Method of measuring taste using two phase radial basis function neural networks, a taste sensor, and a taste measuring apparatus

• US 7,899,765 B2
• Filed: 05/22/2006
• Issued: 03/01/2011
• Est. Priority Date: 05/23/2005
• Status: Active Grant

• Current Assignee: KEIO University
• Original Assignee: KEIO University
• Inventors: Ikeda, Atsushi, Araki, Yoshio, Citterio, Daniel, Ishihara, Saeko, Maruyama, Kenichi, Suzuki, Koji, Hagiwara, Masafumi
• Primary Examiner(s): Vincent; David R
• Application Number:
  US11/920,924
• Publication Number:
  US 20090234196A1
• Time in Patent Office: 1,744 Days
• Field of Search: 706 10- 20
• US Class Current: 706/16
• CPC Class Codes: G01N 33/02 Food
• CPC Class Codes: G01N 33/14 Beverages

Abstract

A method for measuring tastes, which can better simulate the human gustation than known methods, as well as a taste sensor, computer program and an apparatus for measuring tastes, is disclosed. In this method, data processing is carried out by a two-phase radial basis function neural network. That is, by sensors, each of which sensors can quantify at least one component representing, individually or cooperatively, the taste of saltiness, sourness, sweetness, umami or bitterness, to obtain a response value from each sensor, and each of the obtained response values is input to a first phase radial basis function neural network to calculate the concentration of each component from each response value. Then, the concentration of each component is fed into a second phase radial basis function neural network, which correlates the concentration of each component with the intensities of saltiness, sourness, sweetness, umami and bitterness sensed by humans, to calculate the intensities of saltiness, sourness, sweetness, umami and bitterness sensed by humans.

Claims

1. A computer-implemented method for measuring tastes, said method comprising the steps of:

subjecting a test sample to measurements by sensors comprising electrodes, each sensor can quantify at least one component representing, individually or cooperatively, the taste of saltiness, sourness, sweetness, umami or bitterness, to obtain a response value from each sensor;

inputting each of the obtained response values into a first phase radial basis function neural network, which correlates each response value with a concentration of each of the components and calculates the concentration of each component from each response value;

inputting the calculated concentration of each component into a second phase radial basis function neural network, which correlates the calculated concentration of each component and the intensities of saltiness, sourness, sweetness, umami and bitterness sensed by humans, to calculate the intensities of saltiness, sourness, sweetness, umami and bitterness sensed by humans; and

outputting the intensities as the measured taste.

2. The method according to claim 1, wherein said first and second phase radial basis function neural network comprises a basis-auto-optimization algorithm and a weight-repressing decay term-adding algorithm.

3. The method according to claim 1 or 2, wherein Na⁺, K⁺ and Cl⁻

are selected as the components representing saltiness;

H⁺ is selected as the component representing sourness;

glucose and sucrose are selected as the components representing sweetness;

glutamate is selected as the component representing umami; and

caffeine is selected as the component representing bitterness; and

sensors for measuring these eight components are used.

4. The method according to claim 3, wherein said first phase radial basis function neural network comprises:

a radial basis function neural network (1) into which the response value from the sensor measuring Na⁺ is input, and which calculates the Na⁺ concentration;

a radial basis function neural network (2) into which the response value from the sensor measuring K⁺ is input, and which calculates the K⁺ concentration;

a radial basis function neural network (3) into which the response value from the sensor measuring Cl⁻

is input, and which calculates the Cl⁻

concentration;

a radial basis function neural network (4) into which the response value from the sensor measuring H⁺ is input, and which calculates the pH;

a radial basis function neural network (5) into which the response value from the sensor measuring sucrose, the response value from the sensor measuring glucose, and the pH are input, and which calculates the sucrose and glucose concentrations;

a radial basis function neural network (6) into which the response value from the sensor measuring glutamate and the pH are input, and which calculates the glutamate concentration; and

a radial basis function neural network (7) into which the response value from the sensor measuring caffeine is input, and which calculates the caffeine concentration.

5. The method according to claim 1, wherein said intensities of the saltiness, sourness, sweetness, umami and bitterness sensed by humans are those obtained by a method wherein panelists taste five standard samples, which independently represent saltiness, sourness, sweetness, umami and bitterness, respectively;

then the panelists taste a plurality of samples for learning and sensorically evaluate the intensities of said five tastes, respectively; and

the panelists express the evaluated intensities in terms of numerical values by comparing the intensities with the intensity of each taste of said standard samples.

6. The method according to claim 5, wherein said standard samples consist of two standard samples for each taste, which have different concentrations, respectively, and the results of said sensory evaluation of said samples for learning are expressed in terms of values by rating on a 5-point scale which is (1) no taste at all, (2) tastes weaker than the lower concentration standard sample, (3) tastes equally to the lower concentration standard sample, (4) tastes equal to between the lower concentration standard sample and the higher concentration standard sample, and (5) tastes equal to or stronger than the higher concentration standard sample.

7. The method according to claim 1, wherein said second phase radial basis function neural network also correlates the concentration of each component with the standard deviation of intensities of saltiness, sourness, sweetness, umami and bitterness sensed by humans, and calculates variations in the intensities of saltiness, sourness, sweetness, umami and bitterness sensed by humans.

8. The method according to claim 1, said taste sensor comprising sensors, each of which sensors can quantify at least one component representing, individually or cooperatively, the taste of saltiness, sourness, sweetness, umami or bitterness, to obtain a response value from each sensor.

9. An apparatus for measuring tastes, said apparatus comprising said taste sensor according to claim 8 and a computer that stores and executes a computer program that performs the method of claim 8.