WO2022097622A1

WO2022097622A1 - Odor analyzing device, odor analysis method, and odor analysis program

Info

Publication number: WO2022097622A1
Application number: PCT/JP2021/040337
Authority: WO
Inventors: 広明松岡; 祐子元日田
Original assignee: 株式会社レボーン
Priority date: 2020-11-06
Filing date: 2021-11-02
Publication date: 2022-05-12
Also published as: JPWO2022097622A1

Abstract

An odor analyzing device according to one aspect of the present invention is provided with: a measuring unit including n (where n is an integer at least equal to 1) odor sensors; an acquiring unit for acquiring measured values measured by the n odor sensors of the measuring unit; an extracting unit for extracting m (where m is an integer at least equal to 1, and n≥m) main components from among n-dimensional data having the measured value from one odor sensor, acquired by the acquiring unit, as one dimension; a projecting unit for projecting the n-dimensional data in m dimensions on the basis of the main components extracted by the extracting unit; and a rendering unit for rendering the m-dimensional projection data projected by the projecting unit.

Description

Odor analyzer, odor analysis method and odor analysis program

The present invention relates to an odor analyzer, an odor analysis method, and an odor analysis program.

Conventionally, there has been an odor sensor that detects an odor component by depositing a thin film on a crystal oscillator and changing the frequency. For example, Patent Document 1 includes a plurality of heads provided with an organic thin film that selectively adsorbs an odorous substance on at least one surface of a large number of crystal oscillators and surface elastic wave elements via an electrode. , A technique relating to a multi-head type odor sensor for determining an odor substance and an odor quality from a change in the natural vibration of an individual head is disclosed.

Japanese Unexamined Patent Publication No. 05-346384

However, in the conventional technique, although a plurality of measured values measured by a plurality of odor sensors are displayed, it is not possible to determine which odor sensor's measured value should be focused on, so it is possible to recognize the characteristics of the odor to be measured. It could be difficult.

The present invention has been made in view of the above circumstances, and one object of the present invention is to provide an odor analyzer, an odor analysis method, and an odor analysis program that can easily recognize the characteristics of odors.

(1) In order to solve the above problem, the odor analyzer acquires the measured values measured by the measuring unit having n odor sensors (n is an integer of 1 or more) and the n odor sensors of the measuring unit. The main components of m pieces (m is an integer of 1 or more and n ≧ m) from the n-dimensional data whose one dimension is the measured value of one odor sensor acquired by the acquisition unit and the acquisition unit. An extraction unit that extracts n-dimensional data based on the main components extracted in the extraction unit, a projection unit that projects n-dimensional data in m-dimensional form, and a drawing unit that draws m-dimensional projection data projected in the projection unit. Be prepared.

(2) Further, in the odor analyzer of the embodiment, the odor sensor changes the natural frequency due to the adsorption of an odorous substance on the adsorption unit, and the measurement unit measures the natural frequency. May be good.

(3) Further, in the odor analyzer of the embodiment, the projection unit projects the n-dimensional data into the m-dimensional space represented by the principal component vectors of m principal components. You may.

(4) Further, in the odor analyzer of the embodiment, the extraction unit extracts the main component so that the difference between the projection data in the first odor sample and the projection data in the second odor sample is maximized. May be.

(5) Further, in the odor analyzer of the embodiment, the drawing unit may draw by changing the drawing scale of the projection data for each dimension.

(6) Further, in the odor analyzer of the embodiment, the drawing unit may draw the projection data of the reference odor sample and the projection data of the odor sample having a predetermined difference so as to be distinguishable.

(7) In order to solve the above problems, the odor analysis method consists of a measurement step of measuring odor with n odor sensors (n is an integer of 1 or more) and an odor analysis method of n odor sensors in the measurement step. Extract m (m is an integer of 1 or more) principal components from the acquisition step for acquiring the measured value and the n-dimensional data with the measured value of one odor sensor acquired in the acquisition step as one dimension. It includes an extraction step, a projection step of projecting n-dimensional data in m-dimensional based on the main component extracted in the extraction step, and a drawing step of drawing m-dimensional projection data projected in the projection step.

(8) In order to solve the above problems, the odor analysis program uses n odor sensors (n is an integer of 1 or more) to measure the odor on the computer, and n odor sensors in the measurement process. The main components of m (m is an integer of 1 or more) from the acquisition process for acquiring the measured measured values and the n-dimensional data in which the measured value of one odor sensor acquired in the acquisition process is one-dimensional. The extraction process that extracts the Let the computer do it.

According to one embodiment of the present invention, the odor sensor has n (n is an integer of 1 or more), the measured values measured by the n odor sensors are acquired, and the acquired odor sensor is used. M (m is an integer of 1 or more) principal components are extracted from the n-dimensional data whose measured value is one-dimensional, and the n-dimensional data is projected onto the m-dimensional based on the extracted principal components. By drawing the m-dimensional projection data, it is possible to easily recognize the characteristics of the odor.

It is a block diagram which shows an example of the structure of the odor analyzer in Embodiment. It is a figure which shows an example of the appearance of the odor analyzer in an embodiment. It is a figure which shows the 1st display example of the odor analyzer in an embodiment. It is a figure which shows the 2nd display example of the odor analyzer in an embodiment. It is a block diagram which shows an example of the hardware composition of the odor analyzer in Embodiment. It is a flowchart which shows 1st example of the operation of the odor analyzer in Embodiment.

Hereinafter, the odor analyzer, the odor analysis method, and the odor analysis program according to the embodiment of the present invention will be described in detail with reference to the drawings.

First, the configuration of the odor analyzer will be described with reference to FIG. FIG. 1 is a block diagram showing an example of the configuration of the odor analyzer according to the embodiment.

In FIG. 1, the odor analyzer 1 has a function unit of a measurement unit 11, an acquisition unit 12, an extraction unit 13, a projection unit 14, and a drawing unit 15. Each functional part of the odor analyzer 1 will be described as being realized by software. The odor analyzer 1 is, for example, a desktop PC, a notebook PC, a smartphone, a tablet PC, or a head-mounted, glasses-type, or watch-type device. The odor analyzer 1 is communicably connected to the server 2 via the network 9.

The measuring unit 11 has n odor sensors for measuring odors (however, n is an integer of 1 or more). Smell sensors include semiconductor type such as oxide semiconductor type and organic semiconductor type, crystal oscillator type using epoxy resin film, vinyl acetate resin film, Langmuir Projet film, etc. as the sensitive film, and surface acoustic wave. Various ones such as those using a (SAW: Surface Acoustic Wave) filter and a bulk surface acoustic wave (FBAR: Film Bulk Acoustic Resonator) filter can be used. In the following, a case where n odor sensors for measuring the natural frequency that changes due to the adsorption of an odorous substance on the adsorption portion will be described.

The adsorption part is the part that adsorbs odorous substances in the odor sensor. The adsorption portion can be implemented by providing a thin film on a vibrating body such as a crystal oscillator or a piezoelectric element. The thin film can be formed by, for example, thin film forming means such as thin film deposition or sputtering.

The material used for the thin film can be composed of, for example, at least one of amino acids, sugars, bases, aromatic ketones, dihydric alcohols and phenols.

Amino acids include, for example, valine, isoleucine, leucine, methionine, lysine (lysine), phenylalanine, tryptophan, threonine (threonine), histidine, arginine, glycine, alanine, serine, tyrosine, cysteine, aspartic acid, glutamic acid, proline, aspartic acid. , Glutamic acid and the like can be used.

Examples of sugars include α-glucose (dextrose), β-glucose (dextrose), fructose (fruit sugar), galactose, sucrose (sucrose), lactose (lactose), maltose (maltose), and cellobiose amylose (dempun). , Sucrose and the like can be used.

As the base, for example, adenine, guanine, cytosine, uracil, thymine and the like can be used.

The aromatic ketones include, for example, 3,4-methylenedioxypropiophenone, 3-acetyl-6-methoxybenzaldehyde, acridone, 2-acetyl-5-methylfuran, acetylpyrazine, acetylpyridine, and 2-acetyl. Ppyridine, 2-acetylfuran, acetoanisol, acetophenone, acetohexamide, acepromazine, amyodaron, rectinib, antron, isomaltol, indigocarmine, indigo, evastin, emerison, catinone, gallacetophenone, chalcone, xanthone, quinurenin, chloroacetophenone, chloro Facinone, Ketanserin, Saltoprofen, Diazonaphthoquinone, Divapron, Dihydromenaquinone, Sprofen, Seriprolol, Tamthrosin, Thiaprofenic acid, Desaspidine, Tonalide, Tricostatin A, Droperidol, Niticinone, Valerophenone, Pisein, Piseol, Vitamin K , Fantride, phylloquinone, phenylglycol, phenylglycoxylic acid, bupropion, flubendazole, protopin, propaphenone, butyrophenone, propiophenone, floridine, floretin, pungenin, versalide, benziodalone, benzyl, benzanthron, benzbromalon, benzoyl group, Benzoyl, benzophenone, aromatic polyether ketone, N-formylquinurenin, muskketone, methilapon, methylmenaquinone, methocatinone, mebendazole, ramosetron, laroxyphene, roberin and the like can be used.

The dihydric alcohols include dihydric alcohols 1,2-ethanediol, ethylene glycol, 1,2-propanediol, 1,3-propanediol, 2-chloro-1,3-propanediol, and 3-chloro-. 1,2-propanediol, 1,2-butanediol, 1,3-butanediol, 1,4-butanediol, 2,3-butanediol, 2-methyl-1,2-propanediol, 1,5- Pentandiol, 2-methyl-2,3-butanediol, 1,6-hexanediol, 2,5-hexanediol, 2-methyl-2,4-pentanediol, 2,3-dimethyl-2,3-butane Diol, 2-butane-1,4-diol and the like can be used.

As the phenols, phenol, dibutylhydroxytoluene, bisphenol A (BPA), cresol, estrazil, eugenol, guaiacol, picric acid, phenolphthaline, serotonin, dopamine, adrenaline, noradrenaline, timol, tyrosine and the like should be used. Can be done.

The suction part provided with a thin film on the vibrating body has its own natural frequency. When an odorous substance is adsorbed on the thin film provided on the vibrating body, the natural frequency of the adsorbed portion changes. As the amount of odorous substances adsorbed on the thin film increases, the change in natural frequency increases. The odor sensor can measure the amount of odorous substance adsorbed by outputting the natural frequency according to the amount of odorous substance adsorbed.

The measuring unit 11 measures the measured values of n odor sensors. Here, n is an integer of 1 or more. In the present embodiment, a case where the number of odor sensors is nine will be described later, but the number of odor sensors is arbitrary. In this embodiment, the measurement unit 11 does not include an odor sensor, which is hardware, in order to illustrate the case where the measurement unit 11 is implemented by software. However, the measuring unit 11 may be mounted as a sensor unit including an odor sensor.

The acquisition unit 12 acquires the measured values measured by the n odor sensors of the measurement unit 11. For example, when the number of odor sensors is 9, the acquisition unit 12 acquires the measured values of the 9 odor sensors. The acquisition unit 12 may acquire the measured values at a plurality of measurement times. For example, when odors are measured every 10 seconds by nine odor sensors, the acquisition unit 12 acquires 9 × 6 = 54 measured values per minute. In the present embodiment, the term "acquisition" may be a pull-type reception or a push-type reception. Further, when the term "provided" is used, it may be a push-type transmission or a pull-type transmission.

The extraction unit 13 extracts m principal components from the n-dimensional data in which the measured value of one odor sensor acquired by the acquisition unit 12 is one-dimensional (however, m is an integer of 1 or more). There, n ≧ m). For example, when there are nine odor sensors, n = 9, and the extraction unit 13 can extract 9-dimensional data. The extraction unit 13 extracts m (for example, 3) odor sensor measurement values from the 9 odor sensors as the main component.

Here, the principal component is a component that characterizes the odor to be measured, and is a measured value measured by a specific odor sensor. For example, when comparing the odors of sample A and sample B, the principal component is a component that characterizes and distinguishes the odor of sample A and the odor of sample B, respectively. For example, when distinguishing between the odor of mandarin orange and the odor of lemon in nine odor sensors, it is assumed that the seven odor sensors output measured values approximated as the same citrus odor. On the other hand, in the two odor sensors, it is assumed that the odor of mandarin orange and the odor of lemon are distinguished and the measured value is output. In this case, the main component is the measured value of the above two odor sensors. By extracting these two measured values as the main components, it is possible to characterize and distinguish the odors of mandarin oranges and lemons. By extracting the main component, it is possible to easily recognize the characteristics of the odor.

The extraction of the main component is performed by extracting m measured values as the main component from the n measured values. Extraction of the principal component may be performed based on a predetermined setting (setting of which measured value is to be extracted). For example, when the sample to be measured is decided, the main component can be decided in advance. For example, when measuring a change in odor over time of a certain sample, the measured value of an odor sensor that can easily detect the change in odor over time can be determined in advance as a main component. Further, when distinguishing between a real perfume and a fake perfume, the measured value of the odor sensor, which tends to make a difference between the real perfume and the fake perfume, can be determined in advance as the main component. The extraction unit 13 can perform rapid measurement by extracting m predetermined main components.

On the other hand, the extraction of the principal component may be dynamically changed. For example, if the odor of a sample cannot be predicted, it is not possible to determine in advance which odor sensor's measured value is the main component. For example, the extraction unit 13 may be dynamically determined based on the projection data described later. Further, the extraction unit 13 may infer the principal component from the learning result of machine learning of the past data of the sample type and the principal component suitable for the sample type. For example, when measuring the rotten odor of food, the principal component is predicted from the component of the food to be measured by machine learning the combination of the component of the food measured in the past and the component suitable for it as teacher data. You may do it.

The projection unit 14 projects n-dimensional data to m-dimensional based on the principal component extracted by the extraction unit 13 (however, n ≧ m). In the present embodiment, principal component analysis (PCA: Principal Component Analysis) may be performed based on the principal component extracted by the extraction unit 13.

Principal component analysis is a multivariate analysis that synthesizes a small number of uncorrelated variables that best represent the overall variation from a large number of correlated variables. By performing the principal component analysis, the n-dimensional measurement data can be made m-dimensional, and the characteristics of the data can be easily grasped.

The projection unit 14 projects the measured values in the nth-order space onto a vector space composed of an m-dimensional principal component vector (principal component vector) extracted by the extraction unit 13, and vector data of the principal component vector. Generates projection data represented by. For example, when n = 9 and m = 3, the nine-dimensional measured value is projected onto the three-dimensional vector space, and the projected data can be visualized. Further, when m = 2, the nine-dimensional measured value is projected onto the two-dimensional vector plane, and the projected data can be visualized. Further, when m = 1, the nine-dimensional measured value is projected onto the one-dimensional vector straight line, and the projected data can be visualized. By projecting the measured values of the n-dimensional odor sensor into 1 to 3 dimensions to generate projection data, the measured data can be visually recognized and the characteristics of the odor can be easily recognized.

The extraction unit 13 extracts the main component so that the difference between the projection data in the first odor sample and the projection data in the second odor sample is maximized. As described above, the projection data is an m-dimensional principal component vector on which measured values other than the principal components are projected, and is m-dimensional vector data. The difference in the projection data is the difference between the m-dimensional vector data, and can be calculated by subtracting the m-dimensional determinant. Since the projection data shows the characteristics of the odor, the maximum difference from the projection data means the maximum difference in the characteristic odor. That is, the extraction unit 13 extracts the main component so that the difference between the odor of the first odor sample and the odor of the second odor sample is recognized as the maximum. For example, when n = 9 and m = 3, there are 9 × 8 × 7 = 504 combinations of principal components. The extraction unit 13 can calculate the difference in the projection data in each combination and extract the main component of the combination in which the difference is maximum. As for the combination of the principal components, for example, the difference may be calculated by giving a priority based on the past calculation result.

The drawing unit 15 draws the m-dimensional projection data projected by the projection unit 14. Drawing is rendering an image based on the projected data. For example, when m = 1, the drawing unit 15 performs one-dimensional rendering. Further, when m = 2, the drawing unit 15 performs two-dimensional rendering. Further, when m = 3, the drawing unit 15 performs three-dimensional rendering. When m is 4 or more, the drawing unit 15 renders by a predetermined drawing method. The predetermined drawing method is, for example, a drawing method of a figure using a table, a figure, a symbol, coloring, or the like. For example, the drawing unit 15 may represent the four-dimensional projection data by changing the size of the points mapped to the three-dimensional graph. Similarly, the drawing unit 15 may represent the five-dimensional projection data by changing the size and color of the points mapped to the three-dimensional graph. Further, the drawing unit 15 may plot a figure (for example, a figure such as a circle, a square, a triangle, a star, or a character) or a symbol (for example, a JIS symbol) on the graph instead of the point. The rendering of the projection data may include a rendering of a moving image. For example, the drawing unit 15 may render a moving image that dynamically changes the shape or color of the plotted points. Further, the rendering of the projection data may include a rendering of the sound. For example, the drawing unit 15 may render a voice that changes the type and volume of the sound according to the projection data.

The drawing unit 15 changes the drawing scale of the projection data for each dimension and draws. The change of the drawing scale is, for example, the scale of the graph axis. For example, when it is desired to recognize a change in a measured value in a specific principal component vector in detail, the drawing unit 15 enlarges the scale of the graph axis of the principal component vector to make it easier to recognize a smaller change in the measured value. You may do so. Further, the drawing unit 15 may increase the amount of change in the size of the point related to the specific principal component vector to make it easier to recognize a smaller change in the measured value.

The drawing unit 15 may draw the projection data of the reference odor sample and the projection data of the odor sample having a predetermined difference so as to be distinguishable. For example, when the odor analyzer 1 is used to discriminate between a genuine perfume and a fake perfume, the drawing unit selects an odor sample whose projection data has a predetermined difference from the projected data of the reference genuine perfume odor sample. It may be determined to be a fake and drawn. For example, the drawing unit 15 may express the measured value in a specific color or sound an alarm from a speaker (not shown) when it is determined to be a fake.

The setting providing unit 21 provides the odor analyzer 1 with the setting information related to the analysis. The setting information related to the analysis is, for example, information related to the extraction method of the principal component (for example, which sensor's measured value is the principal component, or whether the calculation is preferentially performed as the principal component, etc.). The setting information related to the analysis may be the method of drawing the projection data described above (for example, the drawing angle of the graph, the drawing scale changing method, the coloring method, etc.).

It should be noted that each of the above-mentioned functional units possessed by the odor analyzer 1 shows an example of the function, and does not limit the function possessed by the odor analyzer 1. For example, the odor analyzer 1 does not have to have all the above-mentioned functional parts, and may have some of the functional parts. Further, the odor analyzer 1 may have a function other than the above. For example, the functional unit of the odor analyzer 1 may be realized in the server 2.

Further, as described above, each of the above functional parts has been described as being realized by software. However, at least one or more functional units in the above functional units may be realized by hardware.

Further, any of the above functional units may be implemented by dividing one functional unit into a plurality of functional units. Further, any two or more of the above functional units may be integrated into one functional unit. FIG. 1 shows the functions of the odor analyzer 1 as functional blocks, and does not show, for example, that each functional unit is composed of a separate program file or the like.

Further, the odor analyzer 1 may be a device realized by one housing or a system realized by a plurality of devices connected via a network or the like. For example, the odor analyzer 1 may realize a part or all of its functions by another virtual device such as a cloud service provided by a cloud computing system. That is, the odor analyzer 1 may realize at least one or more of the above-mentioned functional parts in another device.

Next, a display example of the display device of the odor analyzer 1 will be described with reference to FIG.

FIG. 2 is a diagram showing an example of the appearance of the odor analyzer 1 in the embodiment.

In FIG. 2, the odor analyzer 1 has an odor sensor mounting unit 16. The odor sensor mounting unit 16 shown in FIG. 2 illustrates a case where nine odor sensors 161 are mounted. In FIG. 9, a code is assigned to only one sensor, and the code is omitted from the other sensors. The number n of the odor sensors registered in the odor analyzer 1 is an arbitrary integer of 1 or more, and may be, for example, 32 odor sensors. No, FIG. 2 illustrates an odor analyzer 1 in which an odor sensor and other components are mounted on one substrate, but the odor analyzer 1 may have an odor sensor mounting unit 16 independently on another substrate. good.

The odor sensor 161 mounted on the odor sensor mounting unit 16 is a sensor provided with different types of thin films. However, the odor sensor mounting unit 16 may be provided with a plurality of odor sensors 161 provided with the same thin film. For example, when four odor sensors 161 having the same type of thin film are mounted in the odor sensor mounting section 16 of n = 8, the number of odor sensors 161 mounted is 8 × 4 = 32. For example, by acquiring the measured value by using the measured value of the odor sensor provided with the same thin film as an average value or the like, it is possible to reduce the error of the measured value due to the individual difference of the odor sensor 161. Further, by mounting a plurality of odor sensors 161 provided with the same thin film, it is possible to reduce the measurement error due to the mounting position (position for measuring the odor) of the odor sensor 161.

Next, a display example of the projection data drawn by the odor analyzer 1 will be described with reference to FIGS. 3 to 4. FIG. 3 is a diagram showing a first display example of the odor analyzer according to the embodiment.

In FIG. 3, the display screen 1000 displays a graph of two-dimensional projection data in which measured values are projected onto a two-axis principal component vector, where the number m of the principal components is 2 (m = 2). The display screen 1000 is a display example in which the drawing scale of the projection data is changed for each dimension and drawn. The drawing scale is a display magnification of the graph axis, and the display screen 1000 enlarges and displays a specific part of the graph axis. For example, in FIG. 3, it is assumed that the projection data of the main components are distributed between about -2 and 2 (not shown). The display screen 1000 enlarges and displays the portion of −0.75 to 1.5 on the drawing scale of the graph axis on the horizontal axis. Further, the display screen 1000 enlarges and displays the portion of −0.75 to 1.25 on the drawing scale of the graph axis on the vertical axis. By changing the drawing scale of the projection data for each dimension and drawing, it is possible to display the projection data by paying attention to the portion showing the characteristics of the odor, and it is possible to make it easier to visually recognize the characteristics of the odor.

The display screen 1000 shows the measured values of the odor sample A, the odor sample B, and the odor sample C. The odor sample A is a genuine perfume that is manufactured and sold in a regular manner. The odor sample B and the odor sample C are perfumes to be determined whether they are genuine or fake. The measured values of the odor sample A are approximately distributed in the range of the measured value group 1002 with the center point 1001 as the center. The method of calculating the center point in this embodiment uses an arithmetic mean calculated from the average of the simple sums of the measured values. However, as a method for calculating the center point, a weighted average, a geometric mean, a harmonic mean, or the like may be used. Further, among a plurality of sensors, the measured value of a specific sensor may be weighted to calculate the center point. For example, the measured values of three specific odor sensors among the 19 odor sensors may be weighted twice as much as the measured values of the other odor sensors to calculate the center point. The diameter of the center point 1001 indicates the magnitude of the variation. The measured values of the odor sample B are approximately distributed in the range of the measured value group 1004 centered on the center point 1003. The diameter of the center point 1003 indicates the magnitude of the variation. Further, the measured values of the odor sample C are approximately distributed in the range of the measured value group 1010 centering on the center point 1009. The diameter of the center point 1010 indicates the magnitude of the variation.

The display screen 1000 is drawn so that the projection data of the reference odor sample A and the projection data of the odor sample B having a predetermined difference can be discriminated from each other. For example, since the measured value group 1002 and the measured value group 1004 do not overlap, the user can recognize that there is a difference between the measured values. Further, since the magnitude of the variation of the measured value group 1002 and the magnitude of the variation of the measured value group 1004 are different, the user can recognize that there is a difference between the measured values. Further, from the distance between the center point 1001 and the center point 1003, the user can recognize that there is a difference between the measured values. Differences in measured values such as the degree of overlap of the measured value groups, the magnitude of variation, or the distance between the center points can be calculated as numerical values. For example, the degree of overlap of the measured value groups can be calculated by a statistical value called an effect size obtained by dividing the difference between the average values of the measured values by the magnitude of the variation (standard deviation, etc.). Similarly, the display screen 1000 draws the projection data of the reference odor sample A and the projection data of the odor sample C having a predetermined difference so as to be distinguishable.

The display screen 1000 may display the warning display 1005 when the difference between the above-mentioned measured values is equal to or more than (or exceeds) a predetermined predetermined value. The warning display 1005 is, for example, a warning text indicating that the determination target is a fake. The warning display 1005 may change the content of the warning display according to the magnitude of the difference in the measured values. For example, when the determination of whether the difference between the measured values is genuine or fake is in the vicinity of a delicate predetermined value, the warning display 1005 is a display prompting remeasurement or "same by probability of aa%" or "effect size bb" or the like. May be displayed. The user can recognize that there is a difference between the measured values by the warning display. When the difference between the measured values described above is less than (or less than or equal to) a predetermined value, the display screen 1000 is, for example, "same", "genuine", "same with probability of cc%" or "same". You may display "effect size dd" or the like.

Further, the display screen 1000 can display the projection data of a plurality of odor samples so that they can be relatively compared with each other. For example, if the odor of a genuine perfume varies between production lots or the odor changes over time, the projection data of the reference genuine product will change. It may not be possible to judge whether or not the product is genuine based on the difference (absolute value) alone. The display screen 1000 displays each of the projection data of the odor sample A, the projection data of the odor sample B, and the projection data of the odor sample C, and the projection data of the odor sample A and the projection data of any odor sample are displayed on the graph. It is possible to display whether the data is relatively close to each other so that the user can recognize it.

For example, the display screen 1000 can display that the distance between the center point 1001 and the center point 1003 is relatively large compared to the distance between the center point 1001 and the center point 1009. The user can recognize the relative comparison between the distance between the center point 1001 and the center point 1003 and the distance between the center point 1001 and the center point 1009, not the absolute value of the distance between the center point 1001 and the center point 1003. can. As a result, the user can recognize that the difference in the measured values between the odor sample A and the odor sample B is relatively large compared to the difference in the measured values between the odor sample A and the odor sample C. Similarly, the display screen 1000 can display that the distance between the measured value group 1002 and the measured value group 1004 is relatively large compared to the distance between the measured value group 1002 and the measured value group 1010. As a result, the user can recognize that the difference in the measured values between the odor sample A and the odor sample B is relatively large compared to the difference in the measured values between the odor sample A and the odor sample C. As a result, the user can recognize that the odor sample B is a fake and the odor sample C is a genuine product. If it is known from the beginning that the odor sample C is a genuine product having a different production lot or production date, it is possible to determine the odor sample B in consideration of the variation in the production lot and the like. .. Further, the display screen 1000 can display projection data even if the number of odor samples is 4 or more.

Further, the display screen 1000 has a button 1006, a button 1007, and a button 1008. Button 1006 is a button for displaying one-dimensional projection data (data distribution on a straight line) in which a measured value is projected on a one-axis principal component vector, where the number m of the principal components is 1 (m = 1). be. Button 1007 is a button for displaying two-dimensional projection data (data distribution on a plane) in which the measured values are projected onto a two-axis principal component vector, where the number m of the principal components is 2 (m = 2). be. Further, the button 1008 is a button for displaying three-dimensional projection data (data distribution in space) in which the measured value is projected on the three-axis principal component vector, where the number m of the principal components is 3 (m = 3). Is. FIG. 3 shows that the button 1007 is pressed.

Note that FIG. 3 illustrates a method of calculating the average value of the measured values and comparing a plurality of odor samples by comparing the distances of the center points. However, the comparison of odor samples may use a calculation method other than the calculation of the average value. For example, the total value of the measured values may be calculated and the total value may be compared. The total value may be calculated by, for example, simple addition, or may be calculated by weighting the measured value of a specific odor sensor.

FIG. 4 is a diagram showing a second display example of the odor analyzer according to the embodiment.

FIG. 4 displays a graph of two-dimensional projection data in which measured values are projected onto a two-axis principal component vector, where the number m of the principal components is 2 (m = 2), as in FIG. Further, the display screen 2000 enlarges and displays the portion of -1.0 to 1.2 on the drawing scale of the graph axis on the horizontal axis and -0.75 to 1.1 on the drawing scale of the graph axis of the vertical axis. ing.

The display screen 2000 shows a measured value showing a change in odor of odor sample A over time and a measured value showing a change in odor of odor sample B over time. The odor sample A is a food product to which no additive is added, and the odor sample B is a food product to which an additive is added. That is, the display screen 2000 displays the difference in the change in the measured value indicating the change in the odor depending on the presence or absence of the additive.

Further, the display screen 2000 has a button 2003, a button 2004, and a button 2005. Since the functions of the button 2003, the button 2004, and the button 2005 are the same as the functions of the button 1005, the button 1006, and the button 1007 described in FIG. 3, the description thereof will be omitted.

The measured values of the odor sample A vary within the range of the measured value group 2001. Further, in the odor sample B, the measured values vary within the range of the measured value group 2002. The display screen 2000 shows that the measured value group 2001 and the measured value group 2002 do not overlap and the variations are different. From the display on the display screen 2000, the user can visually recognize that there is a difference in the change in odor due to the addition of the additive to the food.

Next, the hardware configuration of the odor analyzer 1 will be described with reference to FIG. FIG. 4 is a block diagram showing an example of the hardware configuration of the odor analyzer 1 according to the embodiment. The hardware configuration of the server 2 is similar to that of the odor analyzer 1, and the description thereof will be omitted.

The odor analyzer 1 has a CPU (Central Processing Unit) 101, a RAM (Random Access Memory) 102, a ROM (Read Only Memory) 103, an I / O device 104, and a communication I / F (Interface) 105. The odor analyzer 1 is an apparatus that executes the information processing program described with reference to FIG.

The CPU 101 controls the user terminal by executing the information processing program stored in the RAM 102 or the ROM 103. The information processing program is acquired from, for example, a recording medium on which the program is recorded, a program distribution server via a network, or the like, installed in the ROM 103, read from the CPU 101, and executed.

The I / O device 104 has an operation input function and a display function (operation display function). The I / O device 104 is, for example, a touch panel. The touch panel enables the user of the information processing terminal 10 to input operations using a fingertip, a stylus, or the like. Even if the I / O device 104 integrates a display device having a display function and an operation input device having an operation input function such as a touch panel, the display device having the display function and the operation input having the operation input function are integrated. It may have a device separately. The display screen of the touch panel can be performed as the display screen of the display device, and the operation of the touch panel can be performed as the operation of the operation input device. The I / O device 104 may be realized by various forms such as a head mount type, a glasses type, and a wristwatch type display.

Communication I / F 105 is an I / F for communication. The communication I / F 105 executes short-range wireless communication such as a wireless LAN, a wired LAN, and infrared rays. Although the figure shows only the communication I / F 105 as the communication I / F, the information processing terminal 10 may have an I / F for each communication in a plurality of communication methods.

Next, the operation of the odor analyzer 1 will be described with reference to FIG. FIG. 6 is a flowchart showing the operation of the odor analyzer 1 in the embodiment.

In FIG. 6, the odor analyzer 1 starts measurement (step S11). The start of measurement is executed, for example, by the user performing a measurement start operation (for example, pressing the start button) on the odor analyzer 1.

After executing the process of step S11, the odor analyzer 1 acquires the measured value (step S12). The acquisition of the measured value may be performed according to the number of odor samples and a predetermined number of measurements. For example, when the number of odor samples is 2, the number of measurements is 3, and the number of sensors is 9 (n = 9), 2 × 3 × 9 = 42 measured values are acquired in the process of step S12.

After executing the process of step S12, the odor analyzer 1 extracts the principal component. The extraction of the principal components can be performed with a preset number of principal components (m).

After executing the process of step S12, the odor analyzer 1 projects n-dimensional data onto the m-dimensional principal component vector based on the m-dimensional principal component extracted by the extraction unit (step S14). The odor analyzer 1 may extract the principal component and project the measured value so that the difference in the projection data in the odor samples to be compared is maximized. In that case, the processes of steps S13 to S14 are repeatedly executed until the maximum value of the difference in the projected data is calculated.

After executing the process of step S14, the odor analyzer 1 draws the m-dimensional projection data projected in step S14 (step S15). The drawing of the projection data may be executed so that the difference in the projection data is maximized. As described above, the difference in projection data differs depending on which measured value is extracted as the main component. In the process of step S15, each projection data may be drawn according to the combination of the extracted principal components.

After executing the process of step S15, the odor analyzer 1 displays the projection data drawn in step S15 on the display device or the like of the odor analyzer 1 (step S16). In the process of step S16, the odor analyzer 1 may display the difference in the projection data calculated in the process of step S14. Further, the odor analyzer 1 may display each projection data according to the combination of the principal components drawn in the process of step S15. The display of the projected data may be performed, for example, on a terminal (not shown) connected via the network 9. After executing the process of step S16, the odor analyzer 1 ends the operation shown in the flowchart.

Note that the illustrated flowchart shows an example of the operation, and does not limit the operation.

Further, the program for realizing the function constituting the apparatus described in the present embodiment is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into the computer system and executed. Therefore, the above-mentioned various processes of the present embodiment may be performed. The "computer system" here may include hardware such as an OS and peripheral devices. Further, the "computer system" includes the homepage providing environment (or display environment) if the WWW system is used. The "computer-readable recording medium" includes a flexible disk, a magneto-optical disk, a ROM, a writable non-volatile memory such as a flash memory, a portable medium such as a CD-ROM, a hard disk built in a computer system, and the like. Refers to the storage device of.

Further, the "computer-readable recording medium" is a volatile memory inside a computer system that serves as a server or client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line (for example, DRAM (Dynamic)). It also includes those that hold the program for a certain period of time, such as Random Access Memory)). Further, the program may be transmitted from a computer system in which this program is stored in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the "transmission medium" for transmitting a program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. Further, the above program may be for realizing a part of the above-mentioned functions. Further, a so-called difference file (difference program) may be used, which realizes the above-mentioned function in combination with a program already recorded in the computer system.

Although the embodiment of the present invention has been described above with reference to the drawings, the specific configuration is not limited to this embodiment and includes various changes within the range not deviating from the gist of the present invention. Is done.

1 Smell analyzer 11 Measurement unit 12 Acquisition unit 13 Extraction unit 14 Projection unit 15 Drawing unit 16 Smell sensor registration unit 2 Server 21 Setting provider 9 Network 101 CPU
102 RAM
103 ROM
104 I / O equipment 105 Communication I / F
1000 Display screen 1001 Center point 1002 Measurement value group 1003 Center point 1004 Measurement value group 1005 Warning display 1006 Button 1007 Button 1008 Button 1009 Center point 1010 Measurement value group 2000 Display screen 2001 Measurement value group 2002 Measurement value group 2003 Button 2004 Button 2005 button

Claims

A measuring unit having n odor sensors (n is an integer of 1 or more),
The acquisition unit that acquires the measured values measured by the n odor sensors of the measurement unit, and the acquisition unit.
Extracts m (m is an integer of 1 or more and n ≧ m) main components from n-dimensional data whose one dimension is the measured value of the one odor sensor acquired by the acquisition unit. Extractor and
A projection unit that projects n-dimensional data in m-dimensionality based on the principal components extracted in the extraction unit, and a projection unit.
An odor analyzer comprising a drawing unit for drawing m-dimensional projection data projected in the projection unit.
In the odor sensor, the natural frequency changes due to the adsorption of odorous substances on the adsorption part, and the natural frequency changes.
The odor analyzer according to claim 1, wherein the measuring unit measures the natural frequency.
The odor analysis according to claim 1 or 2, wherein the projection unit projects the n-dimensional data into the m-dimensional space by projecting the n-dimensional data onto the space represented by the principal component vectors of the m principal components. Device.
Any one of claims 1 to 3, wherein the extraction unit extracts the principal component so that the difference between the projection data in the first odor sample and the projection data in the second odor sample is maximized. The odor analyzer according to item 1.
The odor analyzer according to any one of claims 1 to 4, wherein the drawing unit changes the drawing scale of the projection data for each dimension and draws the drawing.
The odor analyzer according to any one of claims 1 to 4, wherein the drawing unit draws the projection data of the reference odor sample and the projection data of the odor sample having a predetermined difference so as to be distinguishable. ..
A measurement step for measuring odors with n odor sensors (n is an integer of 1 or more),
The acquisition step of acquiring the measured values measured by n odor sensors in the measurement step, and the acquisition step.
An extraction step of extracting m (m is an integer of 1 or more) principal components from n-dimensional data whose one dimension is the measured value of one of the odor sensors acquired in the acquisition step, and the extraction step. A projection step that projects n-dimensional data to m-dimensional based on the principal components extracted in
An odor analysis method including a drawing step of drawing m-dimensional projection data projected in the projection step.
On the computer
Measurement processing that measures odors with n odor sensors (n is an integer of 1 or more),
The acquisition process for acquiring the measured values measured by n odor sensors in the measurement process, and the acquisition process.
An extraction process for extracting m (m is an integer of 1 or more) principal components from n-dimensional data whose one dimension is the measured value of one of the odor sensors acquired in the acquisition process.
Projection processing that projects n-dimensional data to m-dimensional based on the principal components extracted in the extraction process, and
An odor analysis program for causing a computer to perform a drawing process for drawing m-dimensional projection data projected in the projection process.