TW202122795A - Texture evaluation method, texture standard model - Google Patents

Abstract

A texture evaluation method of the present invention includes the following steps: a step of making a template T composed of pattern data that is used as a benchmark for texture evaluation; a step of performing sensory evaluation on a model food F_M for texture evaluation to obtain a sensory evaluation value λ; a step of obtaining measurement data A corresponding to the pattern data of the model food F_M, comparing the measurement data A with each of the pattern data of the template T to calculate similarity S_A; a step of producing a texture standard model M corresponding to types of texture from the sensory evaluation value λ and the similarity S_A; and a step of using the texture standard model M to evaluate the texture of evaluation food F_E.

Description

Taste evaluation method, standard model of taste

The present invention relates to a mouthfeel evaluation method that can perform high-precision mouthfeel evaluation on various foods and a standard model of the mouthfeel for the mouthfeel evaluation.

The deliciousness of food is mainly composed of three points: flavor, aroma and taste. Among these, the taste is composed of extremely slender and complex information. For example, it is difficult to quantify the texture determined by complex factors such as stickiness or melting in the mouth. Therefore, such a taste is sensoryly evaluated by a plurality of evaluators.

In recent years, there has been a method of evaluating the taste using a device having the same structure as the human oral cavity. For example, the tactile sensor of Patent Document 1 below has a flexible layer and a substrate layer laminated, and a magnet is enclosed in the flexible layer. In addition, on both sides of the substrate layer, the opposite side of the flexible layer placement surface that is in contact with the flexible layer is set as the element placement surface, and magnetoresistive elements (GMR) and inductors are placed, and circuits are formed on the element placement surface of the substrate. .

In addition, a tooth-shaped hard member may be provided through the soft layer. When the tooth-shaped member is provided, the tactile sensation applied to the tooth can be detected. In this way, the taste can be evaluated by the device using the tactile sensor (paragraphs 0010, 0012, 0019, 0025, Figure 1).

[Prior Technical Literature]

[Patent Literature]

Patent Document 1: Japanese Patent No. 5187856.

The tactile sensor of Patent Document 1 is a device that compares the texture of the texture based on the load and vibration of the device, and can evaluate texture such as softness or hardness. However, with just such a device, there is a problem that evaluation close to human perception cannot be performed, and the accuracy of determination is insufficient.

The present invention was developed in view of this situation, and aims to provide a texture evaluation method that can perform high-precision texture evaluation of various foods for evaluation.

In order to achieve the above object, the taste evaluation method of the present invention is to evaluate the taste of food for evaluation. The taste evaluation method includes the following steps: a step of making a template. The template is a plurality of virtual data as the index of the taste evaluation, and is It is composed of pattern data; the process of performing sensory evaluation on the model food used for the aforementioned taste evaluation and obtaining the sensory evaluation value; obtaining the measurement data (A) of the aforementioned model food corresponding to the aforementioned pattern data, Steps of comparing the aforementioned measurement data (A) with the aforementioned template data, and calculating the similarity (S _A ); from the aforementioned sensory evaluation value and the aforementioned similarity (S _A ), a standard model of taste corresponding to the taste is made Step; and use the aforementioned standard model of mouthfeel to carry out the step of evaluating the mouthfeel of the food for evaluation.

In the mouthfeel evaluation method of the present invention, a template is first produced, and the template is composed of a plurality of type data as an index for the mouthfeel evaluation. Then compare the measurement data (A) of the food used in the model with the type data, and calculate the similarity (S _A ). In this way, information about which type data is similar to the food used in the model can be obtained.

In addition, the model used for the evaluation of functional foods, the taste evaluation standard model from the functional value and the degree of similarity (S _A) production. Standard model taste were associated with a functional similarity evaluation value (S _A), so the use of the standard model to evaluate the taste of food with the texture evaluation. Thereby, this method can perform high-precision taste evaluation of various foods for evaluation.

The taste evaluation method of the present invention preferably includes the following steps: obtaining the measurement data (B) of the aforementioned food for evaluation corresponding to the aforementioned type data, comparing the aforementioned measurement data (B) with each of the aforementioned type data of the aforementioned template, and calculating the degree of similarity (S _B) of the step; and taste by the standard model evaluated the degree of similarity (S _B), and the taste was assessed in the step of food.

According to this structure, the measurement data (B) obtained for the food for evaluation is compared with the type data, and the similarity (S _B ) is calculated. In this way, information about which type of data the food for evaluation is similar to can be obtained. Taste evaluation standard model is then used with the degree of similarity (S _B) for evaluation of the food. In this way, this method can evaluate the taste of the food for evaluation.

Furthermore, in the taste evaluation method of the present invention, it is preferable that the aforementioned type data is obtained by changing the intensity of a predetermined parameter in an arbitrary manner over time.

The type data of the template is, for example, virtual data obtained by changing the intensity of the parameter in an arbitrary manner over time by a program. By preparing a plurality of this type of data, when comparing the measurement data with each type of data, the type data with a higher degree of similarity can be extracted. Thereby, this method can be used for the taste evaluation of foods for evaluation.

In addition, in the taste evaluation method of the present invention, it is preferable that the step of calculating the similarity degree is by calculating the distance between the type data obtained over time and the corresponding elements of the measurement data (A) or the measurement data (B). Accumulate values.

The waveform of the type data is made by changing the parameters over time, so the data will be graphed. For example, the comparison between the measurement data (B) of the food for evaluation and the type data can be obtained by calculating the cumulative value of the distance between the corresponding elements of the two. The smaller the distance and the more overlap between the two waveforms, it can be evaluated as a similar degree of similarity. high.

Furthermore, in the taste evaluation method of the present invention, it is preferable that the measurement data (A) includes load change data that is a load change obtained when the food for the model is pressed. In this case, it is more preferable that the measurement data (B) includes load change data that is a load change obtained when the food for evaluation is pressed.

The measurement data (A) and (B) include load change data when the model food and the evaluation food are pressed, respectively. For example, when the food for evaluation is pressed twice, the load changes in the first time and the second time according to the type of food. The load change is the unique data of the food, so this method can improve the accuracy of taste evaluation.

Furthermore, in the taste evaluation method of the present invention, it is preferable that the measurement data (A) contains vibration change data that is a vibration change obtained when the food for the model is pressed. In this case, it is more preferable that the aforementioned measurement data (B) include vibration change data belonging to vibration changes obtained when the aforementioned food for evaluation is pressed.

The measurement data (A) and (B) may include vibration change data when the model food and the evaluation food are pressed, respectively. Vibration change is also the unique data of the food for evaluation, so this method helps to improve the accuracy of taste evaluation.

Further, sensory evaluation method of the present invention is preferably in taste by standard model-based explanatory variables as the degree of similarity (S _A) and the target variable is the sensory evaluation value ([lambda]) obtained from the regression analysis and evaluation of prediction formulas.

Taste by using the standard model-based similarity (S _A) and the resulting regression analysis and evaluation of the evaluation value predicted functional formula. Accordingly, the present method may be based programs or the like by the degree of similarity (S _A) and a functional evaluation value inputted to the predictive evaluation, whereby the production can be more easily updated and food model.

In addition, the taste evaluation method of the present invention is preferably that the aforementioned regression analysis can be applied to various linear models, especially the generalized linear mixed model (GLMM) is used.

Regression analysis is a linear model including a generalized linear model (GLM), a generalized linear mixed model (GLMM), etc. By this method, the taste of the food for evaluation can be easily and accurately evaluated by this method.

The mouthfeel standard model of the present invention is used to evaluate the mouthfeel of foods for evaluation. The mouthfeel standard model has: a template, which is composed of a plurality of type data as the aforementioned mouthfeel evaluation index; and the sensory evaluation value is used for the aforementioned the sensory evaluation model for evaluation obtained by using a functional food; degree of similarity (S _A), comparing the model-based food with the data patterns corresponding to the measurement data of (A) with the resulting patterns for each of the template data; and evaluation prediction means, the line corresponding to the functional value and the evaluation type similarity mouthfeel (S _A) obtained from the evaluation prediction.

The mouthfeel standard model of the present invention is provided with a template composed of a plurality of morphological data serving as a mouthfeel evaluation index, compares the measurement data (A) and morphological data of the food used for the model, and calculates the similarity (S _A ). In this way, information about which type data is similar to the food used in the model can be obtained.

Further, the evaluation means based on the prediction model used for food sensory evaluation, evaluated predicted from the functional similarity evaluation value (S _A). Accordingly, the standard model-based texture evaluation value with the functionalized degree of similarity (S _A) in association, and can be evaluated by the evaluation of the taste of food.

According to the present invention, high-precision taste evaluation can be performed on various foods.

S1~S5,1~7: sample

10: Taste evaluation device

11: Contact

13: permanent magnet

15: Middle layer

17: base part

19: Circuit board

21: Magnetoresistive element

23: Inductor components

25: Connector

Fig. 1 is a diagram illustrating a mouthfeel evaluation device used in the mouthfeel evaluation method of the present invention.

Fig. 2 is a view of the circuit board of the taste evaluation device viewed from the bottom side.

Figure 3A shows the measurement results of the load and vibration of the sample S1.

Figure 3B shows the measurement results of the load and vibration of the sample S2.

Figure 3C shows the measurement results of the load and vibration of the sample S3.

Figure 3D shows the measurement results of the load and vibration of the sample S4.

Figure 3E shows the measurement results of the load and vibration of the sample S5.

Fig. 4 is a diagram illustrating the outline of the taste evaluation method of the present invention.

Fig. 5A is a diagram showing a template (type data for load re-use).

Fig. 5B is a diagram showing a template (type data for vibration).

Fig. 6A is a graph showing the average scores of the sensory evaluations of 7 samples.

Figure 6B is a histogram of the average scores of the sensory evaluations of 7 samples.

Figure 7A is the measurement result (area) of a texture analyzer.

Figure 7B is the measurement result (the number of crests) by the texture analyzer.

Figure 8A is the measurement result of the texture analyzer (average decrease).

Figure 8B is the measurement result (peak value) of the texture analyzer.

Fig. 9A shows the load of 7 samples measured by the mouthfeel evaluation device.

Fig. 9B shows the vibration of 7 samples measured by the mouthfeel evaluation device.

Figure 9C is the cumulative sum of vibrations of 7 samples measured by the taste evaluation device.

Hereinafter, one embodiment of the taste evaluation method of the present invention will be described with reference to the drawings.

Figures 1 and 2 show the texture evaluation device 10 used in the texture evaluation method of the present invention. The taste evaluation device 10 shown in FIG. 1 has a structure equivalent to human teeth, gums, and the like. The food texture evaluation device 10 uses various sensors to measure the bite of the food for evaluation in order to evaluate the texture of the food for evaluation to be evaluated.

The mouthfeel evaluation device 10 is mainly composed of the contact portion 11 corresponding to the tooth, the intermediate layer portion 15 corresponding to the gingiva (periodontal ligament), and the base portion 17 corresponding to the alveolar bone. In addition, the base portion 17 is placed on the circuit board 19.

A hard member is used for the contact part 11, and the contact part 11 is formed in a wedge shape, for example. If the contact portion 11 is formed in a wedge shape, it is easier to detect vibration, which is preferable. In addition, a permanent magnet 13 is embedded in the contact portion 11. Next, the magnetoresistive element (GMR: Giant Magneto Resistive effect, giant magneto resistance effect) 21 and the inductor element 23, which are part of a mouth-feel sensor, are provided on the circuit board 19 to obtain the evaluation result when the food for evaluation is pressed against the contact portion 11. Magnetic changes. In addition, the upper surface of the contact portion 11 may be processed into a flat shape to make it easier to press the food for evaluation.

The permanent magnet 13 can be a neodymium magnet (for example, a diameter of 15 mm, a thickness of 1.5 mm, and a surface magnetic flux density of 115 mT). In addition, the magnetoresistive element 21 can detect the position change (displacement amount) of the permanent magnet 13, and the aforementioned change when the food for evaluation is pressed against the contact portion 11 is recorded as a "load".

If the position of the permanent magnet 13 is displaced, an induced electromotive force is generated in response to the change in the magnetic flux density. The inductor element 23 detects this induced electromotive force. The magnitude of the induced electromotive force is more dependent on its speed than the amount of change in magnetic flux density. When the food for evaluation is pressed against the contact portion 11, the aforementioned change detected by the inductor element 23 is recorded as "vibration".

In addition, the intermediate layer 15 uses an elastic body (natural rubber, synthetic rubber, etc.) or a spring, and the base 17 uses a hard member, so that human teeth can be reproduced. Therefore, the values of "load" and "vibration" will vary depending on the type of food used for evaluation.

In addition, FIG. 2 shows a view of the circuit board 19 viewed from the lower side.

As shown in the figure, eight magnetoresistive elements 21 are arranged with the circular center part vacant on the lower surface side of the base part 17 indicated by the circular dotted line. The permanent magnet 13 (not shown) of the contact portion 11 is located at the center of the circle. Therefore, when the food for evaluation is pressed against the contact portion 11, the position change (load) of the permanent magnet 13 can be detected by any magnetoresistive element 21.

In addition, one inductor element 23 is arranged at the center of the circular shape. Thereby, the change (vibration) of the magnetic flux density of the permanent magnet 13 located almost directly above the inductor element 23 can be detected. In addition, the number of magnetoresistive elements 21 and inductor elements 23 can be changed as appropriate. In addition, the signals detected by the magnetoresistive element 21 and the inductor element 23 are relayed by the connector 25 on the circuit board 19 and transmitted to a control unit not shown in the figure.

Next, a measurement example of the load and vibration of the mouthfeel sensor will be described with reference to FIGS. 3A to 3E.

The taste evaluator (hereinafter referred to as the "evaluator") used fried shrimp (5 types of samples S1 to S5) with a high degree of imparting mechanical properties of the taste as the food for evaluation, and measured each sample 8 times. Samples S1 to S5 are fried shrimps that are conditioned under different conditions and coated with different flours.

In FIGS. 3A to 3E, the left vertical axis is force (N), the right vertical axis is voltage (V), and the horizontal axis is time (s). In the samples S1 to S5 corresponding to Figs. 3A to 3E, the load waveform (thick line: load waveform data) and the vibration waveform (solid line: vibration change data) are different, for example, the sample S1 in Fig. 3A and the sample in Fig. 3C S3. In the measurement data of sample S5 in Fig. 3E, a load peak is obviously generated at a predetermined time.

The details will be described later. The evaluator compares it with a template composed of plural time series data prepared from the above-mentioned measurement data in advance, and calculates the similarity. In addition, the evaluator compares the taste standard model made from the measurement data and sensory evaluation value of the food for the model to evaluate the taste (crispy, crispy, crunchy, etc.) of each sample.

Next, the food texture evaluation method of the present invention will be described in detail with reference to FIG. 4. In addition, the content of the template and its production are supplemented with reference to FIGS. 5A and 5B.

First, the evaluator uses model food F _M (usually 6 to 7 types) to create a standard model M of taste. The food F _{M for the} model can be selected from a plurality of foods having a taste similar to that of the food for evaluation.

First, the sensory evaluation value λ of the _{model food F M is} obtained by the sensory evaluation performed by the sensory evaluator. For example, when sensory evaluations were conducted on chicken, with the model-based food F _M using the same chicken. Next, the sensory evaluator chewed the food F _{M for the} model and evaluated which item is close to "crispy", "slightly crispy", "slightly crispy", "crispy", "slightly crispy" and "hard and crispy" By. The sensory reviewer assigns "1" to the closest texture, "2" to the second closest texture, and "3" to the third closest texture, which can be digitized later.

Next, the evaluator measures the food F _{M for the} model using the taste sensor. The above-mentioned mouthfeel evaluation device 10 was used for this measurement. Specifically, the model-based evaluator food F _M with two 11 pressed against the contact portion of the sensory evaluation apparatus 10, and the load information acquired measured vibrations A. In addition, the measurement data A is the waveform data (load change data and vibration change data) of the load and vibration as shown in FIGS. 3A to 3E.

Then, the evaluator compares the measurement data A with the template T prepared in advance. The template T is composed of different types of waveforms as an index of taste evaluation. The type data is virtual data obtained by causing the intensity of a predetermined parameter to generate random numbers and making it change with time in an arbitrary (random) manner.

The template T includes load (refer to FIG. 5A) and vibration (refer to FIG. 5B), and each waveform is one type of data. Comparison with model-based evaluator "load" and FIG measurement data of the food F _M A 5A of the respective data patterns, the similarity degree for the sake of a waveform shape. Further, the evaluator-based model data were compared with each of the patterns for "vibrating" and FIG measuring food F _M of the data A and 5B, in the same manner for the sake of similar waveform shape. _{The similarity S A} obtained from this operation is used to make the standard model M of taste.

Here, when the taste sensor in a measurement model with a food F _M, F _M food model with two pressed against the contact portion 11, it is shown in Figure 5A with a load of data patterns to be formed on the waveforms of time. Specifically, the peaks P1 and P2 of the profile data are the parts corresponding to the first press, and the pressure (force [N]) is changed respectively. In addition, the crests P3 and P4 of the profile data correspond to the second pressing, and the pressures are changed in the same way as the crests P1 and P2. In this way, the evaluator creates 50 types of data with different peaks p1 to p4.

Fig. 5B shows type data for vibration. In taste inductor (inductor element 23) detected by the vibration even if the same kind of model generation timing will be different with the vibration food F _M. Therefore, the evaluator uses the cumulative sum of the number of positive induced electromotive force of vibration (the cumulative sum of vibration) as the vertical axis to create 50 types of data that change over time.

The dynamic time warping method (DTW: Dynamic Time Warping) can be used to quantitatively evaluate the similarity between the measurement data A of the _{food F M for the model and the template T.} In this evaluation, the evaluator measures the distance between the two waveforms (DTW distance), but the more similar the two data, the smaller the DTW distance (higher similarity) and the closer the waveform shape. Here, each waveform is composed of time-series measurement values (loads, etc.) sampled at regular intervals, and each point (requirement of the present invention) is continuously represented. In order to measure the DTW distance, after exhaustively comparing the distance between the elements of each waveform, find the shortest path between the points. Then take the cumulative value of the cumulative required distance as the DTW distance.

In addition, in the similarity between the measurement data A and the template T, the DTW distance can be replaced by, for example, measuring Euclidean distance, Fourier transform distance, autoregressive coefficient distance, or EDR (Edit Distance on Real sequences). Distance and evaluate.

Thereafter, the functional evaluation value λ S _A and the degree of similarity is associated, for example, decided to take the taste of functional evaluators evaluated as "crisp" type of information. Accumulation of this information can complete the taste standard model M that can be used to evaluate various tastes.

In fact, if the explanatory variable is set as the similarity S _A (comparison results s _a1 , s _a2 , ... s _{a50 of} 50 types of data), and the target variable is set as the sensory evaluation value λ, the following evaluation prediction formula can be used And the taste standard model M is obtained.

log{λ/(1-λ)}=β ₀ +β ₁ s _a1 +β ₂ s _a2 +…+β ₅₀ s _a50 +r _i …(Equation 1)

Here, β ₀ , β ₁ , β ₂ ,..., β ₅₀ are coefficients, and r _i is a parameter representing individual differences.

Specifically, based evaluator program or the like by the degree of similarity S _A functional evaluation and the evaluation value is substituted into λ prediction equation (Equation 1), regression analysis to find coefficients _{β 0, β 1, β 2} , ..., β 50 ( And _{the standard deviation of r i} ), and complete the taste standard model M. In addition, the regression analysis is not limited as long as the value of each parameter can be obtained. Various linear models such as generalized linear mixed models (GLMM), generalized linear models (GLM), nonlinear models, and Bayesian hierarchical models (Hierarchical Bayesian Model) are applicable.

In addition, this time the system evaluated six textures of "crispy", "slightly crispy", "slightly crispy", "crispy", "slightly crispy" and "hard and crispy", so the sensory evaluation value λ was 6. That is, the evaluation prediction formula (Equation 1) is also six types, and the evaluator respectively substitutes the similarity S _{A in them} , and creates a taste standard model M corresponding to the taste.

Next, the evaluator for the evaluation was the food F _E, evaluation of measurement data acquired by the load and the vibration of the food F _E using sensory evaluation apparatus 10 B. Measure the load and vibration system of the measurement performed in the same model when using food F _M conditions.

Next, the evaluator compares the measurement data B with the template T. Specifically, the evaluator lines were evaluated by measuring the food material F _E B is the "load", "vibration" and FIGS. 5A, 5B of the information patterns (50 types), and calculates the similarity degree waveform shape. Through this operation, the similarity degree S _B (comparison results s _b1 , s _b2 , ... s _{b50 of} 50 types of data) is obtained. Also, evaluation data of similarity with the template T B F _E measurement of food also can be used such as dynamic time warping method (DTW).

Finally, based evaluator compared with the evaluation of the degree of similarity food F _E S _B M and texture prepared in advance of the standard model. Specifically, the evaluator performs the task of obtaining the model evaluation value λ′ using the following evaluation prediction formula.

log{λ'/(1-λ')}=β ₀ +β ₁ s _b1 +β ₂ s _b2 +...+β ₅₀ s _b50 +r _i …(Equation 2)

_{This time, the evaluator substitutes the similarity SB} (s _b1 , s _b2 ,...s _b50 ) of the food for evaluation into the evaluation prediction formula (Equation 2) to obtain the model evaluation value λ'(coefficient β ₀ , β ₁ , β ₂ , …, β ₅₀ , r _i are the values obtained when the standard model M of texture was created). Model evaluation value λ 'of the concrete example described later, but can be evaluated for the taste evaluation of the food F _E by the above series of operations.

Next, the results of sensory evaluation using 7 types of chicken nugget samples will be described with reference to FIGS. 6A and 6B.

First, FIG. 6A shows the average score (sensory evaluation value) of the sensory evaluation of the samples 1 to 7 evaluated by the sensory evaluator. There are 6 types of taste: "crispy", "slightly crispy", "slightly crispy", "crispy", "slightly hard and crispy" and "hard and crispy". For samples 1 to 7, a panelist (5 people) will test them separately 3 sensory evaluations (10 points full marks). For example, sample 2 has the highest value of "slightly crispy", and it can be said to be a sample that can feel the taste for the sensory reviewer.

The rough definition of each taste is explained here. "Crunchy" refers to the taste that "can be crushed and coated with a lighter force, and the touch will continue when you continue to chew." In addition, "slightly crispy" means "the touch of the powder can be crushed with lighter force when chewing at first", and "slightly crispy" means "if you bite with molars, you can feel the touch of disintegrating into small pieces several times." ". In addition, "crispy" means "feel the disintegration of a harder and firm object, and the tactile sensation will continue when you continue to chew", and "slightly hard and crisp" means "the front teeth can be crushed with lighter force to make the incisors harder." "The touch of things", "hard and brittle" means "the touch of crushed ice products".

In addition, "Proba" in the figure is the test value of whether there is a difference between each item in the one-way analysis of variance (ANOVA), and "Significant" is the multiple comparison test (Tukey-Test, Dukai test) that produces the most significant difference. The significance level of the sample.

Next, FIG. 6B is a histogram showing the average scores of the sensory evaluations of the above-mentioned samples 1 to 7 for each taste. In Sample 1, the average score of "crispy" or "slightly crispy" was higher, but the overall score was lower, and it was judged as "no feature". A lower absolute value of the score indicates a weaker taste. For example, a lower score for "crispy" means that the examiner judges that the taste is weaker.

In addition, from the average score, sample 2 judged to be "slightly crispy" with stronger taste, sample 3 judged to be "crispy" with stronger taste, and sample 4 judged to be "hard and crispy" with stronger taste. Similarly, sample 5 was judged to be "slightly crispy" with a strong texture, sample 6 was judged to be "crunchy" with a strong texture, and sample 7 was judged to be "slightly hard and crisp" with a strong texture. In sample 4, the mouthfeel of "slightly hard and brittle" and "hard and brittle" were scored in the same score, but the mouthfeel of "slightly hard and brittle" of sample 7 was clearer, so sample 4 was judged to be "hard and brittle".

In addition, the difference sign is attached to the samples that have significant differences (P<0.05) in the multiple comparison test. For example, there are significant differences between "a", "b", and "c", but there are no significant differences between "a" and "ab" or between "ab" and "b".

Next, with reference to FIGS. 7A to 8B, the results of the evaluation of the same samples 1 to 7 using a commercially available texture analyzer (TA.XTplus: manufactured by Stable Micro Systems) will be described. The texture analyzer is a device that measures and quantifies data related to the bite, bite, and taste of food. The chicken nugget measurement performed by the texture analyzer was repeated 4 times for samples 1 to 7 respectively.

First, Figure 7A shows the measurement results of the "area" of each sample. "Area" refers to the area S of the product of measurement time (time [sec]) and stress (force [g]), the vertical axis of Fig. 7A is S[g. sec]. The value of "area" is the total value of the rebound force when the texture analyzer presses the sample, so it is generally an indicator of the elasticity or hardness of the sample. In addition, the line segment at the top of the histogram represents the standard deviation of the measured value in the sample.

In the measurement of "area", the taste of "Sample 3 (crispy)", "Sample 6 (crispy)" and "Sample 7 (slightly hard and crisp)" was about 1,000 [g. sec] is a high value, but other tastes are also about 800 [g. The value of sec], no significant difference has been confirmed. Therefore, the taste evaluation of chicken nuggets does not have great reference value.

Next, FIG. 7B shows the measurement results of the "number of crests" of each sample. When the sample is pressed with a texture analyzer, stress separation (wave crest) phenomenon will occur when the sample is crushed. The person who counts this number of times is called "the number of crests". In the measurement of "peak number", the taste of "Sample 6 (crisp)" and "Sample 7 (slightly hard and brittle)" are high values. Generally, the higher the value of "Crest Number", the stronger the "crispy" taste, so it is the indicator.

Next, FIG. 8A shows the measurement result of the "drop off" of each sample. "Drop" refers to how much stress (force [g]) is lost at the peak of the wave. In addition, the "average decrease" is the average value of all the decreases in one measurement. In the measurement of "average drop", the taste of "Sample 7 (slightly hard and crisp)" was a high value. Generally speaking, the "average decrease" is an indicator of brittleness, and the higher the value, the greater force is required for crushing.

Next, FIG. 8B shows the measurement results of the "wave peak value" of each sample. The "peak value" refers to the average value of the maximum stress of each wave peak in one measurement. Regarding the "peak value", although no significant difference has been confirmed for each taste, it is generally an index indicating the hardness of the sample.

In the results of the sensory evaluation, a clear difference was confirmed between "Sample 4 (hard and brittle)" and "Sample 6 (crisp)" (see Fig. 6A and Fig. 6B). However, in the evaluation of texture analyzer, the evaluation The two samples were tested with a significant level of 5% for t test (one-sided, no correspondence), and the results were "area" (significant probability p=0.058%), "peak number" (significant probability p=0.501%), and "average decline" (Significant probability p=0.444%), "Crest" (significant probability p=0.231%) no significant difference has been confirmed. As mentioned above, there is no clear difference between the measured values of the two samples, and as a result, the difference in taste cannot be fully evaluated.

Next, the results of measuring the same samples 1 to 7 by the taste evaluation device 10 of the embodiment of the present invention will be described with reference to FIGS. 9A to 9C.

9A to 9C show the results of measuring the "load", "vibration", and "vibration cumulative sum" of the samples 1 to 7 by the taste evaluation device 10, respectively. Specifically, it is the result of load measurement data (load waveform data) and vibration measurement data (vibration waveform data) obtained by the evaluator pressing each sample onto the contact portion 11 of the taste evaluation device 10 twice.

According to the "load" data in Figure 9A, except for samples 5 and 7, almost similar waveforms are obtained. However, according to the "vibration cumulative sum" data in Figure 9C, although the waveforms of samples 2, 3, and 5 are similar, the waveforms of samples 1, 4, 6, and 7 can be different from those of samples 2, 3, and 5. Characteristic waveform. The result of "vibration accumulation" is that sample 7 is the largest and sample 1 is the smallest.

After that, the evaluator compares the load waveform data and vibration waveform data (the measurement data A in Figure 4 and the various type data (50 types) of the template T) to obtain the similarity (the similarity S _{A in} Figure 4). In other words, the evaluator uses the taste evaluation device 10 to measure the "load" and "vibration cumulative sum" 10 times for samples 1 to 7, respectively, and compare them with each type data (50 types) of the template T to obtain the similarity S _A. and determining the degree of similarity of 10 parts of S _a average.

Next, the evaluator from the resulting degree of similarity and the previous average value S _A of average scores of the sensory evaluation (see FIG. 6A) using generalized linear mixed model (a GLMM), to prepare 6 kinds of evaluation prediction equation (Equation 1), and determining the parameters The values of β ₀ , β ₁ , β ₂ ,..., β ₅₀ , and r _i.

Next, the evaluator used another sample 3'of chicken nuggets prepared and conditioned through the same steps as the above sample 3, and obtained the measurement data of load and vibration by the taste evaluation device 10 under the same conditions as when measuring the sample 3 ( Figure 4 Measurement data B). The measurement of sample 3'was repeated 10 times.

After that, the evaluator compares the measurement data B and each type data (50 types) of the template T, and obtains the similarity (the similarity S _{B in} FIG. 4 ).

In addition, the evaluator uses the obtained similarity _{SB and the parameters β 0} , β ₁ , β ₂ ,..., β ₅₀ , r _{i obtained} by the evaluation prediction formula (Equation 1) based on the previous samples 1 to 7 Value, and find the model evaluation value λ'of sample 3'. The measurement of sample 3'is repeated 10 times, and the evaluator finds the average value of the model evaluation value λ', and uses it as the evaluation result by the taste evaluation model.

Next, referring to Table 1, the evaluation results and sensory evaluation results of the taste evaluation model related to the embodiment of the present invention will be described.

Table 1 shows the evaluation values (sensory evaluation value λ) when the sample 3 was subjected to sensory evaluation, and the evaluation values (model evaluation value λ') of the sample 3'evaluated using the taste evaluation model M of this embodiment.

[Table 1]

The evaluation value of the taste evaluation model M shows the same tendency as the sensory evaluation value λ, and the correlation coefficient between the sensory evaluation value λ and the model evaluation value λ'is 0.97. Generally speaking, the relationship If the absolute value of the number exceeds 0.4, it is correlated, and if it exceeds 0.7, it is strongly correlated. Therefore, sample 3 has a very strong correlation with sample 3'.

Also, similarly to the sample 3', the model evaluation value λ'was also obtained for the samples 4'and 6'corresponding to the samples 4 and 6. The correlation coefficient between the sensory evaluation value λ of the sample 4 and the model evaluation value λ'is 0.86. Similarly, the correlation coefficient between the sensory evaluation value λ of sample 6 and the model evaluation value λ'is 0.87. Therefore, in both samples 5 and 6, the model evaluation value λ'can be integrated with the sensory evaluation value λ.

As mentioned above, in the measurement of the texture analyzer, sample 4 and sample 6 cannot be clearly distinguished, but in the evaluation using the standard model M of taste, the evaluation values of the two samples are significantly different, which is a result that can be evaluated as a distinguishable difference in taste.

As described above, and the results of such combination use of "load" and "vibration" food F _E Evaluation of the texture 10 in the measurement evaluation apparatus of the present invention, whereby the texture can be evaluated with high accuracy.

In addition, foods (fried chicken, croquette, spring rolls, croquette, etc.) whose texture is closer to the above-mentioned chicken nuggets can also be evaluated with the template T produced this time. In addition, if other templates are used, meat buns, fruits, etc. whose texture is completely different from chicken nuggets can be evaluated, and the texture can be evaluated for a wider range of foods for evaluation.

This time prepare 50 types of template T type data, but they are dependent on the evaluated food, so it is not limited to 50 types. Typical of the past, such as making "a little crisp" taste of the waveform data, waveform data and comparing the measured data and evaluation of food F _E, and evaluation of taste. However, by preparing a plurality of pattern data corresponding to the taste according to the taste evaluation device 10, the degree of "slightly crispy" can be obtained. You can also perform more detailed evaluations such as "slightly crispy" closer to "slightly crispy", or "slightly crispy" closer to "crispy".

Further, in the above-described embodiments patterns, so that the model-based evaluator with a food evaluation and food F _M F _E to the press-contacting portion 10 of the sensory evaluation apparatus 11 twice and measured data, but the number may be pressed only once, or may be three the above.

In addition, in the above-mentioned embodiment, the taste evaluation device 10 that obtains measurement data by magnetic change is used, but a device that obtains measurement data by vibration or sound change can be used, for example, a texture testing machine or rheometer ( rheometer), creep meter, etc.

Claims (12)

A taste evaluation method is to evaluate the taste of food for evaluation. The taste evaluation method includes the following steps: In the step of making a template, the aforementioned template is a plurality of virtual data used as an index for the aforementioned taste evaluation, and is composed of morphological data; Perform sensory evaluation on the model food used for the aforementioned taste evaluation and obtain the sensory evaluation value (λ); Obtain the measurement data (A) of the aforementioned model food corresponding to the aforementioned type data, compare the aforementioned measurement data (A) with the aforementioned type data of the aforementioned template, and calculate the step of _{similarity (S A );} Sensory evaluation by the aforementioned value ([lambda]) and the degree of similarity (S _A) in response to the step of making a standard model of the kind of taste texture; and Use the aforementioned standard model of mouthfeel to carry out the procedure of evaluating the mouthfeel of the aforementioned food for evaluation. The taste evaluation method as described in claim 1, further including the following steps: Obtain the measurement data (B) of the aforementioned food for evaluation corresponding to the aforementioned type data, compare the aforementioned measurement data (B) with the aforementioned type data of the aforementioned template, and calculate the step of _{similarity (S B ); and} With the standard model evaluated the taste degree of similarity (S _B), the evaluation step is carried out by evaluation of the taste of the food. The taste evaluation method according to claim 1, wherein the aforementioned type data is obtained by changing the intensity of a predetermined parameter in an arbitrary manner over time. The taste evaluation method according to claim 3, wherein the step of calculating the similarity is by calculating the relationship between the type data obtained over time and the corresponding elements of the measurement data (A) or the measurement data (B) Based on the cumulative value of the distance. The taste evaluation method according to claim 1, wherein the measurement data (A) includes load change data that is a load change obtained when the food for the model is pressed. The taste evaluation method according to claim 5, wherein the measurement data (B) includes load change data that is a load change obtained when the food for evaluation is pressed. The taste evaluation method according to claim 1, wherein the measurement data (A) includes vibration change data that is a vibration change obtained when the food for the model is pressed. The taste evaluation method according to claim 7, wherein the measurement data (B) includes vibration change data that is a vibration change obtained when the food for evaluation is pressed. The requested item obtained from sensory evaluation method of claim 1, wherein the taste will be explained by standard model-based variable is set to the degree of similarity (S _A) and the target variable is set to the sensory evaluation value ([lambda]) of the regression analysis Evaluate the predictive formula. The taste evaluation method according to claim 9, wherein the aforementioned regression analysis can be applied to various linear models. The taste evaluation method according to claim 10, wherein the aforementioned regression analysis uses a generalized linear mixed model (GLMM). A standard model of taste, which is used to evaluate the taste of food for evaluation. The standard model of taste has: The template is composed of multiple types of data as the aforementioned taste evaluation index; The sensory evaluation value (λ) is obtained from the sensory evaluation of the model food used for the aforementioned taste evaluation; The similarity (S _A ) is obtained by comparing the measurement data (A) of the aforementioned model food corresponding to the aforementioned model data with the aforementioned model data of the aforementioned template; and Evaluation prediction means, the line corresponding to the sensory evaluation value ([lambda]), and the like of the type mouthfeel (S _A) obtained from.

Images