TWI551861B - System and method for estimating carbohydrate content - Google Patents

Description

System and method for estimating carbohydrate content

The present invention relates to a system and method for estimating nutrient content in an analyte, and more particularly to a system and method for estimating carbohydrate content.

With the development of science and technology, the lifestyle of modern people has also undergone considerable changes. Although medical technology is very developed, chronic diseases such as asthma, cancer, diabetes, and cardiovascular diseases are prevalent due to changes in the environment and eating habits. Especially in recent years, among developing countries and newly industrialized countries, diabetes incidence and prevalence are rapidly increasing. There are currently about 190 million people with diabetes worldwide, but according to World Health Organization estimates, there will be 330 million people with diabetes worldwide in 2025, and most of them are type 2 diabetes. In Taiwan, diabetes has been the fifth most common cause of death since 1987, and in 2011, diabetes rose to fourth place, the fastest-growing disease in the past two decades.

From the point of view of social costs, the social burden caused by the prevalence of diabetes cannot be underestimated. According to statistics, the number of people with diabetes in the United States is about 28 million, accounting for 8.3% of the total population. In 2012, the cost of medical expenses for diabetes in the United States was $245 billion, of which $176 billion was direct medical cost and $69 billion was diabetes. Indirect costs of medical care (Source: American Diabetes Association (2012)). On the other hand, from 1997 to 2009, the number of diabetic patients in Taiwan increased from 538,000 to 1.23 million, growing more than 2.3 times. In addition, the total medical expenses for diabetes patients have also increased year by year. In 2009, it exceeded 100 billion, accounting for 22% of the total total health insurance expenditure; that is, the Chinese government spends about 300 million yuan a day on diabetes-related medical expenses ( Source: Republic of China Diabetes Association (2011)).

For people with diabetes, controlling blood sugar is very important. Blood sugar is constantly changing due to factors such as diet, exercise, and mood. Since changes in blood sugar are not easily detectable, regular self-monitoring is necessary to ensure that blood glucose levels are controlled within the target range. In addition, by understanding the changes in blood glucose levels (such as tracking and recording the highs and lows of daily blood glucose values), we can more effectively assess the effectiveness of prescriptions such as diet, activities and drugs, and adjust the course of treatment in a timely manner.

However, patients with general diabetes lack the tools to manage the management of personal sugar intake, and it is not easy to control the daily diet. It also causes instability and indirectly increases the medical cost of the country. In view of this, there is a need to develop systems and methods for estimating carbohydrate content for the management of sugar intake by individuals.

The present invention provides a system and method for estimating carbohydrate content, which can be applied to confirm the content of carbohydrates in a test substance, thereby enabling a user (for example, a diabetic patient) to easily manage sugar intake.

The system for estimating carbohydrate content of the present invention comprises: an image data acquisition module for capturing image data of the object to be tested, the image data comprising light field data and identification data of the object to be tested; and a data processing comparison module, The image data acquisition module is coupled to the image data acquisition module, and the data processing comparison module obtains the identity of the object to be tested according to the identity identification data of the object to be tested, and estimates the image according to the light field data. a volume of the object to be tested; and a component integration module coupled to the data processing comparison module, wherein the component integration module obtains the carbohydrate conversion data of the object to be tested by using the identity of the object to be tested, and uses the above The carbohydrate conversion data of the analyte and the volume of the analyte to be tested are used to calculate the carbohydrate content of the analyte.

In an embodiment of the invention, the object identification data includes the two-dimensional/three-dimensional image data of the object to be tested and/or the spectral data of the object to be tested.

In an embodiment of the present invention, the data processing comparison module includes: an image pre-processing module that performs pre-processing on the light field data and the identification data of the object to be tested; and the feature comparison module is coupled to In the image pre-processing module, the feature comparison module obtains the identity of the object to be tested according to the identity identification data of the object to be tested, and the volume estimation module is coupled to the feature comparison module. And estimating the volume of the object to be tested based on the light field data.

In an embodiment of the invention, the image data capture module includes a light field camera, and the volume estimation module performs volume (V) estimation of the object to be tested by the following formula (1): V=fVol (:K, θ, D, H)............(1)

Where K represents the average cross-sectional area of the object to be tested; θ represents the angle between the gyroscope of the sensor of the light field camera; D represents the distance between the sensor and the object to be tested; and H represents the object to be tested. Height, H = cos θ ‧ D.

In an embodiment of the present invention, the image data capture module includes a light field camera, and the object to be tested is contained inside the tableware, and the volume estimation module performs the object to be tested by the following formula (2). Volume (Vol) estimation: Vol = ʃʃ [ z ₀ ( x , y ) - z ( x , y )] dxdy ............ (2)

Wherein, the X, Y, and Z directions are set with the opening edge portion of the tableware as a reference plane, and z ₀ (x, y) represents the distance from the light field camera to the bottom of the tableware in the Z direction, and z(x, y) represents The distance from the light field camera to the surface of the object to be tested in the Z direction.

In an embodiment of the present invention, the feature comparison module includes: an image comparison unit, obtaining the identity of the object to be tested according to the two-dimensional/three-dimensional image data of the object to be tested; and the spectral comparison unit according to the The spectral data of the measuring object obtains the identity of the object to be tested; and the data input unit for the operator to input the identity of the object to be tested.

In an embodiment of the invention, the operator inputs the identity of the object to be tested in a voice or text manner.

In an embodiment of the invention, performing shadowing in the volume estimation module The depth estimation program is used to establish a depth map of the object to be tested, and then combined with the features obtained by the spectral comparison unit to mark the image data to estimate the volume of the object to be tested.

In an embodiment of the invention, the system for estimating carbohydrate content further includes a remote device coupled to at least the component integration module to receive a result calculated by the component integration module.

The method for estimating the carbohydrate content of the present invention comprises the steps of: capturing image data of the object to be tested, the image data comprising light field data and identification data of the object to be tested; and performing identification data according to the identity of the object to be tested in the database Obtaining the identity of the object to be tested according to the comparison; estimating the volume of the object to be tested according to the light field data; and obtaining the carbohydrate conversion data of the object to be tested by using the identity of the object to be tested, and using the object to be tested The carbohydrate content of the analyte to be tested is calculated from the carbohydrate conversion data and the volume of the above-mentioned analyte.

In an embodiment of the invention, the object identification data includes the two-dimensional/three-dimensional image data of the object to be tested and/or the spectral data of the object to be tested.

In an embodiment of the invention, the two-dimensional/three-dimensional image data of the object to be tested is obtained by a light field camera.

In an embodiment of the present invention, the step of obtaining the identity of the object to be tested by comparing the identification data of the object to be tested in the database includes: first obtaining the image according to the two-dimensional/three-dimensional image data of the object to be tested. When the identity of the object to be tested is unable to obtain the identity of the object to be tested based on the two-dimensional/three-dimensional image data of the object to be tested, the identity of the object to be tested is obtained based on the spectral data of the object to be tested.

In an embodiment of the invention, it is not possible to When the three-dimensional image data or the spectral data acquires the identity of the object to be tested, the identity of the object to be tested is obtained by manual determination.

In an embodiment of the present invention, after the identity of the object to be tested is obtained by a manual determination method, the object identification data of the object to be tested is added to the database.

In an embodiment of the invention, the step of estimating the volume of the object to be tested according to the light field data comprises: performing an image depth estimation process to establish a depth map of the object to be tested.

In an embodiment of the present invention, in the case of capturing the image data of the object to be tested by using the light field camera, the step of estimating the volume of the object to be tested according to the light field data includes: performing the following formula (1) Estimation of volume (V) of the object: V = fVol (: K, θ, D, H) (1) where K represents the average cross-sectional area of the object to be tested; θ represents the angle between the gyro of the sensor of the light field camera and the object to be tested; D represents the distance between the sensor and the object to be tested; H represents the height of the object to be tested, H = cos θ ‧ D.

In an embodiment of the present invention, when the image data of the object to be tested is captured by the light field camera, and the object to be tested is contained inside the tableware, the step of estimating the volume of the object to be tested according to the light field data includes: The volume (Vol) estimation of the object to be tested is performed by the following formula (2): Vol = ʃʃ [ z ₀ ( x , y ) - z ( x , y )] dxdy ........... In the formula (2), the X, Y, and Z directions are set with the opening edge portion of the tableware as a reference plane, and z ₀ (x, y) represents the distance from the light field camera to the bottom of the tableware in the Z direction, z ( x, y) represents the distance from the light field camera to the surface of the object to be tested in the Z direction.

Based on the above, the system and method provided by the present invention can be transmitted through The non-contact and non-destructive manner confirms the type of the analyte and obtains data on its carbohydrate content, thereby realizing an interactive intelligent service system that enables users (for example, diabetic patients) to easily manage sugar intake.

The above described features and advantages of the invention will be apparent from the following description.

100‧‧‧System for estimating carbohydrate content

102‧‧‧Image data capture module

102a, 128‧‧‧ light field camera

102b‧‧·Spectrum camera

104‧‧‧Data processing comparison module

106‧‧‧Component Integration Module

108‧‧‧Remote device

110‧‧‧Image pre-processing module

112‧‧‧Feature comparison module

112a‧‧·Image comparison unit

112b‧‧‧spectral comparison unit

112c‧‧‧Data input unit

112d‧‧‧Voice pre-processing unit

112e‧‧‧Voice recognition unit

114‧‧‧ Volume Estimation Module

116‧‧‧Database

118‧‧‧Image data

120‧‧‧Substance ID number

122‧‧‧Study volume data

124, 132‧‧‧Test objects

126‧‧‧ sensor

130‧‧‧Tableware

D‧‧‧Distance

H‧‧‧ Height

Θ‧‧‧ angle

Steps S1, S2, S3, S4, S101, S102, S201, S202, S203, S204, S205, S301, S401‧‧

X, Y, Z‧‧ Direction

1 is a schematic diagram of a system for estimating carbohydrate content, in accordance with an embodiment of the invention.

2 is a schematic flow chart of a method of estimating carbohydrate content of the present invention.

3 is a system flow diagram for estimating carbohydrate content, in accordance with an embodiment of the invention.

FIG. 4 is a schematic diagram of estimating a volume of an object to be tested according to an embodiment of the invention.

FIG. 5 is a schematic diagram of estimating a volume of a sample to be tested according to another embodiment of the invention.

Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings.

1 is an estimated carbohydrate content according to an embodiment of the invention. Schematic diagram of the system of quantities.

Referring to FIG. 1 , the system 100 for estimating carbohydrate content in the embodiment includes an image data capturing module 102 , a data processing comparison module 104 , and a component integration module 106 .

The image data capturing module 102 is configured to capture image data of the object to be tested. The above-mentioned test object is, for example, a food containing only a single component, and may be a mixed food containing a plurality of components, but is not limited thereto. The above image data is, for example, light field data, two-dimensional/three-dimensional image data (ie, two-dimensional or three-dimensional image data), and spectral data. The two-dimensional/three-dimensional image data and the spectral data can be used as the identification data for the object to be tested (hereinafter referred to as "the object identification data"), and the spectral data is, for example, near-infrared spectrum data.

As can be seen from FIG. 1, in the embodiment, the image data capturing module 102 includes a light field camera 102a and a spectral camera 102b. The light field camera 102a can be used to capture the object to be tested and obtain the light field data and the two-dimensional/three-dimensional image data of the object to be tested, and the spectrum camera 102b can be used to obtain the spectral data of the object to be tested, for example, near-infrared spectrum data. However, the present invention is not limited to the above configuration. For example, a general digital camera and a light field camera may be combined to obtain a two-dimensional image of the object to be tested by a general digital camera, and the light field data is obtained by the light field camera as the identification data of the object to be tested. In addition, a camera with both light field data and spectral data capture can be used, so that the information needed to estimate the carbohydrate content can be obtained by a single exposure, that is, the test object can be obtained by one shot. The light field data, 2D/3D image data and spectral data are more conducive to improving the detection speed.

The data processing comparison module 104 is coupled to the image data capturing module 102. In the embodiment, the data processing comparison module 104 includes an image pre-processing module 110, a feature comparison module 112, and a volume estimation module 114.

The image pre-processing module 110 can perform pre-processing on the light field data of the object to be tested and the identification data of the object to be tested. The above pre-processing is, for example, performing array conversion and correction on the above data. After the correction, the image pre-processing module 110 can transmit the data to the feature comparison module 112.

The feature comparison module 112 is coupled to the image pre-processing module 110, and can receive the data processed by the image pre-processing module, and compare the data according to the object identification data to obtain the object to be tested. Identity.

The database 116 is, for example, a collection of data consisting of one or more sub-databases, including spectral, image or food name speech identification data of various foods and nutrients, and volume/weight of various foods and nutrients. Convert data, etc. In addition, each sub-database is, for example, a database built in the data storage device, and may also be a web-based database, and is not limited thereto. Specific examples of the sub-database may include a food nutrition database of the Health Department of the Executive Yuan, a food database provided by the National Institutes of Health, or a food database established by the industry itself.

As shown in FIG. 1 , in the embodiment, the feature comparison module 112 includes an image comparison unit 112a, a spectral comparison unit 112b, and a data input unit 112c. The image matching unit 112a can perform the comparison in the database 116 according to the two-dimensional/three-dimensional image data of the object to be tested to obtain the identity of the object to be tested. The spectral comparison unit 112b can be rooted According to the spectral data of the analyte, the identity of the object to be tested is obtained by comparison in the database 116. The data input unit 112c is available for the operator to input the identity of the above-mentioned object to be tested. The above operator is, for example, the user himself or a professional with food nutrition knowledge at the back end. In addition, the manner of inputting the identity of the object to be tested is not particularly limited, for example, the identity of the object to be tested is specified by voice or text input. The identity of the above-mentioned object to be tested is expressed, for example, by a serial number specified in the database, but is not limited thereto.

As shown in FIG. 1 , when the operator needs to input the identity of the object to be tested by voice, the data processing comparison module 104 may further include a voice pre-processing unit 112d and a voice recognition unit 112e, and the voice pre-processing unit 112d and the voice. The identification unit 112e is integrated into the feature comparison module 112, for example, but is not limited thereto. Specifically, the voice pre-processing unit 112d is coupled to the data input unit 112c and the voice recognition unit 112e, for example, for pre-processing the voice data of the object to be tested captured by the data input unit 112c (for example, for voice). The data is corrected and sampled, etc., and the processed data is transmitted to the speech recognition unit 112e. The voice recognition unit 112e is coupled to, for example, the voice pre-processing unit 112d and the data library 116, and can receive the voice data processed by the voice pre-processing unit 112d, and obtain the voice data according to the voice data in the database 116. The identity of the object to be tested.

The volume estimation module 114 is coupled to the feature comparison module 112, and the volume estimation module 114 can estimate the volume of the object to be tested according to the light field data. The volume estimation program of the object to be tested is, for example, analyzing the light field data by the image depth estimation program, and receiving the characteristic image data returned by the spectral comparison unit 112b for numerical conversion, thereby performing the object to be tested. Volume estimation. Volume estimation after volume estimation is completed The calculation module 114 can communicate the results to the component integration module 106.

The component integration module 106 is coupled to the data processing comparison module 104, and the component integration module 106 obtains the carbohydrate conversion data of the object to be tested by using the identity of the object to be tested, and uses the carbohydrate conversion data of the object to be tested and The volume of the analyte is measured to calculate the carbohydrate content of the analyte.

Additionally, the system 100 for estimating carbohydrate content may further include a remote device 108. The remote device 108 is coupled to the component integration module 106 and the image data capture module 102, and can receive the result calculated by the component integration module 106, and present the result to the user through image display or sound. In addition, the remote device 108 can also transmit the start signal to the image data capturing module 102 through the operation of the user to start the image capturing process. The remote device 108 is, for example, a smart phone, a tablet, a notebook computer or other portable device, but is not limited thereto, and the remote device 108 can also be designed as a stationary device according to requirements.

In addition, in terms of the hardware architecture of the system, the data processing comparison module 104 and the component integration module 106 are integrated into one or more processors, for example, and communicate with the database 116 and the image through the communication link (communication link). The data capture module 102 exchanges information to perform its functions, but is not limited thereto. Those of ordinary skill in the art will appreciate that the present invention can be implemented by the integration of one or more processors, libraries, computing systems, and the like.

2 is a schematic flow chart of a method for estimating carbohydrate content of the present invention, and FIG. 3 is a system flow diagram for estimating carbohydrate content according to an embodiment of the present invention. The system flow diagram shown in FIG. 3 can be performed, for example, by the embodiment of FIG. 1. The system for estimating carbohydrate content is performed, but is not limited thereto.

Hereinafter, the method for estimating the carbohydrate content according to the embodiment of Fig. 1 will be described with reference to Figs. 2 and 3 to explain in detail the method for estimating the carbohydrate content of the present invention.

First, step S1 is performed to extract image data of the object to be tested. For example, the remote device 108 sends a start signal to the image data capturing module 102, and the image data is captured by the object to be tested (step S101), and the image data 118 is obtained. The image data 118 includes light field data, two-dimensional/three-dimensional image data, and spectral data of the object to be tested. The above-described light field data and two-dimensional/three-dimensional image data are obtained by, for example, a light field camera. However, the present invention is not limited thereto, and other types of cameras may be used as long as necessary data can be acquired. The above spectral data is, for example, near-infrared spectroscopy data. Thereafter, for example, the image data 118 is transmitted to the image pre-processing module 110 for pre-processing (step S102). The pre-processing may include performing an array transformation, a color correction, and a calibration on the primary image data 118.

The pre-processed image data 118 is simultaneously transmitted to the feature comparison module 112 and the volume estimation module 114 for subsequent comparison analysis, but is not limited thereto, and the image data 118 may be sequentially transmitted to the feature ratio. The module 112 and the volume estimation module 114 are provided.

Next, in step S2, the identity of the object to be tested is obtained by performing comparison in the database according to the identification data of the object to be tested. The flow of step S2 will be described in detail below.

First, for example, transmitting pre-processed image data 118 to image ratio The unit 112a performs image comparison (step S201). The image comparison unit 112a may first perform image feature extraction on the two-dimensional image in the image data 118. Next, the image matching unit 112a may perform a PCA (Principal Component Analysis) dimensionality reduction process on the obtained image features, and then compare with the data sample features in the database 116. If there is a completely or closest data sample, then Confirming the identity of the object to be tested, for example, obtaining the identity number 120 of the object to be tested. However, the present invention is not limited thereto, and the image matching unit 112a may process and compare the three-dimensional image of the object to be tested according to actual needs, or compare the two-dimensional and three-dimensional images of the object to be tested. Obtain the identity of the object to be tested.

In the above process of image comparison, when the sample feature of the comparison is too low in similarity with the sample feature in the existing database, the identity of the object to be tested cannot be obtained according to the image comparison (step S201). Further, spectral alignment is required (step S202).

In addition, when the results of the image comparison contain more than two types of identity, there may be a need for spectral alignment. For example, for foods with similar appearance or color (such as French fries and cheese bars), it may not be possible to distinguish only by image comparison. In this case, further use of spectral features is needed to identify the object to be tested. .

On the other hand, when the analyte is a mixed food that requires further confirmation of the ingredients, spectral alignment may also be required to achieve more accurate results. For example, the results of the image comparison initially confirm that the test object is fried rice, but the fried rice may include protein-containing meat and carbohydrate-containing corn, green bean kernel, and rice. In this case, it is also necessary to carry out Spectral alignment to further confirm fried rice In the composition.

Specifically, the specific signal notification system may be sent by the image comparison unit 112a to transmit the hyperspectral image data acquired by the spectral camera 102b in the image data 118 to the spectral comparison unit 112b for spectral comparison (step S202). ). The above hyperspectral image is, for example, a near-infrared spectrum image of the object to be tested, but is not limited thereto. The spectral comparison unit 112b can first convert the hyperspectral image data into an N-dimensional space, and then perform convex optimization, and compare with the spectral feature weights of the samples in the database 116. If there is a data sample that is completely consistent or closest, it is confirmed that the identity of the object to be tested is obtained, and the identity number of the object to be tested is obtained 120.

When the identity of the object to be tested cannot be obtained by the above spectral comparison (step S202), the identity of the object to be tested can be further obtained by manual determination (step S203) (step S204). For example, when the identity of the object to be tested cannot be obtained, the specific signal notification system can be sent by the spectral comparison unit 112b, and the image data 118 is transmitted to the operator, and the operator assists in image recognition and classification, and then transmits the data. The input unit 112c specifies the identity of the object to be tested by voice or text input, thereby obtaining the object number 120 of the object to be tested. The above operator is, for example, a professional with food nutrition knowledge at the back end.

In addition, the identity of the object to be tested can also be specified through user assistance. For example, when the identity of the object to be tested cannot be obtained, the specific signal is sent to the remote device 108 by the spectral comparison unit 112b to notify the user at the front end to input the auxiliary feature data. The above auxiliary feature data is, for example, a food name, a shape, a face Information such as color, smell or composition that can be identified by professionals. After the user returns the characteristics of the object to be tested through voice or text input, the system can notify the operator at the back end to specify the identity of the object to be tested. At this time, in addition to the image data, the operator can also refer to the auxiliary feature data input by the user to specify the identity of the object to be tested, thereby giving the object number 120 of the object to be tested. Alternatively, the user can specify the identity of the object to be tested by voice or text, thereby obtaining the object number 120 of the object to be tested.

Further, after the identity of the object to be tested is obtained by the manual determination method, the identity identification data as the object to be tested can be further added to the database (step S205). For example, new identification features of the same food may be added to the database 116, or identification features of new types of food may be created to increase the success rate of identification in the future.

When the above step S2 is performed, step S3 may be simultaneously performed to estimate the volume of the object to be tested based on the light field data. However, those skilled in the art should understand that the order of steps S2 and S3 is not particularly limited as long as the volume of the object to be tested and the identity data can be obtained smoothly.

In step S3, light field data analysis is performed (step S301) to obtain object volume data 122. Specifically, for example, the light field data that has undergone color correction and correction processing in the image pre-processing module 110 can be transmitted to the volume estimation module 114, and the image depth estimation program is performed in the volume estimation module 114 to establish The depth map of the object is combined with the feature obtained by the spectral comparison unit to mark the image data for volume estimation, thereby obtaining the volume data 122 to be tested.

FIG. 4 is a schematic diagram of estimating a volume of an object to be tested according to an embodiment of the invention. Referring to FIG. 4, when volume estimation is to be performed by the sensor 126 of the light field camera, the volume estimation module 114 performs volume (V) of the object to be tested 124 by, for example, the following formula (1). Estimate: V=fVol(:K, θ, D, H)............(1)

Where K represents the average cross-sectional area of the object to be tested 124 (the area filled with hatching in FIG. 4); θ represents the angle obtained by the gyroscope of the sensor 126, and D represents the sensor. The distance between 126 and the object to be tested 124; H indicates the height of the object to be tested 124, H = cos θ ‧ D.

In addition, the average cross-sectional area K is obtained by, for example, numerically converting the characteristic image data returned by the spectral comparison unit; θ, D are measured by, for example, a sensor of the light field camera, and H is determined by θ, D. Seek.

FIG. 5 is a schematic diagram of estimating the volume of the object to be tested according to another embodiment of the present invention, and showing a volume estimation of the object to be tested 132 contained inside the dish 130 by using the light field camera 128.

Referring to FIG. 5, when estimating the volume of the object to be tested 132 contained in the tableware 130, the volume (Vol) estimation of the object to be tested 132 can be performed by using the following formula (2): Vol = ʃʃ [ z ₀ ( x , y )- z ( x , y )] dxdy (2) wherein, as shown in FIG. 5, the opening edge portion of the tableware 130 is set as a reference plane. In the X, Y, and Z directions, z ₀ (x, y) may represent the distance (background value) from the light field camera 128 to the bottom of the tableware 130 in the Z direction, and is expressed in the Z direction by z(x, y). The distance from the light field camera 128 to the surface of the object to be tested 132 (foreground value). Those of ordinary skill in the art will appreciate that the background values and foreground values described above can be obtained by analytically converting the light field data intercepted by the light field camera 128. By subtracting the foreground value from the above background value and performing double integration, an estimated value of the volume of the object to be tested 132 can be obtained. Thereby, the parameter correction can be performed according to the type of the tableware, so that the volume of the object to be tested can be more accurately obtained. In addition, for different shapes, structures or sizes of tableware, different parameters can be designed first to facilitate rapid estimation of the volume of the object to be tested. According to the above manner, even if the object to be tested is placed in the tableware, the volume estimation can be performed efficiently.

After obtaining the volume of the object to be tested and the identity data, step S4 is performed, and the carbohydrate conversion data of the object to be tested is obtained by using the identity of the object to be tested, and the carbohydrate conversion data of the object to be tested and the object to be tested are used. The carbohydrate content of the above test substance was calculated from the volume. Specifically, for example, the object volume data 122 and the object identification number 120 obtained by the step S2 and the step S3 are transmitted to the component integration module 106 to perform a component integration operation (step S401). For example, after the volume/weight conversion rate of the object to be tested is obtained in the database 116 according to the identifier number 120 of the object to be tested, the weight of the object to be tested is calculated by combining the volume data 122 of the object to be tested, and then The analyte identification number 120 obtains the carbohydrate conversion rate of the analyte in the database 116, and calculates the carbohydrate content in the analyte by using the weight of the analyte and the carbohydrate conversion rate.

After obtaining the carbohydrate content information in the test object, the system outputs the result to the remote device 108, for example, and provides the result to the user as a reference by means of voice, text or image.

In summary, the system and method for estimating carbohydrate content provided by the present invention can confirm the type of the analyte and obtain the data of the carbohydrate content through a non-contact and non-destructive manner, thereby realizing Users (eg, diabetics) can easily perform interactive intelligence service systems for sugar intake management.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention, and any one of ordinary skill in the art can make some changes and refinements without departing from the spirit and scope of the present invention. The scope of the invention is defined by the scope of the appended claims.

S1, S2, S3, S4‧‧‧ steps

Claims (17)

A system for estimating carbohydrate content, comprising: an image data acquisition module, capturing an image data of a sample to be tested, the image data comprising a light field data and an object identification data; a data processing ratio The module is coupled to the image data capture module, and the data processing comparison module obtains the identity of the object to be tested according to the identity identification data of the object to be tested, and according to the The light field data is used to estimate the volume of the object to be tested, wherein after the identity of the object to be tested is obtained by manual determination, the identity identification data of the object to be tested is added to the database; and a component integration module is coupled The data processing comparison module, wherein the component integration module obtains the carbohydrate conversion data of the object to be tested by using the identity of the object to be tested, and uses the carbohydrate conversion data of the object to be tested and the object to be tested The carbohydrate content of the analyte was calculated from the volume. The system for estimating carbohydrate content according to claim 1, wherein the object identification data includes two-dimensional/three-dimensional image data of the object to be tested and/or spectral data of the object to be tested. The system for estimating the carbohydrate content according to claim 1, wherein the data processing comparison module comprises: an image pre-processing module, and the light field data and the object identification data of the object to be tested are performed. Pre-processing; a feature comparison module coupled to the image pre-processing module, the feature comparison mode The group obtains the identity of the object to be tested according to the identity identification data of the object to be tested, and a volume estimation module is coupled to the feature comparison module, and estimates the to be tested according to the light field data. The volume of the object. The system for estimating the carbohydrate content according to claim 3, wherein the image data acquisition module comprises a light field camera, and the volume estimation module performs the method by using the following formula (1). Estimation of volume (V) of the object: V = fVol (: K, θ, D, H) (1) where K represents the average cross-sectional area of the object to be tested ; θ represents the angle of the gyro of the sensor of the light field camera to the object to be tested; D represents the distance between the sensor and the object to be tested; H represents the height of the object to be tested, H =cosθ‧D. The system for estimating the carbohydrate content according to claim 3, wherein the image data acquisition module comprises a light field camera, the object to be tested is contained inside a tableware, and the volume estimation module is The volume (Vol) estimation of the object to be tested is performed by the following formula (2): Vol = ʃʃ [ z ₀ ( x , y ) - z ( x , y )] dxdy .......... . . . (2) wherein the X, Y, and Z directions are set with the opening edge of the table as a reference plane, and z ₀ (x, y) is represented by the light field camera to the bottom of the tableware in the Z direction. The distance, z(x, y) represents the distance from the light field camera to the surface of the object to be tested in the Z direction. The system for estimating carbohydrate content according to claim 3, wherein the feature comparison module comprises: an image comparison unit, which is obtained according to the two-dimensional/three-dimensional image data of the object to be tested. The identity of the object to be tested; a spectral comparison unit that obtains the identity of the object to be tested according to the spectral data of the object to be tested; and a data input unit for an operator to input the identity of the object to be tested. The system for estimating the carbohydrate content according to claim 6, wherein the operator inputs the identity of the object to be tested in a voice or text manner. The system for estimating the carbohydrate content according to claim 6, wherein the image depth estimation program is performed in the volume estimation module to establish a depth map of the object to be tested, and then combined with the spectrum comparison unit. The feature marks the image data to estimate the volume of the object to be tested. The system for estimating the carbohydrate content according to claim 1 further includes a remote device coupled to at least the component integration module to receive the result calculated by the component integration module. A method for estimating a carbohydrate content, comprising the steps of: capturing image data of a sample to be tested, the image data comprising a light field data and an identity identification data of the object to be tested; and identifying data according to the identity of the object to be tested The library performs the comparison to obtain the identity of the object to be tested, wherein after obtaining the identity of the object to be tested through manual determination, the identity identification data of the object to be tested is added to the database; and the light field data is used to estimate the a volume of the analyte; and obtaining the carbohydrate conversion data of the analyte by using the identity of the analyte, and using the carbohydrate conversion data of the analyte and the volume of the analyte Calculate the carbohydrate content of the test substance. The method for estimating a carbohydrate content according to claim 10, wherein the object identification data includes two-dimensional/three-dimensional image data of the object to be tested and/or spectral data of the object to be tested. The method for estimating a carbohydrate content according to claim 11, wherein the two-dimensional/three-dimensional image data of the object to be tested is obtained by a light field camera. The method for estimating the carbohydrate content according to claim 11, wherein the step of obtaining the identity of the object to be tested according to the identity identification data of the object to be tested comprises: first Obtaining the identity of the object to be tested by the two-dimensional/three-dimensional image data of the object, and obtaining the identity of the object to be tested according to the two-dimensional/three-dimensional image data of the object to be tested, obtaining the spectrum data of the object to be tested The identity of the object to be tested. The method for estimating the carbohydrate content according to claim 13 , wherein when the identity of the object to be tested cannot be obtained from the two-dimensional/three-dimensional image data or the spectral data of the object to be tested, the method is manually determined. The identity of the object to be tested. The method for estimating a carbohydrate content according to claim 10, wherein the estimating the volume of the analyte according to the light field data comprises: performing an image depth estimation procedure to establish a depth map of the object to be tested. The method for estimating the carbohydrate content according to claim 10, wherein, in the case of capturing the image data of the object to be tested by using a light field camera, estimating the volume of the object to be tested based on the light field data The step includes: estimating the volume (V) of the object to be tested by the following formula (1): V=fVol(:K, θ, D, H) (1) where K represents the average cross-sectional area of the object to be tested; θ represents one of the light field cameras The angle between the gyro of the sensor and the object to be tested; D is the distance between the sensor and the object to be tested; H is the height of the object to be tested, H = cos θ ‧ D. The method for estimating the carbohydrate content according to claim 10, wherein, in the case of capturing the image data of the object to be tested by using a light field camera, and the object to be tested is contained in a tableware, The step of estimating the volume of the object to be tested based on the light field data includes: estimating a volume (Vol) of the object to be tested by the following formula (2): Vol = ʃʃ[ z ₀ ( x , y )- z ( x , y )] dxdy ...... (2) where the X, Y and Z directions are set with the opening edge of the table as the reference plane, z ₀ (x, y ) represents the distance from the light field camera to the bottom of the tableware in the Z direction, and z(x, y) represents the distance from the light field camera to the surface of the object to be tested in the Z direction.