoshii 4-20-19 Minamiaoyama, Minato-ku, Tokyo Inside Yubus Co., Ltd.

Claims (1)

[Claims] 1. A translation means for translating a feeling word or a feeling expression that may have different meanings depending on the environment into information having a specific meaning; the feeling word or a feeling expression, and the information translated by the translation means Storage means for storing translation information that defines the correspondence with the target;
When the texture word or the kansei expression and the object indicated by the information translated by the translation means are different, the translation means comprises a correction means for correcting a translation error generated in the translation information, and the translation means, A kansei expression translation system, characterized in that, based on the translation information modified by the modifying means, the feeling words or the kansei expressions are translated into the information having the specific meaning.