Abstract: Provided is a learning method including (a) preparing a plurality of pieces of data for learning; (b) dividing the plurality of pieces of data for learning into one or more groups to generate one or more input learning data groups; and (c) training M number of machine learning models, wherein (b) includes (b1) dividing the plurality of pieces of data for input into one or more regions to generate, as one of the input learning data groups, a collection of first type divided input data after division belonging to the same region, or (b2) dividing the plurality of pieces of data for learning belonging to one class into one or more groups to generate, as one of the input learning data groups, a collection of second type divided input data after division.

Abstract: An evaluation method for a trained machine learning model includes the steps of (a) inputting evaluation data to the trained machine learning model to generate first explanatory information used for an evaluation of the machine learning model, (b) using a value indicated by each piece of information included in the first explanatory information to generate second explanatory information indicating an evaluation of the trained machine learning model, and (c) outputting the generated second explanatory information.

Abstract: A method according to the present disclosure includes (a) generating N pieces of input data from one target object, (b) inputting the input data to a machine learning model and obtaining M classification output values, one determination class, and a feature spectrum, (c) obtaining a similarity degree between a known feature spectrum group and the feature spectrum for the input data, and obtaining a reliability degree with respect to the determination class as a function of the reliability degree, and (d) executing a vote for the determination class, based on the reliability degree with respect to the determination class, and determining a class determination result of the target object, based on a result of the vote.

Abstract: A method of performing classification processing on classification target data includes: (a) a step of preparing N machine learning models; (b) a step of, when a plurality of pieces of training data are input into the N machine learning models, preparing a known feature vector group obtained from output of at least one specific layer of the plurality of vector neuron layers; and (c) a step of computing, using a selected machine learning model selected from the N machine learning models a similarity, for each class, between the known feature vector group and a feature vector obtained from output of the specific layer when the classification target data is input into the selected machine learning model, and determining a class for the classification target data using the similarity.

Abstract: A quality determination method includes: (a) generating a plurality of pieces of training data by classifying a plurality of pieces of non-defective product data into a plurality of classes; (b) executing learning of a machine learning model using the plurality of pieces of training data; (c) preparing a known feature spectrum group; and (d) executing quality determination processing of inspection data using the machine learning model and the known feature spectrum group. The (d) includes (d1) calculating a feature spectrum related to the inspection data, (d2) calculating a similarity between the feature spectrum and the known feature spectrum group, and (d3) determining the inspection data to be non-defective when the similarity is equal to or greater than a threshold value and determining the inspection data to be defective when the similarity is less than the threshold value.

Abstract: A class discrimination method includes: (a) a step of preparing, for each class, a known feature spectrum group obtained based on an output of a specific layer among a plurality of vector neuron layers when a plurality of pieces of training data are input to a machine learning model; and (b) a step of executing a class discrimination processing of the data to be discriminated using the machine learning model and the known feature spectrum group. The step (b) includes: (b1) a step of calculating a feature spectrum based on an output of the specific layer according to the data to be discriminated to the machine model; (b2) a step for each of the one or more classes; (b3) a step of creating an explanatory text of a class discrimination result for the data to be discriminated according to the similarity; and (b4) a step of outputting the explanatory text.

Abstract: An image processing method includes: an image pickup step of picking up an RGB image of a target object to be picked up, and picking up a spectroscopic image of the target object in a predetermined wavelength range and thus acquiring spectroscopic information peculiar to the target object in the wavelength range; and a display step of displaying a complemented image complemented by superimposing the spectroscopic information on the RGB image.

Abstract: A class determination method includes: step (a): preparing, for each of a plurality of classes, a known feature spectrum group obtained when a plurality of pieces of training data are input to a vector neural network type machine learning model; and step (b): executing, by using the machine learning model and the known feature spectrum group, a class determination processing on data to be determined. The step (b) includes step (b1), calculating a feature spectrum according to an input of the data to be determined to the machine learning model, step (b2), calculating a class similarity between the feature spectrum and the known feature spectrum group related to each of the plurality of classes, and step (b3), determining a class of the data to be determined according to the class similarity.

Abstract: An object detection method includes inputting an input image to a learned machine learning model and generating a similarity image from an output of at least one specific layer, and generating a discriminant image to which at least an unknown label is assigned, by comparing a similarity of each pixel in the similarity image to a predetermined threshold value.

Abstract: A method causes one or more processors to execute a method in which a machine learning model of a vector neural network type is used. The model is learned to reproduce correspondence between first images and a pre-label corresponding to each of the first images, and includes one or more neuron layers. First intermediate data output by the one or more neurons when the first images are input to the learned model is stored in one or more memories in correlation with the neurons. The method includes inputting a second image of an object to the machine learning model and acquiring second intermediate data based on at least one of a second vector and a second activation included in the one or more neurons, calculating a similarity degree between the first and second intermediate data, generating an evidence image corresponding to the similarity degree, and displaying the generated evidence image.

Abstract: A method of making a single processor or a plurality of processors perform classification processing of classification target data using a machine learning model includes the steps of (a) preparing N machine learning models in a memory assuming N as an integer no smaller than 2, and (b) performing the classification processing of the classification target data using the N machine learning models. Each of the N machine learning models is configured so as to classify input data into any of a plurality of classes with learning using training data, and is configured so as to have at least one class different from a class of another of the N machine learning models.

Abstract: An identification method for identifying a target to be measured includes accepting, from an input section, an information representing a condition for acquiring spectral information specific to a target to be measured, capturing, by a spectrometry camera, an image of the target, acquiring the spectral information specific to the target based on the captured image, and identifying the target based on (i) the spectral information and (ii) a database, stored in a memory, containing a plurality of pieces of spectral information corresponding to a plurality of objects. Acquiring the spectral information includes preferentially acquiring the spectral information specific to the target in a specific wavelength region where the target is identifiable.

Abstract: A calibration data acquisition unit (a) acquires Q optical spectra, S evaluation spectra, and a reference spectrum of a target component, (b) extracts R subsets from a set of the optical spectra, (c) performs independent component analysis in which component amounts are treated as independent components on each subset so as to acquire component natural spectra and component calibration spectra, (d) obtains an inner product value between the component calibration spectrum and the evaluation spectrum, (e) obtains a correlation degree between a component amount for the target component and the inner product value with respect to each component calibration spectrum, (f) obtains a similarity between each component natural spectrum and a reference spectrum, (g) selects a component calibration spectrum causing a comprehensive evaluation value based on the correlation degree and the similarity to be greatest as the target component calibration spectrum, and (h) creates a calibration curve.

Abstract: A determination method includes: obtaining measurement data; selecting 0 or more second wavelengths from a plurality of first wavelengths including at least one of a plurality of measurement wavelengths to generate a plurality of individuals, by using a genetic algorithm; inputting, to a first model learned to reproduce a correct answer label of a target object, the measurement data of the target object belonging to a remaining group and a second spectroscopic spectrum determined by the second wavelength to discriminate a label of the target object belonging to the remaining group, for each of the plurality of individuals; and determining whether or not to use the second wavelength as the wavelength of the spectroscopic spectrum for discrimination based on a rate at which the label is correctly discriminated.

Abstract: An apparatus including: memory that stores a machine learning model of a vector neural; and one or more processors that execute an arithmetic operation. The machine model has a plurality of vector neuron layers each including a plurality of nodes. When one of the plurality of vector layers is referred to as an upper layer and a vector layer below is referred to as a lower layer, one or more processors execute outputting one output vector by using output vectors from the plurality of nodes of the lower layer as an input for each node of the upper layer, the outputting including: obtaining a prediction vector, obtaining a sum vector based on a linear combination of the vectors, obtaining a normalization coefficient, and obtaining the output vector of the target node by dividing the sum vector by the norm and multiplying the divided sum vector by the normalization coefficient.

Abstract: A method of causing one or more processors to execute: performing learning of a model that is an algorithm of a vector neural network type to reproduce correspondence between a plurality of first data elements included in a first data set and a pre-label corresponding to each of the plurality of first data elements, in which the model has one or more neuron layers, each of the one or more neuron layers has one or more neuron groups, each of the one or more neuron groups has one or more neurons, and each of the one or more neurons outputs first intermediate data based on at least one of a first vector and a first activation; and inputting the first data set into the learned model and acquiring the first intermediate data output by the one or more neurons by being associated with the neuron.

Abstract: A method for causing one or more processors to execute: performing learning of a first model of a capsule network type including one or more capsule layers each having one or more capsules to reproduce correspondence between a plurality of first data elements included in a first data set and a pre-label corresponding to each of the plurality of first data elements; and inputting the first data set into the learned first model and acquiring first intermediate data based on at least one of a first activation and a first pose included in the one or more capsules, for the one or more capsule layers.

Abstract: A spectroscopic system includes a main body including a light source that radiates light to a light transmissive measurement target, an imaging device that captures an image based on transmitted light having passed through the measurement target, and a spectroscopy section that is provided in an optical path between the light source and the imaging device and selectively transmits light that belongs to a specific wavelength region, and an attachment that includes an optical path changer that changes the direction of the optical path of the light outputted from the light source and is so attached to the main body as to form a placement space which is located between the optical path changer and the main body and in which the measurement target is placed.

Abstract: An image processing method includes: an image pickup step of picking up an RGB image of a target object to be picked up, and picking up a spectroscopic image of the target object in a predetermined wavelength range and thus acquiring spectroscopic information peculiar to the target object in the wavelength range; and a display step of displaying a complemented image complemented by superimposing the spectroscopic information on the RGB image.

Abstract: A display method according to the present disclosure includes an imaging step of imaging inherent spectroscopic information provided to an object to be measured as a first image, and taking an image of the object different from the spectroscopic information as a second image, and a display step of inevitably displaying the second image and selectively displaying the first image out of the first image and the second image.