Estimation device and estimation method

Abstract

To stabilize the accuracy of freshness estimation regardless of the food's photographic environment. [Solution] The estimation device 2 of this disclosure includes an image acquisition unit 101 that acquires a first image of food under a first lighting condition and a second image of food under a second lighting condition, an image generation unit 102 that generates a difference image by subtracting the pixel values ​​between the first image and the second image, and an estimation unit 104 that outputs an estimation result that estimates the freshness of the food by inputting the difference image into a learning model.

Description

【Technical Field】 【0001】 One aspect of the present disclosure relates to an estimation device and an estimation method. 【Background Art】 【0002】 Conventionally, a device for determining the freshness of food based on an image of the food has been used (see Patent Document 1 and Patent Document 2 below). The above device predicts the freshness of food from image statistics such as the average value of luminance and the standard deviation. 【Prior Art Documents】 【Patent Documents】 【0003】 【Patent Document 1】 Japanese Patent Application Laid-Open No. 2016-851,17 【Patent Document 2】 Japanese Patent Application Laid-Open No. 2023-130,969 【Summary of the Invention】 【Problems to be Solved by the Invention】 【0004】 In the above conventional device, the environment at the time of photographing the food may affect the determination value of freshness. That is, the estimation accuracy of freshness tends to depend on the photographing environment. 【0005】 Therefore, an object of the present disclosure is to provide an estimation device and an estimation method capable of stabilizing the estimation accuracy of freshness regardless of the photographing environment of food. 【Means for Solving the Problems】 【0006】 The estimation device of the present disclosure includes an image acquisition unit that acquires a first image obtained by imaging food in a first illumination state and a second image obtained by imaging food in a second illumination state, an image generation unit that generates a difference image by differentiating pixel values between the first image and the second image, and an estimation unit that outputs an estimation result obtained by estimating the freshness of the food by inputting the difference image into a learning model. 【0007】 The estimation method of this disclosure comprises: an image acquisition step in which an estimation device acquires a first image of food under a first lighting condition and a second image of food under a second lighting condition; an image generation step in which the estimation device generates a difference image by subtracting pixel values ​​between the first image and the second image; and an estimation step in which the estimation device outputs an estimation result that estimates the freshness of the food by inputting the difference image into a learning model. [Effects of the Invention] 【0008】 According to this disclosure, the accuracy of freshness estimation can be stabilized regardless of the food's photographic environment. [Brief explanation of the drawing] 【0009】 [Figure 1] Figure 1 is a block diagram showing the configuration of the estimation system of this disclosure. [Figure 2] Figure 2 shows an example of differential image data generated by the image generation unit 102 in Figure 1. [Figure 3] Figure 3 shows the first images of coffee beans with multiple roasting levels that are to be estimated by the estimation device 2 in Figure 1. [Figure 4] Figure 4 shows the directory structure of the memory area for storing training image data in the image storage unit 106 of Figure 1. [Figure 5] Figure 5 is a flowchart showing the procedure for generating a dataset using estimation system 1. [Figure 6] Figure 6 is a flowchart showing the machine learning procedure using estimation system 1. [Figure 7] Figure 7 is a flowchart showing the procedure for the estimation process performed by estimation system 1. [Figure 8] Figure 8 shows an example of the hardware configuration of the estimation device 2 according to one embodiment of the present disclosure. [Modes for carrying out the invention] 【0010】 Embodiments of this disclosure will be described with reference to the attached drawings. Where possible, the same parts will be denoted by the same reference numerals, and redundant descriptions will be omitted. 【0011】 Figure 1 is a diagram showing the device configuration of the estimation system according to this embodiment. The estimation system 1 shown in Figure 1 includes terminals 10A and 10B, and estimation device 2, which are configured to communicate with each other via a network including a wireless communication network and a fixed communication network. The estimation system 1 is a data processing device that estimates the freshness of food using images of food acquired by terminal 10A. 【0012】 The food targeted by Estimation System 1 is a tangible substance consumed by humans or other organisms, and is a processed product of naturally occurring plants and animals. For example, coffee beans are used as the food targeted by Estimation System 1 in this embodiment. 【0013】 Terminal 10A is a device used by a user who wants to obtain an estimate of the freshness of food. Terminal 10B is a device used by a user who configures the estimation system 1. Terminal 10B may be the same device as terminal 10A. Terminals 10A and 10B are, for example, personal computers, smartphones, tablet terminals, feature phones, etc., and are devices equipped with an imaging device such as a CMOS (Complementary Metal Oxide Semiconductor) camera and an illumination device (light source) that illuminates the object being imaged when imaging is performed by the imaging device. Terminals 10A and 10B may also be control circuits integrated with food processing equipment (coffee maker, coffee grinder, etc.). Although only one terminal 10A and one terminal 10B are shown in Figure 1, the estimation system 1 may include any number of terminals 10A and terminals 10B, two or more. 【0014】 Terminal 10A is a device used by a user who attempts to obtain freshness information regarding the freshness of coffee beans F1, which are food. Terminal 10B is a device used by a user who prepares coffee beans F2, which are food with known freshness, and attempts to machine-learn a learning model used for freshness estimation. 【0015】 The estimation device 2 is a device that estimates the freshness of coffee beans using a learning model. The learning model used by the estimation device 2 is a neural network model (e.g., ResNet described in Internet URL: “https: / / arxiv.org / abs / 1512.03385”) that can output freshness information representing freshness in response to the input of an image of coffee beans, and is a model including an input layer, a hidden layer, and an output layer. 【0016】 In this embodiment, the estimation device 2 enables freshness estimation processing using a learning model. This learning model may be stored in the estimation device 2, or may be stored in another device connected to the estimation device 2 via a network and configured such that information can be exchanged between the estimation device 2 and the other device. Although only one estimation device 2 is illustrated in FIG. 1, the estimation system 1 may include a plurality of estimation devices 2. 【0017】 The estimation device 2 is configured to include, as functional components, an image acquisition unit 101, an image generation unit 102, a selection unit 103, an estimation unit 104, a learning unit 105, an image storage unit 106, and a learning model storage unit 107. Hereinafter, the functions of each functional unit of the estimation device 2 will be described in detail. 【0018】 The image acquisition unit 101 acquires target image data obtained by imaging the coffee bean F1, which is the object of freshness estimation, from the terminal 10A, and stores the acquired target image data in the image storage unit 106. That is, the image acquisition unit 101 acquires, as the target image data, a first image in which the coffee bean F1 is in the first illumination state by the terminal 10A and a second image in which the coffee bean F1 is in the second illumination state by the terminal 10A. The first illumination state and the second illumination state are states in which the illuminance at the timing of imaging food is different from each other. In this embodiment, the first illumination state is a state in which the lighting device is turned on, and the second illumination state is a state in which the lighting device is turned off (the same applies hereinafter). 【0019】 Here, the image acquisition unit 101 may passively acquire the target image data from the terminal 10A used by the user who intends to acquire the freshness information regarding the coffee bean F1, or may actively acquire the target image data. When the image acquisition unit 101 actively acquires the data, it may remotely control the imaging device and the lighting device at the time of imaging the coffee bean F1 in the terminal 10A, and control to transmit the target image data acquired in the terminal 10A according to the control to the estimation device 2. 【0020】 In addition, the image acquisition unit 101 acquires training image data, which is training data for machine learning of the learning model used for the estimation process, from the terminal 10B, and stores the acquired training image data in the image storage unit 106. Also in this case, the image acquisition unit 101 acquires, as the training image data, a first image in which the coffee bean F2 is in the first illumination state by the terminal 10B and a second image in which the coffee bean F2 is in the second illumination state by the terminal 10B. At this time, the image acquisition unit 101 may passively acquire the training image data from the terminal 10B used by the user who intends to set up the estimation system 1, or may actively acquire the training image data. When the image acquisition unit 101 actively acquires the data, it may remotely control the imaging device and the lighting device at the time of imaging the coffee bean F2 in the terminal 10B, and control to transmit the training image data acquired in the terminal 10B according to the control to the estimation device 2. 【0021】 The image generation unit 102 processes the target image data acquired by the image acquisition unit 101 to generate differential image data. Specifically, it generates a first partial image and a second partial image by cropping (cutting out) the same area containing the image of the coffee bean F1 from a first image and a second image, respectively, of the same coffee bean F1 captured by the terminal 10A at the same magnification. The image generation unit 102 then generates target differential image data by assigning the difference value obtained by differentiating the pixel values ​​of each pixel between the first partial image and the second partial image to each pixel. 【0022】 Furthermore, the image generation unit 102 processes the training image data acquired by the image acquisition unit 101 in the same manner as the generation of the target difference image data described above, and generates training difference image data based on the first partial image and the second partial image. In this case, it is preferable that the training image data that forms the basis of the training difference image data is captured at the same magnification as the target image data and cropped to the same pixel size. 【0023】 Figure 2 shows an example of differential image data generated by the image generation unit 102. The differential image data G3 is generated as the difference between the first image G1, which was captured with the illumination device turned on, and the second image G2, which was captured with the illumination device turned off. Generally, foods such as coffee beans tend to release oil onto their surface as they deteriorate and lose freshness. By acquiring differential image data, it is possible to obtain an image in which the oil on the surface of the food is emphasized by light reflection. 【0024】 Here, the estimation device 2 is capable of estimating the freshness of multiple types of food, namely coffee beans. Specifically, the estimation device 2 is capable of estimating the freshness of coffee beans processed to multiple degrees. Figure 3 shows an example of the first image of coffee beans processed to multiple degrees that are the target of estimation by the estimation device 2. Here, eight roasting levels of coffee beans are exemplified: "light roast," "cinnamon roast," "medium roast," "high roast," "city roast," "full city roast," "French roast," and "Italian roast." The training difference image data generated by the image generation unit 102 includes image data of the multiple types of coffee beans that are the target of estimation. 【0025】 Figure 4 shows the directory structure of the memory area for storing training image data in the image storage unit 106. As shown, the directory " / dataset" for storing training image data stores images of coffee beans F2 at different roasting levels, corresponding to the number of days elapsed since roasting, indicated by directories "day_1", "day_2", etc., such as "day 1", "day 2", etc., which indicate freshness. In this embodiment, images of coffee beans F2 from "day 1" to "30 days" are stored. Each directory "day_1", "day_2", etc., indicating the number of days elapsed, is configured with a directory "01_light" that stores multiple images of coffee beans F2 with a "light roast" level, a directory "02_cinnamon", etc., that stores multiple images of coffee beans F2 with a "cinnamon roast" level, and a directory "05_city" that stores multiple images of coffee beans F2 with a "city roast" level. The directory for storing images of coffee beans F2 at each roasting level will have two directories: “with_flash” for the first image and “without_flash” for the second image. The “with_flash” and “without_flash” directories will store the number of first and second images of multiple sets of coffee beans F2 required for machine learning. 【0026】 The learning unit 105 pre-trains a learning model corresponding to multiple roasting levels using the training difference image data generated by the image generation unit 102 based on the training image data stored in the image storage unit 106. Specifically, the learning unit 105 acquires training difference image data generated based on the training image data of coffee beans F2 of the same roasting level at multiple intervals of elapsed days. Then, the learning unit 105 tags the elapsed days on the training difference image data and generates a training dataset, a test dataset, and a validation dataset. Next, the learning unit 105 adjusts the parameters of the learning model by performing machine learning using these datasets and stores the generated learning model parameters in the learning model storage unit 107. Furthermore, the learning unit 105 repeats the above generation of the learning model using the training difference image data generated based on the training image data of coffee beans F2 of multiple roasting levels, and stores the parameters of the learning model corresponding to multiple roasting levels in the learning model storage unit 107. 【0027】 When the freshness estimation process for coffee beans F1 is executed, the selection unit 103 selects a learning model to be used for the estimation process based on the information indicating the degree of roasting of coffee beans F1 received from terminal 10A, and reads the parameters of the selected learning model from the learning model storage unit 107. For example, if information indicating a degree of roasting of "light roast" is received from terminal 10A, the selection unit 103 selects a learning model that has been trained using a dataset of coffee beans F2 with a degree of roasting of "light roast," and reads its parameters. 【0028】 When the freshness estimation process for coffee beans F1 is performed, the estimation unit 104 uses the learning model selected by the selection unit 103 to estimate the number of days elapsed since the roasting date of coffee beans F1, and outputs an estimation result indicating the estimated number of days elapsed. The output of the estimation result may be performed by transmitting it to an external device such as a terminal 10A, or by storing it in the data storage unit within the estimation device 2 so that it can be accessed from the outside. Specifically, the estimation unit 104 inputs the target difference image data generated by the image generation unit 102 into a learning model whose parameters have been read out by the selection unit 103, and obtains the output value output from the learning model accordingly as the estimation result. 【0029】 The procedure for generating a dataset using the estimation system 1 configured as described above, the procedure for machine learning using the estimation system 1, and the procedure for estimation processing using the estimation system 1 will be explained below. Figures 5, 6, and 7 are flowcharts illustrating the procedure for generating a dataset using the estimation system 1, the procedure for machine learning, and the procedure for estimation processing, respectively. 【0030】 The dataset generation process shown in Figure 5 is initiated, for example, by instructions received from the user of terminal 10B. First, the image acquisition unit 101 of the estimation device 2 repeatedly acquires training image data from terminal 10B, targeting coffee beans F2 of multiple roasting levels, in the form of combinations of a first image under a first lighting condition and an image under a second lighting condition (step S101). Subsequently, it is determined whether or not images have been taken for 30 days (step S102), and if images have not been taken for 30 days (step S102; No), the acquisition of training image data by the image acquisition unit 101 is repeated. 【0031】 On the other hand, if images have been taken for 30 days (Step S102; Yes), the image acquisition unit 101 organizes the training image data by roasting degree, number of days elapsed since roasting, and lighting conditions, and the organized training image data is stored in the image storage unit 106 in multiple directories (Step S103). With this, the preparation of the training dataset is complete (Step S104). 【0032】 The machine learning process shown in Figure 6 begins after the dataset generation process shown in Figure 5. First, the image generation unit 102 of the estimation device 2 reads the training dataset stored in the image storage unit 106 (step S201). Next, the image generation unit 102 crops the first and second images that constitute each training image data included in the dataset to generate the first partial image and the second partial image (step S202). Then, the image generation unit 102 generates training difference image data based on the first and second partial images, targeting the training image data (step S203). Subsequently, it is determined whether the training difference image data has been generated for all roasting levels and all elapsed days (step S204). If it has not been generated for all roasting levels and all elapsed days (step S204; No), the generation of training difference image data by the image generation unit 102 is repeated. This completes the preprocessed dataset (step S205). 【0033】 Next, the learning unit 105 of the estimation device 2 divides the completed preprocessed dataset into a training dataset, a test dataset, and a validation dataset (step S206). Finally, the learning unit 105 uses these datasets to generate multiple learning models capable of estimating the number of days elapsed for coffee beans at multiple roasting levels using machine learning, and the parameters of the generated learning models are stored in the learning model storage unit 107 (step S207). 【0034】 The estimation process shown in Figure 7 is initiated, for example, by instructions received from the user of terminal 10A. First, the user of terminal 10A inputs information indicating the degree of roasting of the coffee beans F1 to be estimated for freshness to the estimation device 2 (step S301). Next, the image acquisition unit 101 of the estimation device 2 acquires target image data from terminal 10A, consisting of a combination of a first image under a first lighting condition and a second image under a second lighting condition, with respect to the coffee beans F1 (step S301). Subsequently, the image generation unit 102 crops the first and second images included in the target image data to generate a first partial image and a second partial image (step S303). Then, the image generation unit 102 generates target difference image data based on the first partial image and the second partial image, with respect to the target image data (step S304). 【0035】 Next, the selection unit 103 selects a learning model corresponding to the roasting degree of coffee beans F1 based on the information indicating the roasting degree, and the parameters of the selected learning model are read from the learning model storage unit 107 (step S305). Then, the estimation unit 104 inputs the target difference image data into the learning model with the read parameters set (step S306). Finally, the estimation unit 104 outputs the output value of the learning model as the estimation result (step S307). 【0036】 Next, the effects of the acquisition system of this disclosure will be explained. According to the estimation system 1 of this disclosure, the number of days elapsed since roasting, which is the freshness of the coffee beans, is estimated by inputting a difference image obtained by differentiating a first image of coffee beans F1 taken under a first lighting condition and a second image taken of coffee beans F1 taken under a second lighting condition into a learning model. Since the difference image is less affected by the shooting environment, the accuracy of freshness estimation can be stabilized regardless of the shooting environment. 【0037】 Coffee beans lose their flavor as they lose freshness, so to enjoy delicious coffee, consumers need to either choose fresh beans or adjust their brewing method according to the freshness. Coffee bean freshness can be measured in two ways: the time elapsed since processing from coffee cherries to green beans, and the time elapsed since roasting. Generally, roasted coffee beans are more popular with consumers than green beans. Therefore, understanding the freshness of roasted coffee beans is important for consumers. 【0038】 One conventional method for estimating the freshness of roasted coffee beans involves observing the bloom (swell) during coffee extraction. Since fresh beans contain more carbon dioxide than older beans, it is possible to estimate freshness by the degree of bloom during extraction. However, while it is desirable to determine freshness before extraction, this method cannot judge freshness before extraction. Another conventional method involves measuring the aroma of coffee beans. This method utilizes the property that fresh beans emit a strong aroma, while beans that have lost their freshness have a diminished aroma. However, using this method requires specialized equipment such as a gas chromatograph, and it is difficult for non-experts to obtain and use such specialized and expensive equipment. 【0039】 On the other hand, in the estimation system 1 of this embodiment, the property that oil seeps to the surface of coffee beans as they deteriorate is utilized to obtain an image that emphasizes light reflection due to the oil on the surface of the coffee beans by taking the difference between an image taken with the lights on and an image taken with the lights off. This makes it possible to accurately estimate the freshness of the coffee beans (specifically, the number of days elapsed since roasting). 【0040】 Furthermore, according to this disclosure, the freshness of coffee beans can be easily estimated using only a terminal and data processing device, without the need for specialized equipment, allowing even non-experts to select fresh coffee beans. Also, since the freshness can be estimated before brewing, if the purchased coffee beans are not fresh enough, measures such as re-roasting or adjusting the brewing method can be taken, making it possible to always enjoy delicious coffee. Moreover, the estimation system 1 can be applied to the IoT (Internet of Things), for example, by adding a coffee bean freshness estimation function to a coffee pot, it becomes possible to automatically notify the user when the coffee beans have deteriorated. 【0041】 In the estimation device 2 of this disclosure, the image acquisition unit 101 acquires a first image of the coffee beans taken with the light source that illuminates the coffee beans turned on, and a second image of the coffee beans taken with the light source turned off. This makes it possible to generate a difference image that highlights the gloss on the surface of the coffee beans, and by using this difference image, the freshness of the coffee beans can be estimated with high accuracy. 【0042】 Furthermore, in the estimation device 2 of this disclosure, the estimation unit 104 outputs an estimation result using a learning model selected from among multiple learning models based on information representing the degree of processing of the coffee beans. In this case, the freshness can be estimated using a learning model that has estimation criteria corresponding to the degree of processing of the coffee beans, thereby further improving the accuracy of freshness estimation. 【0043】 Furthermore, in the estimation device 2 of this disclosure, multiple learning models are pre-trained using difference images acquired for coffee beans of multiple processing degrees, and the estimation unit 104 selects a learning model trained using difference images of coffee beans corresponding to the processing degree information. In this case, multiple learning models corresponding to each of the processing degrees of multiple coffee beans can be constructed, and by selecting a learning model from among them that corresponds to the processing degree of the target coffee bean, the accuracy of freshness estimation can be further improved. 【0044】 The acquisition device described herein has the following configuration. 【0045】 [1] An image acquisition unit that acquires a first image of food under a first lighting condition and a second image of the food under a second lighting condition. An image generation unit generates a difference image by subtracting pixel values ​​between a first image and a second image, An estimation unit that outputs an estimation result that estimates the freshness of the food by inputting the difference image into a learning model, An estimation device equipped with the following features. 【0046】 [2] The image acquisition unit acquires a first image of the food taken with the light source illuminating the food turned on, and a second image of the food taken with the light source turned off. The estimation device described in [1] above. 【0047】 [3] The estimation unit outputs the estimation result using a learning model selected from among multiple learning models based on the information representing the type of food. Estimation device as described in [1] or [2] above 【0048】 [4] The aforementioned multiple learning models have been pre-trained using the difference images acquired for multiple types of food, The estimation unit selects a machine learning model that uses the difference images of the food of the type corresponding to the information. The estimation device described in [3] above. 【0049】 [5] The aforementioned type is the degree of processing. The estimation device described in [3] or [4] above. 【0050】 The block diagram used in the description of the above embodiment shows functional units. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one device that is physically or logically coupled, or it may be realized using two or more physically or logically separated devices that are directly or indirectly connected (for example, using wired or wireless connections). A functional block may be realized by combining the above one device or the above multiple devices with software. 【0051】 Functions include, but are not limited to, judgment, decision, judgment, calculation, calculation, processing, derivation, investigation, exploration, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, assumption, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating (mapping), and assigning. For example, a functional block (configuration part) that enables transmission is called a transmitting unit or transmitter. As mentioned above, the method of implementation is not particularly limited. 【0052】 For example, the estimation device 2 in one embodiment of the present disclosure may function as a computer that performs the processing of the present disclosure. Figure 8 is a diagram showing an example of the hardware configuration of the estimation device 2 according to one embodiment of the present disclosure. The estimation device 2 described above may be physically configured as a computer device including a processor 1001, memory 1002, storage 1003, communication device 1004, input device 1005, output device 1006, bus 1007, etc. Note that the estimation device 2 may be configured as a computer device including at least one processor such as a CPU or GPU, may be configured as a computer device including multiple processors, or may be configured as including multiple computer devices. 【0053】 In the following explanation, the term "device" can be replaced with "circuit," "device," "unit," etc. The hardware configuration of estimated device 2 may include one or more of the devices shown in the figure, or it may be configured to omit some of the devices. 【0054】 Each function in the estimation device 2 is realized by loading predetermined software (programs) onto hardware such as the processor 1001 and memory 1002, which causes the processor 1001 to perform calculations, control communication by the communication device 1004, and control at least one of data reading and writing in the memory 1002 and storage 1003. 【0055】 The processor 1001 controls the entire computer, for example, by running the operating system. The processor 1001 may be composed of a central processing unit (CPU) that includes interfaces with peripheral devices, control units, arithmetic units, registers, etc. For example, the image acquisition unit 101, image generation unit 102, selection unit 103, estimation unit 104, learning unit 105, etc., described above may be implemented by the processor 1001. 【0056】 Furthermore, the processor 1001 reads programs (program code), software modules, data, etc., from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes accordingly. The program used is one that causes the computer to execute at least a part of the operations described in the above embodiment. For example, the image acquisition unit 101, image generation unit 102, selection unit 103, estimation unit 104, and learning unit 105 may be implemented by a control program stored in the memory 1002 and running on the processor 1001, and other functional blocks may be implemented similarly. The above-described processes have been explained as being executed by one processor 1001, but they may be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may be transmitted from a network via a telecommunications line. 【0057】 Memory 1002 is a computer-readable recording medium and may consist of at least one of the following: ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. Memory 1002 may also be called a register, cache, main memory, etc. Memory 1002 can store executable programs (program code), software modules, etc., for carrying out a method according to one embodiment of the present disclosure. 【0058】 Storage 1003 is a computer-readable recording medium and may consist of at least one of the following: an optical disc such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disc, a digital multipurpose disc, a Blu-ray® disc), a smart card, flash memory (e.g., a card, a stick, a key drive), a floppy® disk, a magnetic strip, etc. Storage 1003 may also be called an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, server, or other suitable medium including at least one of memory 1002 and storage 1003. 【0059】 The communication device 1004 is hardware (transceiver / receiver device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as a network device, network controller, network card, communication module, etc. The communication device 1004 may be configured to include high-frequency switches, duplexers, filters, frequency synthesizers, etc., in order to implement at least one of frequency division duplex (FDD) and time division duplex (TDD). For example, the image acquisition unit 101 and estimation unit 104 described above may be implemented by the communication device 1004. 【0060】 The input device 1005 is an input device that accepts input from an external source (e.g., a keyboard, mouse, microphone, switch, button, sensor, etc.). The output device 1006 is an output device that outputs to an external source (e.g., a display, speaker, LED lamp, etc.). The input device 1005 and the output device 1006 may be configured as an integrated unit (e.g., a touch panel). 【0061】 Furthermore, each device, such as the processor 1001 and memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or different buses may be configured for each device. 【0062】 Furthermore, the estimation device 2 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), or an FPGA (Field Programmable Gate Array), and some or all of each functional block may be realized by such hardware. For example, the processor 1001 may be implemented using at least one of these hardware components. 【0063】 Information notification is not limited to the embodiments described herein and may be carried out by other means. For example, information notification may be carried out by physical layer signaling (e.g., DCI (Downlink Control Information), UCI (Uplink Control Information)), upper layer signaling (e.g., RRC (Radio Resource Control) signaling, MAC (Medium Access Control) signaling, broadcast information (MIB (Master Information Block), SIB (System Information Block))), other signals, or combinations thereof. RRC signaling may also be called RRC messages, and may be, for example, RRC Connection Setup messages, RRC Connection Reconfiguration messages, etc. 【0064】 The processing procedures, sequences, flowcharts, etc., of each aspect / embodiment described herein may be reordered, provided they are consistent with each other. For example, the methods described herein present various step elements in an exemplary order and are not limited to that specific order. 【0065】 Input and output information may be stored in a specific location (e.g., memory) or managed using a management table. Input and output information may be overwritten, updated, or appended to. Output information may be deleted. Input information may be sent to other devices. 【0066】 The determination may be made by a value represented by 1 bit (0 or 1), by a boolean value (true or false), or by a numerical comparison (for example, a comparison with a predetermined value). 【0067】 Each aspect / embodiment described herein may be used individually, in combination, or switched between as needed during implementation. Furthermore, notification of specific information (e.g., notification that "X is") is not limited to explicit notification, but may also be implicit (e.g., by not providing such notification). 【0068】 Although the present disclosure has been described in detail above, it will be clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the intent and scope of the present disclosure as defined by the claims. Accordingly, the descriptions in the present disclosure are illustrative and not intended to be restrictive in any way. 【0069】 In other words, the estimation device 2 according to the modified version of this disclosure may operate to estimate the freshness of coffee beans F1, which are a blend of coffee beans with multiple degrees of processing (roasting). In this modified version, the estimation device 2 operates as follows: In the estimation process shown in Figure 7, in step S301, information on the coffee bean with the highest degree of roasting blended in coffee beans F1 is input. In step S303, the color difference (RGB value or HSV value) in the target image data is calculated, the coordinates or range of the darkest part are identified, and a partial image of a size that can be input into the learning model is cropped from those coordinates or range. With this modified version, even if the coffee beans to be estimated are a blend of multiple degrees of roasting and there is unevenness in the rate of deterioration of freshness, the freshness can be estimated by focusing on the darker roasted beans. 【0070】 Software should be broadly interpreted to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, execution threads, procedures, functions, and so on, whether they are called software, firmware, middleware, microcode, hardware description languages, or by any other name. 【0071】 Furthermore, software, instructions, information, etc., may be transmitted and received via a transmission medium. For example, if software is transmitted from a website, server, or other remote source using at least one of wired technologies (such as coaxial cable, fiber optic cable, twisted pair, or digital subscriber line (DSL)) and wireless technologies (such as infrared or microwave), then at least one of these wired and wireless technologies is included in the definition of a transmission medium. 【0072】 The information, signals, etc. described in this disclosure may be represented using any of the various different techniques. For example, the data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof. 【0073】 In addition, terms used in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings. For example, at least one of the channel and symbol may be a signal (signaling). Also, a signal may be a message. Furthermore, a component carrier (CC) may be called a carrier frequency, cell, frequency carrier, etc. 【0074】 Furthermore, the information, parameters, etc., described in this disclosure may be expressed using absolute values, relative values ​​from a given value, or other corresponding information. For example, wireless resources may be indicated by an index. 【0075】 The names used for the parameters described above are not restrictive in any way. Furthermore, formulas and other expressions using these parameters may differ from those expressly disclosed in this disclosure. Various channels (e.g., PUCCH, PDCCH, etc.) and information elements can be identified by any suitable name, and therefore, the various names assigned to these various channels and information elements are not restrictive in any way. 【0076】 In this disclosure, terms such as "Mobile Station (MS)," "user terminal," "User Equipment (UE)," and "terminal" may be used interchangeably. 【0077】 A mobile station may also be referred to by those skilled in the art as a subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, or some other appropriate term. 【0078】 As used in this disclosure, the terms “determining” and “determining” may encompass a wide variety of actions. “Determining” may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, searching, or inquiring (e.g., searching in a table, database, or other data structure), or ascertaining. “Determining” may also include, for example, receiving (e.g., receiving information), transmitting (e.g., sending information), inputting, outputting, or accessing (e.g., accessing data in memory). Furthermore, "judgment" and "decision" can include considering something as having been "judged" or "decided" after resolving, selecting, choosing, establishing, comparing, etc. In other words, "judgment" and "decision" can include considering something as having been "judged" or "decided" after some action. Also, "judgment (decision)" can be reinterpreted as "assuming," "expecting," or "considering." 【0079】 The terms “connected,” “coupled,” or any variation thereof, mean any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are “connected” or “coupled” with each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, “connection” may be reinterpreted as “access.” As used in this disclosure, two elements may be considered to be “connected” or “coupled” with each other using at least one of one or more wires, cables, and printed electrical connections, and, in some non-limiting and non-exclusive examples, electromagnetic energy having wavelengths in the radio frequency domain, microwave domain, and optical (both visible and invisible) domain. 【0080】 In this disclosure, the phrase "based on" does not mean "based solely on" unless otherwise specified. In other words, the phrase "based on" means both "based solely on" and "based at least on." 【0081】 Any reference to elements using designations such as “first,” “second,” etc., as used in this disclosure does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient way to distinguish between two or more elements. Accordingly, references to first and second elements do not imply that only two elements may be adopted, or that the first element must precede the second element in any way. 【0082】 Where the terms “include,” “including,” and their variations are used in this disclosure, these terms are intended to be inclusive, as is the term “comprising.” Furthermore, the term “or” as used in this disclosure is not intended to mean exclusive OR. 【0083】 In this disclosure, if articles are added by translation, such as a, an, and the in English, this disclosure may include the fact that the noun following these articles is plural. 【0084】 In this disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "combine" may be interpreted similarly to "different." [Explanation of symbols] 【0085】 1... Estimation system, 10A, 10B... Terminals, 2... Estimation device, 101... Image acquisition unit, 102... Image generation unit, 103... Selection unit, 104... Estimation unit, 105... Learning unit, 106... Image storage unit, 107... Learning model storage unit.

Claims

[Claim 1] An image acquisition unit that acquires a first image of food under a first lighting condition and a second image of the food under a second lighting condition, An image generation unit generates a difference image by subtracting pixel values ​​between a first image and a second image, An estimation unit that outputs an estimation result that estimates the freshness of the food by inputting the difference image into a learning model, An estimation device equipped with the following features. [Claim 2] The image acquisition unit acquires a first image of the food taken with the light source illuminating the food turned on, and a second image of the food taken with the light source turned off. The estimation device according to claim 1. [Claim 3] The estimation unit outputs the estimation result using a learning model selected from among multiple learning models based on the information representing the type of food. The estimation device according to claim 1. [Claim 4] The aforementioned multiple learning models have been pre-trained using the difference images acquired for multiple types of food, The estimation unit selects a machine learning model that uses the difference images of the food of the type corresponding to the information. The estimation device according to claim 3. [Claim 5] The aforementioned type is the degree of processing. The estimation device according to claim 3. [Claim 6] The estimation device performs an image acquisition step of acquiring a first image of the food under a first illumination state and a second image of the food under a second illumination state, The estimation device performs an image generation step of generating a difference image by differentiating pixel values ​​between a first image and a second image, The estimation step involves the estimation device inputting the difference image into a learning model to output an estimation result that estimates the freshness of the food, An estimation method comprising the following:

Images