Patent Application
20210325343
The present invention relates to an information processing apparatus, an information processing method, a learned model generation method, and a program.
There is a technology for performing object identification based on odor of the object. For example, JP 2017-161300 discloses an odor identification system for identifying a passenger seated on a seat of a vehicle using a neural network based on a signal detected by an odor detection apparatus installed in a sheet.
However, the invention according to JP 2017-161300 has not identified in consideration of the conditions when the odor is measured.
In one aspect, the present invention aims to provide an information processing apparatus or the like capable of preferably identifying an object from the odor of an object.
An information processing apparatus according to one aspect comprises a first acquisition unit configured to acquire odor data measuring an odor of an object, a second acquisition unit configured to acquire an acquisition condition of the odor data, and an identification unit configured to identify the object from the odor data and the acquisition condition acquired by the first and second acquisition units based on the learned model that has learned the odor data and the acquisition condition of the object and the object corresponding to the odor data.
In one aspect, an object can be suitably identified from the odor of the object.
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
The information processing apparatus 1 is an information processing apparatus that can perform various information processing and send/receive information, such as server computer or personal computer. In the present embodiment, it is assumed that the information processing apparatus 1 is a server computer, and it will be read as server 1 in the following description for the sake of brevity. The server 1 is a server computer that analyzes and manages odor data of objects arbitrarily uploaded by users, and performs machine learning to learn the odor data and generates an identification model 141 to identify objects from the odor data. Then, the server 1 identifies the object using the generated identification model 141. For instance, if the object is human, the server 1 identifies the person based on the measured odor data of exhalation odor of the person. Note that person identification is only one example of the use of the system, and the system may be used to identify other objects.
The terminal 2 is a terminal apparatus used by users of this system, such as personal computer, tablet, or smartphone. The server 1 obtains odor data of objects from the odor sensor 3 via the terminal 2, performs identification by the identification model 141, and outputs the identification results to the terminal 2.
The odor sensor 3 is a sensor that measures the odor of an object, and converts gas molecules corresponding to odor components into numerical values that can be processed by a computer. For example, the odor sensor 3 is an odor sensor using the QCM (Quartz Crystal Microbalance) method, which converts odor components into frequency data by using the oscillation of a quartz crystal when gas molecules are adsorbed on the surface of the quartz crystal. The use of an odor sensor using the quartz crystal as the odor sensor 3 makes it possible to obtain odor data that is closer to the human sense of odor than that of semiconductor-type odor sensors that measure the resistance of semiconductors or biosensors that use FETs (field effect transistors). The odor sensor 3 is connected to the terminal 2 of the user, and the server 1 obtains the odor data measured by the odor sensor 3.
The odor sensor 3 itself may be equipped with a communication module so that the server 1 can obtain odor data directly from the odor sensor 3. The odor sensor 3 is not limited to the odor sensor with the quartz crystal, but may also be semiconductor type odor sensor, FET biosensor, or the like.
The auxiliary storage 14 is a non-volatile storage area such as large-capacity memory or hard disk, and stores the program P and other data necessary for the controller 11 to execute processing. Further, the auxiliary storage 14 stores the identification model 141, user DB 142, odor DB 143, domain DB 144, and learning DB 145. The identification model 141 is a model for object identification based on the odor data. The identification model 141 is, for example, a model related to neural network to be described later. The identification model 141 is assumed to be used as a program module that is part of artificial intelligence software. The user DB 142 is a database that stores information of each user who uses this system. Odor DB 143 is a database that stores odor data acquired from the odor sensors 3. The domain DB 144 is a database that stores information of domains (odor categories) to be described later. The learning DB 145 is a database that stores learned parameters obtained by learning odor data, such as weights that are set in the identification model 141 (neural network).
The auxiliary storage 14 may be an external storage connected to the server 1. The server 1 may be a multi-computer comprises multiple computers, or may be a virtual machine virtually built by software.
Further, in the present embodiment, the server 1 is not limited to the above configuration, but may include, for example, an input unit to receive operation input, a display unit to display images, etc. In addition, the server 1 may has a reading unit that reads and executes a portable storage medium 1 a such as CD (Compact Disk)-ROM or DVD (Digital Versatile Disc)-ROM, and reads and executes the program P from the portable storage medium 1a. Alternatively, the server 1 may read the program P from a semiconductor memory 1b.
The odor DB 143 includes an acquisition data ID column, an odor name column, an acquisition date and time column, an acquisition user column, an acquisition sensor column, a data column, and an acquisition condition column. The acquisition data ID column stores the acquisition data ID, which is an identifier of the odor data acquired from the odor sensor 3. The odor name column, the acquisition date and time column, the acquisition user column, the acquisition sensor column, the data column, and the acquisition condition column store the odor name (object name), the acquisition date and time, the user name from which the odor data was acquired, the name of the odor sensor 3 from which the odor data was acquired, the odor data, and the acquisition condition of the odor data, respectively, in correspondence with the acquisition data ID. The acquisition condition column stores, for example, the domain name and subdomain name to be described later, the cleaning time when the odor sensor 3 was cleaned before measurement, the suction time of the odor(gas), the residence time in the gas chamber in the odor sensor 3, the cleaning time of the odor sensor 3 after measurement, the location information of the place where the odor was measured, and the weather information of the place. Various information that defines the acquisition conditions of the odor data will be described in detail later.
The domain DB 144 includes a domain ID column, a creation date and time column, a creation user column, a domain name column, a subdomain name column, and an acquisition condition column. The domain ID column stores the domain ID, which is an identifier of a domain indicating category of odor. The creation date and time column, the creation user column, the domain name column, the subdomain name column, and the acquisition condition column store the date and time when the user registering (creating) the information of the domain in the system, the registered user name, the domain name, the subdomain name, and the measurement conditions (acquisition conditions) in the odor sensor 3, respectively, in correspondence with the domain ID. The acquisition conditions store, for example, pre-cleaning time before odor measurement, suction time, residence time, and post-cleaning time after odor measurement.
The learning DB 145 includes a learning ID column, a creation date and time column, a creation user column, a odor name column, a domain name column, a subdomain name column, and a data column. The learning ID column stores the learning ID, which is an identifier of the learned parameter (such as the weight of the neural network) obtained by learning the odor data. The creation date and time column, the creation user column, the odor name column, the domain name column, the subdomain name column, and the data column store the creation date and time when the learning (creation) of the identification model 141 was executed, user name of the user who performed the learning of the identification model 141, the odor name that was learned, the domain name and subdomain name indicating the acquisition conditions of learned odor data, and learned parameters, respectively, in correspondence with the learning ID.
The controller 21 has an arithmetic processing unit such as CPU and executes various information processing and control processing by reading and executing program stored in the auxiliary storage 23. The main storage 22 is a temporary storage area such as RAM and temporarily stores data necessary for the controller 21 to execute arithmetic processing. The auxiliary storage 23 is a non-volatile storage area such as ROM (Read-Only Memory) and stores program and other data necessary for the controller 21 to execute processing. The communication unit 24 is a communication module for executing processing related to communication and sends/receives information to/from the outside. The display unit 25 is a display screen such as liquid crystal display and displays images given by the controller 21. The input unit 26 is an operation interface such as mechanical key and receives operation input from the user.
The identification model 141 has an input layer that accepts input of time-series data, an intermediate layer (hidden layer) that extracts features from the data input to the input layer, and an output layer that outputs identification results based on the features. The input layer has neurons that accept input of data at each of multiple consecutive time points, and the neurons in the input layer pass the input data to the neurons in the intermediate layer. The intermediate layer receives input data from the neurons in the input layer and executes operation based on the input data. In LSTM, the neurons in the intermediate layer are called LSTM blocks, and they temporarily store the results of their own operations. When executing an operation regarding the input data at the next time point, the operation is executed by referring to the operation result of the input data at the previous time point. By referring to the operation result at the previous time point, the operation of the next time point is executed from the time-series data up to the most recent time point. The output layer has neurons that operate the output value based on the operation result in the intermediate layer, and outputs the identification result of the object.
In the present embodiment, the identification model 141 is assumed to be a model with one output in correspondence with multiple inputs (Many-To-One), but it may have multiple outputs. Further, the identification model 141 propagates the operation result of the intermediate layer only in one direction (One-directional) from the past time point to the future time point, but it may be a model that propagates operation result in both directions (Bi-directional).
Although the identification model 141 is described as LSTM in the present embodiment, it may be a model based on other learning algorithms, for example, other deep learning such as CNN (Convolution Neural Network), SVM (Support Vector Machine), decision tree, or the like.
In the present embodiment, the server 1 generates the identification model 141 that inputs the odor data measured the odor of the object and outputs the identification result of the object corresponding to the input odor data. The odor data input to the identification model 141 is the data measured by the odor sensor 3 described above, and is a time series data of frequency measured by QCM method. For example, the odor sensor 3 measures the gas (odor) inhaled into the gas chamber in the odor sensor 3 between measurement time of several seconds to several tens of seconds and obtains the time series data of the frequency over the measurement time. The server 1 uses the data as odor data.
For example, the server 1 receives an upload of odor data measured by the odor sensor 3 from the terminal 2 of each user, and stores the odor data in the odor DB 143 in correspondence with the odor name indicating the object for which the odor was measured and the acquisition condition of the odor data to be described later. The server 1 uses the odor data stored in the odor DB 143 as training data to generate the identification model 141.
That is, the server 1 sequentially inputs the time series odor data measured by the odor sensor 3 to each neuron in the input layer of the identification model 141 according to the time series order at the time of the measurement, and obtains the output value indicating the identification result of the object from the output layer after the operation in the intermediate layer. In the present embodiment, the server 1 considers the object corresponding to the odor data input to the input layer as a two-class classification problem to determine whether or not the object corresponds to a specific object to be learned, and outputs a probability value indicating whether or not it corresponds to the object. By performing the identification of a single object per a single identification model 141, the identification accuracy can be improved compared to the case of multi-class classification in which a single identification model 141 is used to identify multiple objects.
The output value from the identification model 141 may not be a probability value but a value that indicates whether or not it corresponds to the object by a binary value (0 or 1). Further, in the present embodiment, two-class classification is used to determine whether or not the object corresponds to one object, but the model may be used for multi-class classification by simultaneously learning the odor data of multiple objects as training data.
Here, the server 1 performs learning by using not only the odor data measured by the odor sensor 3 but also the acquisition condition of the odor data input to the identification model 141. The acquisition condition of the odor data is information that indicates the condition under which the odor data was measured, and includes, for example, text data that indicates the category of the measured odor, state information that indicates the state of the odor sensor 3 from which the odor data was acquired, and environmental information related to measurement environment when the odor was measured.
The odor category is information that indicates, for example, type of object for which the odor was measured, or the state of the object at the time of odor measurement (for example, if the object is food, the number of days since the food was purchased). These are just examples, and the odor category may be defined arbitrarily. In the present embodiment, the odor category is set by the user who has measured the odor of an object by entering an arbitrary text. In the system, the odor category arbitrarily set by the user is called as “domain”.
For example, the server 1 obtains the domain name arbitrarily input by the user as indicating the odor category when receiving an upload of the odor data from the terminal 2. Specifically, the server 1 acquires the domain name that indicates the odor category and the subdomain name that indicates a further detailed category than the domain name. As an example, the type name of the object for which the odor was measured (“human” in the example of person identification) is input as the domain name, and a more detailed type name than the domain name (for instance, the name of a person) is input as the subdomain name. In this way, the domain is set by the user who uploads the odor data by entering arbitrary text.
The state information is information indicating the state of the odor sensor 3 that has measured the odor of the object, and includes, for example, the pre-cleaning time, the suction time, the residence time, and the post-cleaning time described above. The pre-cleaning time is time to clean the odor sensor 3 before odor measurement. The suction time is time for the odor sensor 3 to inhale gas (odor). The residence time is time when gas is retained in the gas chamber of the odor sensor 3 for measurement. The post-cleaning time is time to clean the odor sensor 3 after odor measurement. In this way, the state information is used to indicate the acquisition state of the odor data and the maintenance state of the odor sensor 3.
The environmental information is information related to the measurement environment when the odor is measured, and includes, for example, location information and weather information. The location information is geographical information of the measurement location, such as name of area where the odor was measured and GPS (Global Positioning System) coordinate values. The weather information is data indicating the weather condition of the measurement location at the time of odor measurement, for example, “sunny”, “rain”, or the like.
In addition to location and weather, the environmental information may also include information related to date and time of measurement (for instance, season).
In addition to the odor data, the server 1 inputs the acquisition condition of the odor data, such as the domain name and the state information, to the identification model 141 for learning. By inputting the acquisition condition of the odor data as well, it is possible to make the learning processing closer to human sense. For example, humans are expected to make different judgments (identifications) when smelling a known object (domain) and when smelling an unknown object. Further, they are also expected to make different judgments based on environment (location, weather, or the like) in which they smell the object. Thus, by using the acquisition condition of the odor data as input, it is possible to reproduce the identification result that is closer to human sense.
For instance, the server 1 inputs various information that defines the acquisition condition of the odor data as category variables into the identification model 141. For example, the server 1 provides a layer for category variable input (not shown) in the input layer of the identification model 141 that is different from the layer for odor data input. The server 1 inputs categorical variables indicating such as the domain name, the subdomain name, the state information, or the environmental information to the layer for category variable input and learns them together with the odor data.
The server 1 inputs the odor data for training and the acquisition condition (categorical variable) for training to the identification model 141 and obtains object identification result from the output layer. The server 1 compares the acquired identification result with the correct object (odor name) and optimizes parameters such as weights between the neurons using the error back propagation method so that they approximate each other. In this way, the server 1 obtains the optimal parameters, that is, learned parameters, by identifying objects with identification model 141.
The server 1 stores the learned parameters obtained from the above learning in the learning DB 145 in correspondence with the object (odor name) to be learned and the acquisition condition of the odor data (domain name, or the like). In this way, the server 1 stores the data of the identification model 141 (learned model) generated by machine learning in the learning DB 145.
The server 1 receives upload of the odor data of various objects from the terminal 2 of each user and performs the above machine learning in response to requests from the users. The server 1 stores the learned parameters of the identification model 141 generated in correspondence with the request from each user in the learning DB 145. Thus, the server 1 manages the data of multiple identification models 141 that have learned the odor data of different objects in the learning DB 145.
The server 1 identifies objects based on the identification model 141 in response to request from the terminal 2. Specifically, the server 1 first receives selection input of learned parameter set in the identification model 141 from each learned parameter stored in the learning DB 145 from the terminal 2. The server 1 sets the selected learned parameter to the identification model 141. In this way, the server 1 receives the selection input of the identification model 141 to be used for object identification from the terminal 2.
Next, the server 1 receives selection input from the terminal 2 to select the odor data of the object to be identified among the odor data stored in the odor DB 143. Of course, the server 1 may newly acquire the odor data of the object to be identified from the terminal 2 of the user instead of the odor data already stored in the odor DB 143.
The server 1 reads the selected odor data and the acquisition condition of the odor data from the odor DB 143 and inputs them to the identification model 141 in which the learned parameters are set. Then, the server 1 obtains the identification result of the object corresponding to the input odor data as an output from the identification model 141. Specifically, as described above, the server 1 obtains the probability value of whether or not the object to be identified corresponds to the object learned in the identification model 141. The server 1 outputs the identification result to the terminal 2 for and display the identification result.
From the above, according to the present embodiment, it is possible to identify objects appropriately based on odor by using not only the odor data but also the acquisition conditions of the odor data as input to the identification model 141.
The controller 11 of the server 1 acquires training data for generating the identification model 141 from the odor DB 143 (Step S11). The training data is the data in which the correct object (odor name) is associated with the odor data of the object measured by the odor sensor 3 and the acquisition condition of the odor data. The acquisition condition includes the domain name and subdomain name arbitrarily input by the user as text indicating the odor category, the state information indicating the state of the odor sensor 3 such as pre-cleaning time and suction time, and the environmental information such as location information and weather information related to measurement environment of the odor.
Based on the training data, the controller 11 generates the identification model 141 that outputs the identification result of the object using the odor data and the acquisition condition of the odor data as inputs (Step S12). Specifically, as described above, the controller 11 generates a neural network (LSTM) as the identification model 141. The controller 11 inputs the odor data and the category variable indicating the acquisition condition to the identification model 141, and obtains the identification result of that identified the object corresponding to the odor data as an output. The controller 11 compares the acquired identification result with the correct object and generates the identification model 141 by optimizing parameters such as weights between neurons, that is, learned parameters, so that the two are approximated.
The controller 11 stores the learned parameters related to the generated identification model 141 in the learning DB 145 in correspondence with the object (odor name) to be learned and the acquisition condition of the odor data (domain name, etc.) (Step S13) and ends the series of processing.
The controller 11 receives the selection input to select the identification model 141 to be used for identifying the object from the terminal 2 (step S33). Specifically, as described above, the controller 11 receives the selection input of the learned parameter to be set in the identification model 141. The controller 11 sets the selected learned parameter to the identification model 141, and inputs odor data and acquisition condition of the object to the identification model 141 to identify the object (Step S34). The controller 11 outputs the identification result to the terminal 2 (Step S35) and ends the series of processing.
From the above, according to the first embodiment, it is possible to identify objects appropriately based on odor.
Further, according to the first embodiment, it is possible to identify objects more appropriately by using the state information indicating the state of the odor sensor 3 as data acquisition condition for input to the identification model 141.
According to the first embodiment, it is possible to identify objects more appropriately based on odor by using the odor sensor 3 with the quartz crystal.
Further, according to the first embodiment, it is possible to identify objects more appropriately by using the environmental information indicating the measurement environment at the time of odor measurement as data acquisition condition for input to the identification model 141.
In addition, according to the first embodiment, it is possible to provide various options (identification model 141) to users by generating a plurality of identification models 141 that perform two-class classification and allowing the users to select the identification model 141 to be used from a plurality of identification models 141.
In the present embodiment, an embodiment related to UI (User Interface) screen for users to use the identification system described above will be illustrated. Descriptions that overlap with those in the first embodiment are indicated with the same signs and are omitted hereinafter.
Specifically, the terminal 2 receives the selection input of the odor sensor 3 to be used for odor measurement in a sensor selection column 81. When the odor sensor 3 is selected, the acquisition condition of the odor data such as domain name or subdomain name are displayed as default. For example, the server 1 receives registration of the domain name, the subdomain name, and the state information corresponding to the odor sensor 3 used by the user in advance from the user, and stores them in the domain DB 144 in correspondence with the user ID. When the odor sensor 3 of the user is selected in the sensor selection column 81, the domain name or the like registered by the user is set as acquisition condition by default. The setting of the domain name or the like changed by the user may be received on the measurement screen in
Further, the terminal 2 receives setting input related to the environmental information such as location information and weather information in each column such as “location”, “GPS information”, and “weather”. Finally, the terminal 2 receives the input of the odor name indicating the object to be measured in the odor name input column 82, and starts the odor measurement by the odor sensor 3 corresponding to the operation input to an execution button 83.
The server 1 acquires the measured odor data described above along with the acquisition condition set on the measurement screen, the odor name (correct object), and the like. The server 1 stores each acquired data in the odor DB 143.
For example, as shown in
Specifically, based on the operation input to an odor data selection column 101, the terminal 2 receives the selection input to select the odor data to be learned from the odor data stored in the odor DB 143. For example, when the operation input to the odor data selection column 101 is received, the terminal 2 pops up a list of the odor data similar to the list screen in
When the odor data to be learned is selected, the domain name and subdomain name corresponding to the selected odor data are displayed as default in a domain selection column 102 and a subdomain selection column 103. The terminal 2 receives setting changes to change the default displayed domain name and subdomain name corresponding to the operation input to the domain selection column 102 and the subdomain selection column 103. In this way, the terminal 2 receives selection input of the acquisition condition (domain name, etc.) of the odor data to be learned along with the odor data.
Moreover, the terminal 2 receives text input of the odor name indicating the object to be learned in an odor name input column 104. In this way, the terminal 2 receives the input of the correct name of the object (correct object) corresponding to the odor data selected in the odor data selection column 101.
In correspondence with the operation input to an execution button 105, the terminal 2 requests the server 1 to perform machine learning based on various information input described above. According to the request from the terminal 2, the server 1 performs machine learning to learn the selected odor data and the acquisition condition and the correct object corresponding to the odor data and generates the identification model 141. The server 1 stores the learned parameters (weights, etc.) of the generated identification model 141 in the learning DB 145 corresponding to the odor name, domain name, and the like.
In the learning screen in
Specifically, the terminal 2 receives the selection input of the learned parameters to be set to the identification model 141 in a learned parameter selection column 111. For example, when the terminal 2 receives the operation input to the learned parameter selection column 111, it displays a pop-up list of information on each learned parameter (data of each identification model 141) stored in the learning DB 145. Specifically, the terminal 2 displays a list of the odor name, the domain name, the subdomain name, or the like stored in the learning DB 145 corresponding to each learned parameter. The terminal 2 receives the selection input to select any of each learned parameter displayed on the list. In this way, the terminal 2 receives the selection input of the identification model 141 used to identify the object.
Next, the terminal 2 receives the selection input of one or more odor data to be identified in an odor data selection column 112. Similar to the learning screen, when receiving the operation input to the odor data selection column 112, the terminal 2 pops up a list of the odor data similar to the list in
When the odor data to be identified is selected, it is preferably that the server 1 determines whether or not the acquisition condition of the odor data (domain name, etc.) is coincident with the acquisition condition of the odor data to be learned in the selected identification model 141 described above, and outputs an error if they do not match. In this way, it is possible to avoid inappropriate situations such as, for instance, using an identification model 141 with a domain name different from the object to be identified.
The terminal 2 receives the operation input to an execution button 113 and requests the server 1 to identify the object. In correspondence with the request from the terminal 2, the server 1 sets the learned parameters selected in the learned parameter selection column 111 to the identification model 141, and inputs the odor data selected in the odor data selection column 112 to the identification model 141 to identify the object. Specifically, the server 1 identifies each of the one or more odor data selected in the odor data selection column 112, and outputs the identification result (probability value) of each odor data to the terminal 2. The terminal 2 displays the identification result of each odor data output from the server 1.
On the determination screen, an identification tab 114 and a search tab 115 are displayed at the top of the screen. When the identification tab 114 is selected, the terminal 2 makes the server 1 to perform the identification of the object according to the procedure described above. On the other hand, when the search tab 115 is selected, the terminal 2 makes the server 1 to perform the identification (search) of the object using one or more identification models 141.
Specifically, the terminal 2 receives the selection input to select one or more learned parameters corresponding to a selection input to the learned parameter selection column 111 on the determination screen similar to that of
Based on the identification results in each identification model 141, the server 1 outputs an identification result (search result) to identify which of the objects corresponding to each identification model 141 corresponds to the object according to the input odor data. For instance, the server 1 ranks the objects according to the probability values output from each identification model 141, and outputs the object names in descending order of the probability values. Alternatively, the server 1 may output the object name with the highest probability value as the search result.
As described in the first embodiment, it is possible to prepare an identification model 141 capable of multi-class classification by having one identification model 141 learn the odor data of multiple objects, but the identification accuracy can be improved by making the identification model 141 as a model that performs two-class classification. By combining a plurality of these identification models 141, it is possible to determine the correspondence between a specific odor and a plurality objects, that is, multi-class classification can be performed suitably.
The controller 11 of the server 1 determines whether or not to measure the odor data in response to the operation input in the terminal 2 (step S201). If it is determined that measurement of the odor data is to be performed (S201: YES), the controller 11 receives the odor name indicating the object to be measured (correct object) as well as the setting input of the acquisition condition of the odor data from the terminal 2 (Step S202). The terminal 2 performs measurement by the odor sensor 3 in response to the operation input from the user, and the controller 11 of the server 1 acquires odor data from the odor sensor 3 via the terminal 2 and stores the odor data in the odor DB 143 in correspondence with the acquisition condition or the like set in step S202 (step S203).
After executing the processing of step S203, or in the case of NO in step S201, the controller 11 determines whether or not to perform the learning of the odor data in response to the operation input in the terminal 2 (step S204). If it is determined that learning is to be performed (S204: YES), the controller 11 receives a selection input from the terminal 2 to select the odor data to be learned (Step S205). Specifically, as described above, the controller 11 receives setting input of the odor data to be learned as well as the acquisition condition of the odor data, the correct object, or the like.
The controller 11 performs machine learning based on the selected odor data, the acquisition conditions of the odor data, and the input correct object, and generates the identification model 141 (Step S206). The controller 11 stores the learned parameters of the generated identification model 141 in the learning DB 145 (Step S207).
After performing the processing of step S207, or in the case of NO in step S204, the controller 11 determines whether or not to identify the odor data in response to the operation input in the terminal 2 (step S208). If it is determined that the identification of the odor data is to be performed (S208: YES), the controller 11 receives the selection input to select the identification model 141 to be used for identification of the odor data (step S209). Specifically, as described above, the controller 11 receives the selection input to select the learned parameter set in the identification model 141.
The controller 11 receives the selection input to select the odor data to be identified (Step S210). The controller 11 sets the learned parameter selected in step S209 to the identification model 141 and inputs the odor data selected in step S210 to identify the object (step S211). The controller 11 outputs the identification result to the terminal 2 (Step S212). After performing the processing of step S212, or in the case of NO in step S208, the controller 11 ends the series of processing.
Overall, according to the second embodiment, it is possible to provide users with a platform that can learn and identify odor data with simple operations.
The embodiments described in the present invention are presented as examples and are not intended to limit the scope of the invention. The scope of the present invention is not in the sense described above but is indicated by the claims, and it is intended to include all modifications within the scope of the claims and the equivalent scope thereof.
This is a U.S. National Phase Application under 35 U.S.C. 371 of International Application No. PCT/JP2019/047343, filed on Dec. 4, 2019, which claims priority to U.S. Patent Application No. 62/775,561, filed on Dec. 5, 2018. The entire disclosures of the above applications are expressly incorporated by reference herein.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/JP2019/047343 | 12/4/2019 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 62775561 | Dec 2018 | US |
