Method, system, and program for evaluating the moisture content of processed shrimp products.

Abstract

This technology provides a way to quickly and objectively evaluate the moisture content (water retention, etc.) of processed shrimp products in a manner that is robust enough for inter-lot comparisons, without relying on measurements involving destruction or heating, or on the subjective judgment of the measurer. [Solution] T2 relaxation data of a sample is acquired using a low-field nuclear magnetic resonance spectrometer, and a T2 distribution 41 is calculated by reconstitution processing. The T2 distribution is divided into a first region 42, a second region 43, and a third region 44 using boundary values ​​T2A (45) and T2B (46). Areas A1 (50), A2 (51), and A3 (52) of each region are calculated, and indices J1 and J2 are calculated. If indices J1 and J2 each meet the judgment criteria, the product is judged to conform to the quality standard; otherwise, it is judged not to conform to the quality standard, and the indices and judgment results are output.

Description

Technical Field

[0001] The present invention relates to a quality evaluation technology in the food field, and particularly to a method, system, and program for evaluating the moisture state of shrimp processed products, which are aquatic processed products, using T2 relaxation data obtained by low-field nuclear magnetic resonance measurement and the T2 distribution calculated by applying an inversion process to the T2 relaxation data, and calculating and outputting a moisture state index based on the evaluation.

Background Art

[0002] Shrimp processed products are widely used in business applications such as out-of-home dining, in-home dining, and school lunches. The yields after thawing and heating, elasticity (texture), and retention of flavor are important quality requirements. In particular, fluctuations in the water retention state affect the quality through increased thawing drip, heating loss, softening or hardening of the texture, and elution of flavor components.

[0003] Conventionally, the moisture state and water retention of shrimp processed products have been evaluated by measuring the moisture content by a drying method, measuring the amount of water separated by centrifugation or pressurization, measuring the thawing drip rate and heating loss rate, physical property tests such as texture profile analysis (TPA), and sensory evaluation. However, these evaluations involve destruction or heating of the sample, require time and labor for measurement, and are easily affected by factors such as the measurer, measurement conditions, and sample shape. Furthermore, there is room for improvement from the viewpoints of speed, reproducibility, and objectivity when applied to business lot management and in-process management.

[0004] On the other hand, in the food field, a technology for evaluating the mobility of water molecules by nuclear magnetic resonance (NMR) measurement and estimating the moisture state in food is known. For example, it has been proposed to use the distribution of T2 relaxation times (T2 distribution) by low-field NMR to grasp peak components that reflect differences in the binding state and mobility of water and utilize them for quality evaluation.

[0005] However, conventional evaluations based on T2 distribution tend to be qualitative in their interpretation of peaks, and results can vary depending on differences in measurement conditions (sample temperature, filling amount, pulse sequence, ECHO interval, number of ECHOs, etc.) and reaction conditions. Therefore, standardizing the definition and calculation procedure of the index is a challenge when using it for inter-lot comparisons or as an objective control indicator within a process. Accordingly, there is a need for a technology that can reproducibly evaluate the moisture content of processed shrimp products by dividing the T2 distribution into predetermined regions and calculating and outputting a moisture content index based on the area of ​​those regions. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] CN 106885816 A (Publication of Chinese Patent Application, Publication Date: 2017-06-23) "Non-destructive Testing Method for Injecting Adhesive into Aquatic Products" [Patent Document 2] CN 110596176 A (Chinese Patent Application Publication, Publication Date: 2019-12-20) "Method for Detecting Moisture Content in Aquatic Products Using Low Magnetic Field Nuclear Magnetic Resonance Technology" [Patent Document 3] CN 105606637 A (Chinese Patent Application Publication, Publication Date: 2016-05-25) "Method for Detecting Water and Fat Content in Abalone Using Low Magnetic Field Nuclear Magnetic Resonance Technology" [Patent Document 4] CN 105548234 A (Chinese Patent Application Publication, Publication Date: 2016-05-04) "Method for measuring moisture and fat content of Kiguchi without damaging the material" [Patent Document 5] Japanese Patent Publication No. 2008-203230 (JP2008203230A) (Publication Date: 2008-09-04, Title of Invention: IDENTIFICATION AND QUANTIFICATION METHOD OF MARBLING AND NON-INVASIVE MEASURING APPARATUS OF MARBLING) [Overview of the project] [Problems that the invention aims to solve]

[0007] Conventionally, the moisture content and water retention of processed shrimp products have been evaluated by measuring moisture content, water separation, thawing drip rate and heat loss rate, physical property tests, and sensory evaluation. However, these methods involve destroying or heating the sample, require time and effort for measurement, and are susceptible to influences from the operator, measurement conditions, and sample shape. Furthermore, although evaluation using the T2 distribution obtained by low-field nuclear magnetic resonance measurement is known, if the definition and calculation procedure of the index are not standardized, it is difficult to ensure reproducibility and objectivity in inter-lot comparisons and in-process control. Therefore, the present invention aims to provide a technology that can objectively evaluate the moisture content of processed shrimp products with good reproducibility by dividing the T2 distribution obtained by low-field nuclear magnetic resonance measurement into predetermined regions and calculating and outputting a moisture content index based on the area of ​​those regions.

[0008] Patent Document 1 focuses on non-destructively detecting aquatic products that have undergone so-called injection treatment (injection of gel, etc.). It describes a technology that obtains feature quantities from acquired nuclear magnetic resonance data through reflection processing, etc., and identifies normal / abnormal or injected substances, etc., through multivariate analysis, etc. Therefore, the configuration and operation of standardizing the T2 distribution by boundary values ​​to divide it into regions in order to control the moisture state (water retention, etc.) of processed shrimp products within the process, and determining whether or not it meets quality standards from the ratio of the region area, cannot be derived from Patent Document 1.

[0009] Patent documents 2 to 4 primarily focus on estimating "contents" such as water content and fat content by constructing regression models (e.g., PLSR) based on reference measurements such as drying methods, using relaxation data or T2 distributions obtained from low-field nuclear magnetic resonance measurements. While these may presuppose calibration and statistical learning of numerous samples for model construction, they do not disclose or suggest an explainable and standardized quality gate configuration suitable for in-process control, such as fixing the T2 distribution into 1st to 3rd regions using boundary values ​​T2A and T2B, and determining the specifications using a threshold table based on ratio indices (J1, J2) of the areas A1 to A3 of each region.

[0010] Patent Document 5 describes a technique for identifying and quantifying marbling in animal or fish meat using a signal based on the difference in relaxation characteristics between fat and muscle. Therefore, it differs from the present invention in its technical challenges and indicator design, which involves dividing the T2 distribution into regions for the purpose of evaluating the moisture content, calculating the relative ratio and bias of the long T2 side components as area ratios (J1, J2), and determining whether or not the quality standard is met based on these ratios.

[0011] As described above, the prior art does not present any technology that provides as an integrated method / system / program a series of frameworks, including (i) targeting the moisture content of processed shrimp products, (ii) standardizing the region division of the T2 distribution by fixing or calibrating the measurement conditions and boundary values, (iii) using ratio indices J1 and J2 calculated from region areas A1, A2, and A3, and (iv) performing standard determination (quality gate) within the process by a simple comparison process based on a threshold table.

[0012] The present invention does not require regression models or inference using multivariate analysis. By defining the region divisions of the T2 distribution (division → area → index) and standardizing the index calculation procedure, and performing specification judgments based on the calculated indices (J1, J2) and thresholds, it ensures objectivity, reproducibility, and explainability suitable for in-process control and acceptance inspection. Furthermore, by using index J1, which represents the ratio normalized by the total area, and index J2, which represents the bias of the third region component relative to the first and second regions, in combination for specification judgments, the influence of fluctuations in evaluation values ​​caused by small variations in measurement conditions or differences in sample filling amounts can be reduced, thereby improving the stability of pass / fail judgments. [Means for solving the problem]

[0013] To solve the above problems, the present invention provides a method for acquiring T2 relaxation data from a sample of processed shrimp using a low-field nuclear magnetic resonance spectrometer, calculating a T2 distribution by applying a reflection process to the T2 relaxation data, dividing the T2 distribution into at least a first region, a second region, and a third region, calculating areas A1, A2, and A3 corresponding to each region, and calculating and outputting at least one moisture state index based on areas A1, A2, and A3.

[0014] Furthermore, the present invention provides an embodiment in which at least one of A3 / (A1+A2+A3) and A3 / (A1+A2) is used as the moisture state index, and an embodiment in which the classification is performed by boundary values ​​T2A and T2B of T2, and the boundary values ​​T2A and T2B are values ​​stored in the storage unit or values ​​set by device calibration.

[0015] Furthermore, the present invention provides a configuration in which the acquisition of T2 relaxation data is performed by fixing measurement conditions including at least one of the sample temperature, sample mass (filling amount), pulse sequence, ECHO interval, and number of ECHOs, and the calculation of the T2 distribution is performed by fixing the reflection conditions.

[0016] Furthermore, the present invention provides a system for performing the above method as a quality evaluation system, comprising a low-field nuclear magnetic resonance spectrometer and a calculation unit that calculates the T2 distribution by applying reflection processing to the T2 relaxation data acquired by the low-field nuclear magnetic resonance spectrometer, calculates the region divisions of the T2 distribution and areas A1, A2, and A3, and calculates and outputs a moisture state index based on areas A1, A2, and A3.

[0017] Furthermore, the present invention provides a program that allows a computer to apply a reflection process to T2 relaxation data to calculate a T2 distribution, divide the T2 distribution into at least a first region, a second region, and a third region, calculate areas A1, A2, and A3 corresponding to each region, and calculate and output a moisture state index based on areas A1, A2, and A3. [Effects of the Invention]

[0018] According to the present invention, T2 relaxation data is obtained from a sample of processed shrimp by low-field nuclear magnetic resonance measurement, a reflection process is applied to the T2 relaxation data to calculate the T2 distribution, the T2 distribution is divided into at least a first region, a second region, and a third region, and the areas A1, A2, and A3 of each region are calculated, and a moisture state index based on these is calculated and output, thereby enabling a non-destructive and objective evaluation of the moisture state of processed shrimp.

[0019] According to the present invention, the T2 distribution is divided into predetermined regions and output as indices based on areas A1, A2, and A3. This enables quantitative evaluation suitable for lot-to-lot comparisons, compared to qualitative peak interpretation which is prone to subjectivity of the measurer.

[0020] Furthermore, when threshold determination is performed using only J1, slight fluctuations in sample filling volume and measurement conditions may prevent sufficient constraint on the bias in the contribution of the third region component, potentially leading to misjudgments. In contrast, the present invention uses J2 in combination and requires both to be within the standard range, thereby constraining the "relative ratio" and "bias" of the long T2 side component on separate axes, improving the reproducibility of in-process pass / fail determination. According to the present invention, by using at least one of A3 / (A1+A2+A3) and A3 / (A1+A2) as a moisture state index, it is possible to reflect changes in the relative ratio of components with high water molecule mobility, thus making it possible to easily grasp differences in the moisture state of processed shrimp products. In particular, index J1(A3 / (A1+A2+A3)) is a ratio normalized by the total area, so it is easy to stabilize as a relative evaluation. On the other hand, index J2(A3 / (A1+A2)) represents the bias of the third region component relative to the first and second regions, so even in regions where J1 is small, it is easy to grasp the increase or decrease of the third region component as a relative ratio. Preferably, by using both J1 and J2 for standard determination, evaluation can be performed from both the relative ratio and bias aspects, and the reproducibility of the determination can be improved.

[0021] According to the present invention, by classifying the boundary values ​​T2A and T2B of T2 as values ​​stored in the memory unit or values ​​set by device calibration, the classification process of the T2 distribution can be standardized and the reproducibility of the evaluation results can be improved.

[0022] According to the present invention, by fixing measurement conditions including at least one of sample temperature, sample mass, pulse sequence, ECHO interval, and number of ECHOes to obtain T2 relaxation data, and fixing inversion conditions to calculate the T2 distribution, it is possible to suppress variations in evaluation values caused by fluctuations in the measurement conditions or inversion conditions, and perform stable evaluation suitable for in-process management or incoming inspection.

[0023] According to the present invention, by providing the above evaluation method as a system or program, calculation of the T2 distribution, region classification, area calculation, and calculation / output of the moisture state index can be executed as a series of processes, contributing to the acceleration and labor saving of the quality evaluation work of shrimp processed products.

Brief Description of Drawings

[0024] [Figure 1] It is a flowchart showing an example of the processing procedure of the moisture state evaluation method for shrimp processed products according to the present invention. This flowchart includes a process of obtaining T2 relaxation data 40, a process of applying an inversion process to the T2 relaxation data 40 to calculate the T2 distribution 41, a process of setting boundary values T2A (45) and T2B (46) and dividing the T2 distribution 41 into a first region (T2 < T2A), a second region 43 (T2A ≤ T2 < T2B), and a third region 44 (T2B ≤ T2), a process of calculating areas A1 (50), area A2, and area A3 (52) corresponding to each region, a process of calculating index J1 and index J2 based on the areas, and a process of outputting the index and the standard determination result, including at least in the order of "classification → area → index". [Figure 2] An example of the T2 distribution 41 obtained by low-field nuclear magnetic resonance measurement, and a conceptual diagram showing the concept of dividing the T2 distribution 41 into a first region 42 (T2 < T2A), a second region 43 (T2A ≤ T2 < T2B), and a third region 44 (T2B ≤ T2) by boundary values T2A (45) and T2B (46), calculating areas A1 (50), area A2 (51), and area A3 (52) corresponding to each region, and evaluating the relative ratio and bias of the third region 44 component by index J1 and index J2. [Figure 3] This is an explanatory diagram showing an example of how the T2 distribution 41 is divided by the boundary values ​​T2A(45) and T2B(46) of T2. [Figure 4] This is a block diagram showing an example configuration of a system for evaluating the moisture content of processed shrimp products according to the present invention, comprising at least a low-field nuclear magnetic resonance spectrometer 10 (including a control unit 11, a receiving unit 12, a magnetic field generating unit 13, and a sample holder 14), a calculation unit 20, a storage unit 21, and an output unit 22. The diagram shows the data flow in which T2 relaxation data 40 acquired by the low-field nuclear magnetic resonance spectrometer 10 is input to the calculation unit 20, the calculation unit 20 calculates the T2 distribution 41, area A1 (50), area A2 (51), area A3 (52), index J1 and index J2, and the output unit 22 outputs the index and standard judgment results. [Figure 5] This flowchart shows an example of a processing procedure that performs calculation of the T2 distribution 41, region division, calculation of area A1(50), area A2(51), area A3(52), and calculation and output of a moisture state index using the program according to the present invention. This flowchart includes at least the steps of "division → area → index", which are performed in the order [Modes for carrying out the invention]

[0025] <Overall Structure> Embodiments of the present invention will be described below with reference to the drawings. The same or corresponding parts are denoted by the same reference numerals, and redundant explanations will be omitted as appropriate. Furthermore, descriptions of the figures included in the brief descriptions of the drawings will be omitted to avoid redundancy, and the figure numbers will be referred to as necessary.

[0026] The present invention provides a basic configuration for evaluating the moisture content of processed shrimp products, which involves using a low-field nuclear magnetic resonance spectrometer 10 to acquire T2 relaxation data 40 from a sample of processed shrimp products, calculating a T2 distribution 41 from the T2 relaxation data 40, dividing the T2 distribution 41 into at least a first region 42, a second region 43, and a third region 44, calculating the areas A1(50), A2(51), and A3(52) corresponding to each region, and calculating and outputting a moisture content index 60 based on these areas.

[0027] In the present invention, the classification can be performed using boundary values ​​T2A(45) and T2B(46) of T2, and these boundary values ​​T2A(45) and T2B(46) may be set using values ​​stored in the storage unit 21, or they may be set by device calibration.

[0028] The present invention can be configured as a moisture state evaluation system (Claim 5) comprising a low-field nuclear magnetic resonance apparatus 10, a calculation unit 20, a storage unit 21, and an output unit 22, as an embodiment of the above method. The calculation unit 20 performs calculation of the T2 distribution 41, region division, calculation of area A1 (50), area A2 (51), area A3 (52), and calculation and output of the moisture state index 60.

[0029] The present invention can also be provided as a program (claim 6) that causes a computer 30 to perform the above processing. In this case, the processor 31 of the computer 30 performs the calculation of the T2 distribution 41, region division, calculation of area A1 (50), area A2 (51), area A3 (52), and calculation and output of the moisture state index 60 according to the program stored in the memory 32.

[0030] The following sections will explain, with specific examples, the evaluation target, the configuration of the measuring device, the measurement conditions, the acquisition of T2 relaxation data 40, the calculation of the T2 distribution 41, the setting of region divisions and boundary values, the calculation of area, the calculation and output of moisture state indices, and the ensuring of reproducibility by fixing the measurement conditions.

[0031] <Definition of Terms> In this specification, "shrimp processed products" refers to processed products made from shrimp such as Pacific white shrimp and kuruma shrimp, which are distributed or provided to the public, regardless of whether they are shelled or peeled, whether the veins have been removed or not, whether they are cooked or uncooked, whether they are seasoned or not, and whether they are frozen or not.

[0032] In this specification, "low-field nuclear magnetic resonance apparatus 10" means an apparatus that measures nuclear magnetic resonance signals in a static magnetic field formed by a permanent magnet or electromagnet, and includes a transmitting and receiving system (RF coil, etc.) for exciting and receiving nuclear magnetic resonance signals, and is capable of acquiring data relating to at least T2 relaxation.

[0033] In this specification, "T2 relaxation data 40" refers to a time-series signal relating to spin-spin relaxation acquired by a low-field nuclear magnetic resonance spectrometer 10, and includes, for example, an echo train signal acquired by the CPMG method.

[0034] In this specification, "spin-spin relaxation" refers to the phenomenon (transverse relaxation) in which the transverse magnetization of nuclear spins excited in a static magnetic field decays due to the interaction between spins (including fluctuations in the local magnetic field), and its time constant is called the T2 relaxation time. For example, the echo train signal obtained by the CPMG method shows a decrease in echo amplitude over time in response to this decay of transverse magnetization, and therefore this echo train signal can be used as T2 relaxation data 40.

[0035] In this specification, "T2 distribution 41" refers to the distribution of signal intensity or contribution rate with respect to T2 relaxation time, calculated by reflection processing or the like based on the T2 relaxation data 40.

[0036] In this specification, "regional division" means dividing the T2 distribution 41 into multiple regions according to the range of T2 relaxation times, and in this embodiment, it is divided into at least a first region 42, a second region 43, and a third region 44.

[0037] As used herein, the "boundary values T2A(45), T2B(46)" are the threshold values of the T2 relaxation time used for the region classification. The boundary value T2A(45) defines the boundary between the first region 42 and the second region 43, and the boundary value T2B(46) defines the boundary between the second region 43 and the third region 44. As used herein, the "first region 42, second region 43, and third region 44" are regions obtained by dividing the T2 distribution 41 with the boundary values T2A(45) and T2B(46) on the T2 axis. For example, in the first region 42, T2 < T2A; in the second region 43, T2A ≤ T2 < T2B; and in the third region 44, T2B ≤ T2.

[0038] As used herein, the "areas A1(50), A2(51), A3(52)" refer to the integral values (the total signal contribution within the region) of the distribution curves corresponding to the first region 42, second region 43, and third region 44, respectively, in the T2 distribution 41 after the region classification.

[0039] As used herein, the "moisture state" refers to the distribution of the mobility or binding state of moisture in the sample (the composition ratio and bias of the free water component and the bound water component) reflected in the T2 distribution 41. As used herein, the "moisture state index" is an index calculated based on the areas A1(50), A2(51), A3(52), and includes, for example, at least one of A3 / (A1 + A2 + A3) and A3 / (A1 + A2). Also, as used herein, the "quality standard" refers to the pass / fail criteria predetermined in the incoming inspection, in-process control, or shipment determination. Further, as used herein, the "standard determination" refers to the process of storing the pass / fail criteria related to the quality standard in the storage unit 21 as the threshold values (upper limit value / lower limit value or tolerance range) of the indicators J1 and J2, and comparing the calculated indicators with the threshold values to determine compliance / non-compliance with the quality standard.

[0040] As used herein, the "measurement conditions" are the conditions used for acquiring the T2 relaxation data 40, and include, for example, at least one of the sample temperature, sample mass (filling amount), pulse sequence, ECHO interval, and number of ECHOes. [[ID=第十九]]

[0041] In this specification, "performing measurements under fixed conditions" means obtaining T2 relaxation data 40 by maintaining at least a portion of the measurement conditions within a predetermined value or range in a lot-to-lot comparison or continuous evaluation.

[0042] In this specification, “inversion conditions” refers to settings related to the inversion process for calculating the T2 distribution 41 by applying the inversion process to the T2 relaxation data 40, and includes, for example, at least one of the following: regularization conditions, number of inversion iterations, initial value settings, and distribution discretization conditions.

[0043] <Standardization of verb terminology> In this specification, claims, and drawings, verbs representing processes shall be used in at least the following senses. In particular, the same process shall, in principle, be described using the verbs "acquire," "calculate," "set," "classify," "determine," and "output," and the same process shall not be mixed with other expressions such as "estimate," "calculate," and "request." (1) Acquisition: Obtain T2 relaxation data 40 using a low-field nuclear magnetic resonance spectrometer 10. (2) Calculation: The T2 distribution 41, area A1 (50), area A2 (51), area A3 (52) for each region, and indices J1, J2, etc. shall be calculated and obtained by processing such as the counter-processing. (3) Settings: Define and use values ​​such as boundary values ​​T2A(45), T2B(46), judgment criteria, and thresholds. (4) Classification: Divide the T2 distribution 41 into the first region 42, the second region 43, and the third region 44 based on the boundary values ​​T2A (45) and T2B (46). (5) Judgment: Determine compliance with the standard by comparing indicators J1 and J2 with the judgment criteria (including standard judgment). (6) Output: The calculated indicators, area, boundary values, and judgment results shall be displayed, recorded, printed, or transmitted to external parties.

[0044] <Evaluation target> The products to be evaluated in this embodiment are processed shrimp products, including, for example, shrimp with or without shells, shrimp with or without deveining, cooked or uncooked shrimp, shrimp with or without seasoning, and frozen or unfrozen products.

[0045] The shrimp products subject to evaluation include, but are not limited to, frozen products used in commercial distribution, and include, for example, those that have undergone processing such as brine immersion, seasoning, water retention treatment, coating treatment, or boiling treatment before freezing.

[0046] The form of the sample to be evaluated may be, for example, a solid sample, a sample made by mixing multiple samples, or a sample whose shape has been adjusted by crushing, cutting, etc. However, when performing lot-to-lot comparisons or continuous evaluations, it is preferable to standardize the sample shape, mass (filling amount), and temperature conditions.

[0047] The sampling unit to be evaluated may be, for example, a sample randomly drawn from a predetermined number of individuals per lot, or a sample mixed to a predetermined weight. The sampling method and sampling location may be set according to the purpose of the evaluation (e.g., receiving inspection, in-process control, shipping decision).

[0048] In this embodiment, the moisture state of the target is evaluated based on the T2 distribution 41 obtained by low-field nuclear magnetic resonance measurement, and a moisture state index obtained by region division and area calculation is output.

[0049] <Configuration of the measuring device> In this embodiment, the measuring device used to evaluate the moisture content of processed shrimp products is a low-field nuclear magnetic resonance apparatus 10. The low-field nuclear magnetic resonance apparatus 10 comprises a magnetic field generating unit 13 that forms a static magnetic field, a sample holder 14 that houses the sample, a transmitting and receiving system (including transmitting and receiving elements such as RF coils, but not shown) that excites and receives nuclear magnetic resonance signals, a receiving unit 12 that acquires the received signal, and a control unit 11 that controls the measurement operation.

[0050] The magnetic field generating unit 13 forms a static magnetic field using a permanent magnet or electromagnet and applies a predetermined magnetic field strength to the sample placed in the sample holder 14. The sample holder 14 has a structure that holds the sample in a predetermined position and ensures the reproducibility of the sample position during measurement.

[0051] The transmitting and receiving system, under the control of the control unit 11, applies RF pulses according to a predetermined pulse sequence to the sample in the static magnetic field and receives nuclear magnetic resonance signals from the sample. The receiving unit 12 amplifies, detects, and converts the received signal into digital data through A / D conversion, etc., and acquires it as T2 relaxation data 40.

[0052] The low-field nuclear magnetic resonance spectrometer 10 may be equipped with operating means (not shown) for receiving settings for measurement conditions (e.g., pulse sequence, ECHO interval, number of ECHOs, repetition time, etc.), and may also be equipped with a communication interface for outputting measurement results to the outside.

[0053] In this embodiment, the T2 relaxation data 40 acquired by the low-field nuclear magnetic resonance spectrometer 10 is input to the calculation unit 20, where the calculation unit 20 performs calculation of the T2 distribution 41, region division, calculation of area A1 (50), area A2 (51), area A3 (52), and calculation and output of the moisture state index 60. The calculation unit 20 may be built into the low-field nuclear magnetic resonance spectrometer 10 or may be configured as an external computer 30.

[0054] The calculation unit 20 includes, for example, a processor 31 such as a CPU, GPU, or DSP, and performs calculation processing based on a program or parameters stored in the storage unit 21. The storage unit 21 may store setting values ​​such as boundary values ​​T2A(45) and T2B(46), as well as measurement conditions. The output unit 22 outputs the moisture state index 60, or the evaluation result based thereon, via a display device, printer, file, or communication.

[0055] <Measurement conditions> In this embodiment, the measurement conditions used to acquire the T2 relaxation data 40 are preferably fixed in part, from the viewpoint of ensuring the reproducibility of the evaluation and the comparability between lots. The measurement conditions include, for example, at least one of the following: sample temperature, sample mass (filling amount), pulse sequence, ECHO interval, number of ECHOs, and repetition time (REPETITIONTIME).

[0056] Since the sample temperature affects the T2 relaxation time, it is preferable to adjust the temperature to a predetermined temperature (e.g., refrigeration temperature range or room temperature range) at the time of measurement and to suppress temperature fluctuations during measurement. The sample temperature may also be controlled by temperature control of the sample holder 14 or by maintaining a constant temperature before measurement.

[0057] Since the sample mass (filling amount) affects the filling state and signal intensity within the RF coil, it is preferable to standardize the mass or volume to a predetermined level. When mixing multiple samples, it is preferable to standardize the degree of homogenization and the mixing ratio.

[0058] The pulse sequence is a signal acquisition method used to acquire T2 relaxation data 40, and for example, the CPMG method can be used. The ECHO interval and number of ECHOs affect the resolution of the T2 distribution 41 and the acquireable T2 range, so they should be set according to the measurement target, and it is preferable to use the same conditions when performing lot-to-lot comparisons.

[0059] The repetition time is set to suppress the effects of longitudinal relaxation and ensure the stability of the measurement, and may be set to a value above a predetermined range, for example.

[0060] The re-examination conditions are settings related to the re-examination process for calculating the T2 distribution 41 from the T2 relaxation data 40, and include, for example, the discretization conditions of the distribution, the regularization conditions, the number of iterations, and the convergence determination conditions. The re-examination conditions may be set to predetermined values ​​by the device or calculation unit 20 and calibrated according to the purpose of evaluation, but it is preferable to use the same conditions when performing inter-lot comparisons.

[0061] By fixing at least a part of the above measurement conditions and fixing the inversion conditions, variations in the T2 distribution 41 and the moisture state index 60 caused by variations in the measurement conditions or inversion conditions are suppressed, enabling stable evaluation suitable for in-process management, incoming inspection, or shipment determination, etc.

[0062] <Acquisition of T2 relaxation data 40> In this embodiment, the T2 relaxation data 40 is acquired using the low-field nuclear magnetic resonance apparatus 10 with a sample of the processed shrimp product accommodated in the sample holder 14. The sample is placed in the measurement space after adjusting the measurement conditions (such as sample temperature, mass (filling amount), etc.) to predetermined values or a predetermined range.

[0063] The control unit 11 of the low-field nuclear magnetic resonance apparatus 10 applies an RF pulse according to the set pulse sequence and receives the nuclear magnetic resonance signal from the sample. As the pulse sequence, for example, the CPMG method can be used, by which an echo train signal is acquired and the T2 relaxation data 40 is obtained as a time-series signal related to T2 relaxation.

[0064] The receiving unit 12 digitizes the received signal by amplification, detection, A / D conversion, etc., and records it in the storage unit 21 or an external storage medium as the T2 relaxation data 40. The T2 relaxation data 40 may include the correspondence between the echo number and the signal intensity, as well as the measurement conditions (such as ECHO interval, number of ECHOes, repetition time, etc.) as accompanying information.

[0065] When performing inter-lot comparison or continuous evaluation, the acquisition of the T2 relaxation data 40 is preferably performed with the measurement conditions fixed. For example, it is preferable to acquire it with the same pulse sequence, the same ECHO interval, and the same number of ECHOes. Also, if necessary, the same sample may be measured multiple times and statistical processing such as outlier removal or averaging may be performed.

[0066] The T2 relaxation data 40 acquired as described above is used by the arithmetic unit 20 to calculate the T2 distribution 41.

[0067] <Calculation of T2 distribution 41> In this embodiment, the T2 distribution 41 is calculated by the calculation unit 20 performing a re-examination process on the T2 relaxation data 40 acquired by the low-field nuclear magnetic resonance spectrometer 10. The calculation unit 20 receives the T2 relaxation data 40 (for example, an echo train signal obtained by the CPMG method) as input and calculates the T2 distribution 41 as the distribution of signal contributions to the T2 relaxation time using a predetermined re-examination algorithm.

[0068] As a counter-reaction algorithm, for example, the inverse Laplace transform, the least squares method with non-negative constraints, an iterative solution method using regularization, or a similar method can be used. The calculation unit 20 calculates the T2 distribution 41 based on the counter-reaction conditions (for example, the discretization condition of the distribution, the regularization condition, the number of iterations, the convergence criterion, etc.).

[0069] The T2 distribution 41 is represented as a distribution in which the signal intensity or contribution rate of each T2 component is shown with respect to the axis of T2 relaxation time (e.g., logarithmic axis). The T2 distribution 41 may be observed as multiple peaks or shoulders to reflect differences in the mobility of the water molecules being measured.

[0070] When performing lot-to-lot comparisons or continuous evaluations, it is preferable to fix the reaction conditions used to calculate the T2 distribution 41. For example, by keeping the number of discretization points of the distribution, the T2 range, the regularization parameter, and the number of iterations constant, differences in the distribution shape caused by differences in reaction conditions can be suppressed, and the reproducibility of the moisture state index 60 can be improved.

[0071] The calculated T2 distribution 41 is used in the next step to calculate the region divisions and areas A1(50), A2(51), and A3(52).

[0072] <Setting of region divisions and boundary values ​​T2A(45), T2B(46)> In this embodiment, the arithmetic unit 20 divides the calculated T2 distribution 41 into at least a first region 42, a second region 43, and a third region 44. The region division is performed based on the range of the T2 relaxation time. For example, using the boundary value T2A (45) and the boundary value T2B (46), it can be divided such that T2 < T2A is the first region 42, T2A ≤ T2 < T2B is the second region 43, and T2B ≤ T2 is the third region 44.

[0073] The boundary values T2A (45) and T2B (46) may be set as predetermined values held in the storage unit 21 or set by device calibration from the viewpoint of ensuring the reproducibility of the region division. For example, calculate T2 41 for a standard sample or a sample of a reference lot, determine T2A and T2B based on a predetermined peak or distribution shape, and save the determined values in the storage unit 21 for use in subsequent evaluations.

[0074] As a method for determining the boundary values T2A (45) and T2B (46), for example, a method of setting the valley (local minimum value) between peaks in the T2 distribution 41 as the boundary, a method of setting a predetermined T2 range as a fixed value, or a method of correcting the boundary following the movement of the peak position can be used.

[0075] When setting the boundary values T2A (45) and T2B (46) by device calibration, it is preferable to calculate the T2 distribution 41 using a standard sample or a representative sample after fixing the measurement conditions (sample temperature, ECHO interval, number of ECHOs, etc.) and the inversion conditions, etc., and derive the boundary values from the distribution. Thereby, the fluctuation of the boundary setting due to the difference between devices or the difference in measurement conditions can be reduced.

[0076] Also, when performing inter-lot comparison or continuous evaluation, it is preferable to operate with the boundary values T2A (45) and T2B (46) fixed for a certain period. For example, it may be an operation to perform re-calibration only when device maintenance or measurement condition changes are made and update the boundary values.

[0077] By setting boundary values ​​T2A(45) and T2B(46) as described above and performing region division, the subsequent area calculation and moisture status index calculation 60 can be standardized without relying on the qualitative peak interpretation of the T2 distribution 41.

[0078] <Calculation of areas A1 (50), A2 (51), and A3 (52)> In this embodiment, the calculation unit 20 calculates the areas A1(50), A2(51), and A3(52) corresponding to the first region 42, second region 43, and third region 44, respectively, based on the T2 distribution 41 after region division.

[0079] Areas A1(50), A2(51), and A3(52) may be obtained by summing the signal contributions of each point on the discretized T2 axis for each region, or they may be calculated by numerical integration such as the trapezoidal rule. Areas A1(50), A2(51), and A3(52) are obtained by integrating (or summing) the distribution curve (or discretized distribution values) of the T2 distribution 41 within the T2 range of each region. For example, if the calculation unit 20 represents the T2 distribution 41 with predetermined discrete points (T2i), the value equivalent to the area may be calculated by summing the distribution values ​​of the discrete points belonging to each region.

[0080] The T2 distribution 41 is the distribution of signal intensity or contribution rate with respect to T2 relaxation time. Generally, components with shorter T2 times reflect a state in which molecular motion is restricted, while components with longer T2 times reflect a state in which molecular motion is relatively free. Therefore, the areas A1(50), A2(51), and A3(52) of the regions divided by T2A and T2B can be used as values ​​representing the sum of the signal contributions belonging to the respective states. In particular, the area A3(52) of the third region 44 can be used as an indicator representing the contribution of components on the longer T2 side (components with relatively high motility).

[0081] The integration method used to calculate the area is not particularly limited; for example, numerical integration such as the trapezoidal rule, a simple sum of discrete values, or a similar method can be used. However, when performing lot-to-lot comparisons or continuous evaluations, it is preferable to use the same discretization conditions and the same integration method.

[0082] Areas A1(50), A2(51), and A3(52) may be normalized as needed with respect to the overall contribution calculated as the total area AT = A1 + A2 + A3. For example, to reduce the effect of variability in the overall signal intensity of the T2 distribution, areas A1(50), A2(51), and A3(52) can be used as normalized areas obtained by dividing them by AT.

[0083] Normalization using the total area AT (calculation of J1) is performed to enable evaluation based on the relative ratio between regions, even when the overall signal intensity of the T2 distribution 41 fluctuates due to factors such as sample mass (filling amount), filling state within the RF coil, and receiving gain. This reduces the impact of variations in absolute signal intensity in inter-lot comparisons and in-process judgments.

[0084] Furthermore, prior to calculating the area, noise reduction processing, baseline correction processing, or outlier removal processing may be applied to the T2 distribution 41. This reduces the influence of noise, particularly in the low-signal region or on the long T2 side, and improves the stability of areas A1(50), A2(51), and A3(52).

[0085] The areas A1(50), A2(51), and A3(52) calculated as described above are used in the next step to calculate and output the moisture status index 60.

[0086] <Calculation and output of moisture status index> In this embodiment, the calculation unit 20 calculates a moisture state index 60 based on the calculated areas A1(50), A2(51), and A3(52). The moisture state index 60 is at least one index obtained from the combination of areas A1(50), A2(51), and A3(52), and includes, for example, at least one of the indices shown in equation (1) and equation (2).

[0087] (Equation 1) J1 = A3 / (A1 + A2 + A3) (Formula 2) J2=A3 / (A1+A2) Furthermore, if A1+A2 falls below a predetermined threshold, the calculation of J2 may be skipped, or A1+A2 may be replaced with a predetermined minimum value for calculation, or other processing may be performed to ensure numerical stability.

[0088] Here, J1 is the ratio (0 to 1) of the area A3 (52) of the third region 44 to the total area AT, and is an index representing the relative ratio of the components of the third region 44. On the other hand, J2 is the ratio of A3 to the sum of the first region 42 and the second region 43 (A1 + A2), and is an index representing the degree to which the components of the third region 44 are biased (relatively dominant) compared to other regions. J2 can also be expressed as J2 = A3 / (AT - A3) = J1 / (1 - J1), making it easier to grasp the change as a relative ratio even in regions where J1 is small.

[0089] Preferably, both J1 and J2 are calculated and output as the moisture state index 60, and these are used in combination for specification determination. This allows for understanding the relative ratio of the 44 components in the third region using J1, and understanding the bias toward the first region 42 and the second region 43 on a separate axis using J2, thereby improving the stability of specification determination. However, depending on the application, only one of J1 or J2 may be calculated and output. In addition, if necessary, the areas A1 (50), A2 (51), A3 (52), and the total area AT = A1 + A2 + A3 may also be output.

[0090] The calculation unit 20 outputs the calculated moisture content index 60 via the output unit 22. The output format is not particularly limited and includes, for example, displaying it as a numerical value on a display device, recording it in a file (e.g., CSV format), printing it, or transmitting it to an external device via communication. The calculation unit 20 may also determine whether the processed shrimp product conforms to quality standards based on pre-set criteria (for example, the allowable range of index J1 and index J2 (greater than or equal to the lower limit and less than or equal to the upper limit), or the upper limit, or the lower limit), and output the result of this standard determination. The criteria can be managed as values ​​stored in, for example, the storage unit 21 and used as quality gates such as acceptance determination, in-process determination, or shipment determination.

[0091] Furthermore, the calculation unit 20 may calculate a moisture state index 60 for multiple samples and output statistical quantities (mean, standard deviation, etc.) together. This makes it easier to understand variations within the same lot or to compare between lots.

[0092] Furthermore, the output unit 22 may output the moisture state index 60 along with the measurement conditions (sample temperature, sample mass (filling amount), pulse sequence, ECHO interval, ECHO count, etc.), reaction conditions, and boundary values ​​T2A (45) and T2B (46) as supplementary information. This ensures traceability of the evaluation results.

[0093] As described above, by calculating and outputting the moisture status index 60, it becomes possible to quantitatively understand the moisture status of processed shrimp products without relying on the qualitative interpretation of the T2 distribution 41.

[0094] <Ensuring reproducibility by fixing measurement conditions> In this embodiment, the T2 distribution 41 and the moisture state index 60 may be affected by measurement conditions (sample temperature, sample mass (filling amount), pulse sequence, ECHO interval, ECHO count, etc.) and reaction conditions. Therefore, for applications such as inter-lot comparison, in-process control, acceptance inspection, or shipment judgment, it is preferable to fix at least some of the measurement conditions and reaction conditions during operation.

[0095] For example, if the sample temperature fluctuates, the T2 relaxation time may change due to changes in the mobility of water molecules, and the peak position or shape of the T2 distribution 41 may change. Also, if the sample mass (filling amount) fluctuates, the filling state and signal intensity in the RF coil may change, which may affect the areas A1(50), A2(51), and A3(52) of the T2 distribution 41. Therefore, it is preferable to maintain the sample temperature and sample mass (filling amount) within predetermined values ​​or ranges.

[0096] Furthermore, the ECHO interval and number of ECHOs affect the obtainable T2 range and resolution, and the response conditions (regularization conditions, discretization conditions, etc.) can affect the smoothness of the distribution and peak separation. Therefore, when performing lot-to-lot comparisons, it is preferable to use the same pulse sequence, the same ECHO interval and number of ECHOs, and the same response conditions.

[0097] As a concrete example of operating with fixed measurement conditions, for instance, the setting values ​​for the measurement conditions can be saved in the memory unit 21, and the system can be configured to automatically recall these settings and execute the measurement at the start of the measurement. Alternatively, the system can be configured to restrict changes to measurement conditions to administrator privileges only, thereby preventing changes by general users.

[0098] Furthermore, if substantial changes in measurement conditions may occur due to equipment maintenance, parts replacement, or changes in the measurement environment, it is preferable to perform calibration using a standard or representative sample and reset the boundary values ​​T2A(45), T2B(46), and response conditions. This reduces deviations in evaluation values ​​caused by differences between equipment or changes over time.

[0099] As described above, by fixing at least some of the measurement conditions and performing calibration and readjustment as necessary, the reproducibility of the T2 distribution 41 and the moisture state index 60 can be improved, making it possible to stably evaluate the moisture state of processed shrimp products.

[0100] <Example of the processing procedure for the evaluation method> Below, an example of the processing procedure for evaluating the moisture content of processed shrimp products according to the present invention will be described with reference to Figure 1.

[0101] (S101) Prepare samples of the processed shrimp product to be evaluated, and adjust the sample temperature and sample mass (filling amount) to the specified value or range. If necessary, take multiple samples from the same lot.

[0102] (S102) The sample is placed in the sample holder 14 of the low-field nuclear magnetic resonance spectrometer 10, and the measurement conditions (pulse sequence, ECHO interval, number of ECHOs, repetition time, etc.) are set or recalled to perform LF-NMR measurement and acquire T2 relaxation data 40. If necessary, it is recorded in the storage unit 21 or an external storage medium.

[0103] (S103) The calculation unit 20 applies inversion to the acquired T2 relaxation data 40 to calculate the T2 distribution 41. The conditions for inversion (discretization conditions, regularization conditions, etc.) are preferably fixed when performing lot-to-lot comparisons.

[0104] (S104) The calculation unit 20 sets boundary values ​​T2A(45) and T2B(46) on the T2 axis of the T2 distribution 41. For example, (S104a) the boundary values ​​T2A(45) and T2B(46) may be set by instrument calibration, or (S104b) the boundary values ​​T2A(45) and T2B(46) may be set by reading fixed values ​​held in the storage unit 21.

[0105] (S105) The calculation unit 20 divides the T2 distribution 41 into a first region 42, a second region 43, and a third region 44 using boundary values ​​T2A(45) and T2B(46). The boundary values ​​T2A(45) and T2B(46) may be values ​​stored in the storage unit 21, or values ​​set by device calibration.

[0106] (S106) The calculation unit 20 calculates the areas A1(50), A2(51), and A3(52) corresponding to each region. If necessary, the total area AT=A1+A2+A3 may be calculated and normalization may be performed.

[0107] (S107) The calculation unit 20 calculates a moisture state index 60 based on the areas A1 (50), A2 (51), and A3 (52). For example, J1 = A3 / (A1 + A2 + A3) and J2 = A3 / (A1 + A2) are calculated.

[0108] (S108) The calculation unit 20 compares the calculated indices J1 and J2 with a preset judgment criterion (for example, judgment thresholds J1TH and J2TH held in the storage unit 21) and makes a judgment on whether the processed shrimp product conforms to the quality standard. For example, it may be determined to pass if (J1 is within the standard) and (J2 is within the standard), and to fail if either is outside the standard. (S109) The output unit 22 outputs the indicators J1 and J2, as well as the standard determination result. The output format may be display, file recording, printing, or transmission by communication. If necessary, the area A1 (50), area A2 (51), area A3 (52), boundary values ​​T2A (45), T2B (46), and measurement conditions may be output as supplementary information.

[0109] (Optional) When evaluating multiple samples, steps (S101) to (S109) above may be repeated for each sample, and the statistical values ​​(mean, standard deviation, etc.) of the obtained moisture state index 60 may be calculated and output.

[0110] The above processing procedure allows for the quantitative determination of the moisture content of processed shrimp products as a moisture content index 60, without relying on a qualitative interpretation of the T2 distribution 41.

[0111] <Example System Configuration> The following describes an example of the configuration of a system for evaluating the moisture content of processed shrimp products according to the present invention, with reference to Figure 4.

[0112] The moisture state evaluation system of this embodiment comprises a low-field nuclear magnetic resonance (NMU) spectrometer 10, a calculation unit 20, a storage unit 21, and an output unit 22. These may be configured as an integrated device, or the NMU spectrometer 10 may be configured by communicating with an external computer 30.

[0113] As described above, the low-field nuclear magnetic resonance spectrometer 10 comprises a magnetic field generator 13, a sample holder 14, a transmitting / receiving system, a receiving unit 12, and a control unit 11, and acquires nuclear magnetic resonance signals from a sample to generate T2 relaxation data 40.

[0114] The calculation unit 20 calculates the T2 distribution 41, divides regions based on boundary values ​​T2A (45) and T2B (46), calculates areas A1 (50), A2 (51), and A3 (52), and calculates a moisture state index 60 based on the T2 relaxation data 40 input from the low-field nuclear magnetic resonance spectrometer 10. The calculation unit 20 includes a processor 31, such as a CPU, and performs calculation processing according to the program and parameters stored in the storage unit 21.

[0115] The memory unit 21 stores measurement conditions (sample temperature, sample mass (filling amount), pulse sequence, ECHO interval, number of ECHOs, repetition time, etc.), response conditions, and set values ​​such as boundary values ​​T2A (45) and T2B (46). It may also store acquired T2 relaxation data 40, calculated T2 distribution 41, area A1 (50), area A2 (51), area A3 (52), and moisture state index 60, as well as evaluation results.

[0116] The output unit 22 outputs the moisture status index 60 calculated by the calculation unit 20. The output unit 22 includes, for example, a display device, a printing device, a function to save to an external storage medium, or a communication interface, and can display the evaluation result as a numerical value, output it as a file, or transmit it to an external device.

[0117] The system of this embodiment may be configured to store the measurement conditions and boundary values ​​T2A(45) and T2B(46) as predetermined settings in the storage unit 21, and to automatically recall these settings at the start of measurement to execute the evaluation process. This suppresses variations in settings by the measurer and improves the reproducibility of the evaluation results.

[0118] Furthermore, the system of this embodiment may include a function to aggregate the evaluation results of multiple samples and calculate and output statistical quantities such as the average value and standard deviation on a lot-by-lot basis. In addition, traceability may be ensured by recording and outputting supplementary information such as the measurement date and time, operator, and device identification information.

[0119] <Examples of programs and processing procedures> In this embodiment, the program according to the present invention is executed by a computer 30 to calculate the T2 distribution 41, divide the region, calculate areas A1 (50), A2 (51), and A3 (52), and calculate and output the moisture state index 60 based on the T2 relaxation data 40 acquired by the low-field nuclear magnetic resonance spectrometer 10.

[0120] The computer 30 includes a processor 31, a memory 32, and an input / output interface 33. The memory 32 stores the program and setting information such as measurement conditions, boundary values ​​T2A(45), T2B(46), and reflection conditions. The input / output interface 33 is capable of performing input / output processing, including communication with the low-field nuclear magnetic resonance apparatus 10, input / output to an external storage medium, and output to a display device or the like.

[0121] Hereinafter, an example of the processing procedure of the program according to the present invention will be described with reference to Figure 5.

[0122] (S201) The computer 30 acquires or receives T2 relaxation data 40 from the low-field nuclear magnetic resonance spectrometer 10 and stores it in the storage unit 21 or memory 32. If necessary, it also acquires supplementary information such as measurement conditions and boundary values ​​T2A(45) and T2B(46).

[0123] (S202) The computer 30 performs a re-processing operation on the stored T2 relaxation data 40 and calculates the T2 distribution 41. The re-processing operation is performed according to the set re-processing conditions (discretization conditions, regularization conditions, number of iterations, etc.).

[0124] (S203) The computer 30 divides the T2 distribution 41 into a first region 42, a second region 43, and a third region 44 based on the boundary values ​​T2A (45) and T2B (46).

[0125] (S204) The computer 30 calculates the areas A1 (50), A2 (51), and A3 (52) corresponding to each region, and calculates the total area AT = A1 + A2 + A3 as needed.

[0126] (S205) The computer 30 calculates the moisture content index 60 based on the areas A1 (50), A2 (51), and A3 (52). For example, it calculates J1 = A3 / (A1 + A2 + A3) and J2 = A3 / (A1 + A2).

[0127] (S206) The computer 30 compares the calculated indicators J1 and J2 with a preset judgment criterion (for example, judgment thresholds J1TH and J2TH held in the storage unit 21 or memory 32) and makes a judgment on whether the processed shrimp product conforms to the quality standard. For example, it may be judged as passing if (J1 is within the standard) and (J2 is within the standard), and as failing if either is outside the standard. (S207) The computer 30 outputs the indicators J1 and J2, as well as the standard determination result. The output may be in the form of display, file recording, printing, or transmission by communication. If necessary, the computer may also output additional information such as area A1 (50), area A2 (51), area A3 (52), boundary values ​​T2A (45), T2B (46), measurement conditions, and calculation date and time.

[0128] (S208) When evaluating multiple samples or multiple lots, the computer 30 may repeat the processes described in (S201) to (S207) and calculate and output statistical values ​​(mean, standard deviation, etc.) of the obtained moisture state index 60.

[0129] The above processing procedure allows the computer 30 to perform the calculation of the T2 distribution 41 and the calculation and output of the moisture status index 60 in a single sequence, contributing to the acceleration and labor saving of evaluation work.

[0130] <In-process judgment using threshold table> In this embodiment, a determination threshold table (hereinafter also referred to as the "threshold table") used for in-process determination can be set based on boundary values ​​T2A(45), T2B(46) and moisture state indices J1, J2. The threshold table includes, for example, (1) boundary values ​​T2A(45), T2B(46), (2) determination thresholds J1TH and J2TH, and (3) a group of thresholds for grade classification as needed, and is held in the storage unit 21 or memory 32.

[0131] The judgment criteria (threshold) may be set based on the distribution of indicators J1 and J2 of multiple samples that have been previously confirmed to be acceptable. For example, methods include determining an acceptable range using the mean and standard deviation, or setting a predetermined percentile value of the group of acceptable products as the threshold. Threshold tables can be established based on a small number of calibration samples, without requiring the training of a regression or machine learning model that requires a large number of representative samples. For example, a predetermined number of conforming and non-conforming products are selected (e.g., 3 or more samples each, preferably 3 to 10 samples), and the judgment thresholds J1TH and J2TH are determined based on the distribution or statistics (mean, quantile, etc.) of J1 and J2 calculated from these samples.

[0132] The calculation unit 20 or computer 30 compares J1 and J2 of the sample obtained during the process with the judgment thresholds included in the threshold table and performs a specification judgment (pass / fail) or grade classification. For example, if both indicator J1 and indicator J2 are within the judgment criteria (acceptable range), it is judged as pass; otherwise, it is judged as fail. When performing grade classification, the thresholds for J1 and / or J2 may be set to multiple levels.

[0133] Furthermore, because it does not depend on a learning model, the indicators (J1, J2) and thresholds used for judgment can be presented directly within the process. Even if unexpected outliers occur during operation, it is easy to track where the changes occurred—in the measurement conditions, response conditions, boundary values, or area components—making it easy to identify the cause and correct the problem. This embodiment is characterized by the fact that it does not require the estimation of regression equations or inference of trained models for content estimation, and that the quality gate can be constructed by a simple comparison process based on a threshold table. Therefore, it is possible to make a quick determination within the process without requiring a training process for a large number of samples, and the basis for the determination (J1, J2, and threshold) can be presented in an explainable form. Furthermore, by using two indicators, J1 and J2, for the determination, and configuring the specification determination based on whether both meet the determination threshold, it is possible to perform normalization of the overall contribution and evaluation of bias between regions on separate axes, thereby increasing the robustness of the determination compared to cases that rely on a single indicator.

[0134] The threshold table may be set or updated in accordance with individual differences in equipment, changes in measurement conditions, significant fluctuations in raw material lots, etc. From the viewpoint of minimizing the update frequency and ensuring in-process operability, the boundary values ​​T2A(45), T2B(46) and reaction conditions may be fixed, and only the judgment thresholds J1TH and J2TH may be updated.

[0135] <Variation> The above embodiments are examples of the present invention, and the present invention is not limited thereto. Examples of modifications are given below.

[0136] (Variations of the evaluation target) The evaluation target shrimp products can be applied regardless of their form, such as peeled, with shell, cooked or uncooked, seasoned or not, frozen or not frozen. Furthermore, it may be applied not only to shrimp products but also to other processed seafood products (fish, shellfish, fish roe, etc.) that require similar moisture content evaluation.

[0137] (Modifications of the number of regions and division method) The T2 distribution 41 is shown as being divided into at least a first region 42, a second region 43, and a third region 44, but it may be divided into more than three regions depending on the evaluation purpose. In addition, the boundary values ​​T2A (45) and T2B (46) may be fixed values, or they may be automatically set or corrected based on the shape of the T2 distribution 41 (peak position or trough position).

[0138] (Variations of area calculation and preprocessing) When calculating areas A1(50), A2(51), and A3(52), the numerical integration method is not limited to the trapezoidal rule; a simple sum of discrete values, integration using spline interpolation, etc., may also be used. Furthermore, preprocessing such as smoothing, baseline correction, noise reduction, and outlier removal may be applied to the T2 distribution 41.

[0139] (Variations of moisture state index) As examples of moisture state index 60, A3 / (A1+A2+A3) and A3 / (A1+A2) have been given, but the index is not limited to these. For example, other ratio indices such as A1 / (A1+A2+A3), A2 / (A1+A2+A3), A2 / (A1+A3), or indices obtained by linear combination, logarithmic transformation, or standardization of areas A1(50), A2(51), and A3(52) may be used.

[0140] (Variations of measurement conditions and measurement method) Although the CPMG method is given as an example of measurement conditions, other pulse sequences that can acquire T2 relaxation data 40 may be used. In addition, measurement conditions such as sample temperature, sample mass (filling amount), ECHO interval, ECHO number, and fixed elements such as reaction conditions may be appropriately selected according to the evaluation purpose, and it is not necessary to fix all of them.

[0141] (Variations in system configuration) The calculation unit 20 may be built into the low-field nuclear magnetic resonance apparatus 10, or it may be configured as an external computer 30. Furthermore, the output unit 22 is not limited to a display device, but may also transmit to a cloud server, store in a database, or provide output in cooperation with other systems.

[0142] (Variations in program delivery format) The program of the present invention may be provided by recording it on a recording medium, or by distributing it via a communication line. The program may also be provided by dividing it into multiple modules, or by incorporating it as an additional function into existing analysis software.

[0143] Various modifications are possible without departing from the spirit of the present invention, including the modifications described above. [Explanation of Symbols]

[0144] 10 Low-field nuclear magnetic resonance apparatus 11 Control Unit 12 Receiver 13 Magnetic field generation section 14. Sample holder 20 Arithmetic section 21 Memory section 22 Output section 30 Computers 31 processors 32 memory 33 Input / Output Interfaces 40 T2 mitigation data 41 T2 distribution 42 First area 43 Second area 44 Third area 45 Boundary value T2A 46 Boundary value T2B 50 Area A1 51 Area A2 52 Area A3 60 Moisture Status Index

Claims

1. A method for evaluating the moisture content of processed shrimp products, T2 relaxation data was obtained for the shrimp processed product sample using a low-field nuclear magnetic resonance spectrometer. The T2 distribution is calculated by applying a counter-processing to the aforementioned T2 relaxation data. Boundary values ​​T2A and T2B (T2A < T2B) are set on the T2 axis of the T2 distribution. Based on the boundary values ​​T2A and T2B, the T2 distribution is divided into a first region (T2 < T2A), a second region (T2A ≤ T2 < T2B), and a third region (T2B ≤ T2). The areas A1, A2, and A3 corresponding to the first, second, and third regions are calculated, Based on the areas A1, A2, and A3, the indices J1 = A3 / (A1 + A2 + A3) and J2 = A3 / (A1 + A2) are calculated. If indicator J1 satisfies a predetermined judgment criterion and indicator J2 satisfies a predetermined judgment criterion, the processed shrimp product is judged to conform to the quality standard; if indicator J1 does not satisfy the judgment criterion, or if indicator J2 does not satisfy the judgment criterion, the processed shrimp product is judged to not conform to the quality standard. A method characterized by outputting the aforementioned indicators J1 and J2, as well as the standard determination result.

2. The method according to claim 1, characterized in that the judgment criteria include an acceptable range for each of index J1 and index J2 that is greater than or equal to the lower limit and less than or equal to the upper limit.

3. The method according to 1 or 2, characterized in that the boundary value T2A and the boundary value T2B are fixed values ​​stored in the storage unit or values ​​set by device calibration.

4. The method according to claim 1, characterized in that the acquisition of the T2 relaxation data is performed with fixed measurement conditions including at least one of the sample temperature, sample mass, pulse sequence, ECHO interval, and number of ECHOs, and the calculation of the T2 distribution is performed with fixed reflection conditions.

5. A system for evaluating the moisture content of processed shrimp products, Low-field nuclear magnetic resonance apparatus, The system comprises a calculation unit, the calculation unit calculates the T2 distribution by applying reflection processing to the T2 relaxation data acquired by the low-field nuclear magnetic resonance spectrometer, Boundary values ​​T2A and T2B (T2A < T2B) are set on the T2 axis of the T2 distribution, and the T2 distribution is divided into a first region (T2 < T2A), a second region (T2A ≤ T2 < T2B), and a third region (T2B ≤ T2) based on the boundary values ​​T2A and T2B. The areas A1, A2, and A3 corresponding to the first, second, and third regions are calculated, Based on the areas A1, A2, and A3, the indices J1 = A3 / (A1 + A2 + A3) and J2 = A3 / (A1 + A2) are calculated. If indicator J1 satisfies a preset criterion and indicator J2 satisfies a preset criterion, the product is judged to conform to the standard; otherwise, it is judged to be non-conforming to the standard, and indicators J1 and J2, as well as the standard judgment result, are output. A system characterized by the following features.

6. The computer calculates the T2 distribution by applying a counter-processing to the T2 relaxation data. Boundary values ​​T2A and T2B (T2A < T2B) are set on the T2 axis of the T2 distribution, and the T2 distribution is divided into a first region (T2 < T2A), a second region (T2A ≤ T2 < T2B), and a third region (T2B ≤ T2) based on the boundary values ​​T2A and T2B. The areas A1, A2, and A3 corresponding to the first, second, and third regions are calculated, Based on the areas A1, A2, and A3, the indices J1 = A3 / (A1 + A2 + A3) and J2 = A3 / (A1 + A2) are calculated. If indicator J1 satisfies the predetermined criteria and indicator J2 satisfies the predetermined criteria, the product is judged to conform to the standard; otherwise, it is judged to be non-conforming to the standard. A program that outputs the aforementioned indicators J1 and J2, as well as the standard determination result.

Images