Data processing device, data processing method and data processing program

Abstract

To provide a data processor, a method for processing data, and a data processing program that can increase the accuracy.SOLUTION: According to an embodiment, the data processor includes an acquisition unit and a processing unit. The acquisition unit can acquire first data including time-series image data. The processing unit can derive first feature information based on a multi-dimensional array of n-dimensions (n is an integer of at least 3) based on the first data acquired by the acquisition unit. The first axis of the multi-dimensional array relates to time.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] An embodiment of the present invention relates to a data processing device, a data processing method, and a data processing program. [Background technology]

[0002] For example, the state of the evaluation target is evaluated by processing target image data etc. Evaluation with higher accuracy is desired. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-91756 A Summary of the Invention [Problem to be solved by the invention]

[0004] The embodiments of the present invention provide a data processing device, a data processing method, and a data processing program that can improve accuracy. [Means for solving the problem]

[0005] According to an embodiment of the present invention, a data processing device includes an acquisition unit and a processing unit. The acquisition unit is capable of acquiring first data including time-series image data. The processing unit is capable of deriving first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data acquired by the acquisition unit. A first axis of the multidimensional array relates to time. [Brief description of the drawings]

[0006] [Figure 1] FIG. 1 is a schematic diagram illustrating a data processing device according to the first embodiment. [Diagram 2] FIG. 2 is a flowchart illustrating the operation of the data processing device according to the first embodiment. [Diagram 3]FIG. 3 is a schematic diagram illustrating the operation of the data processing device according to the first embodiment. [Figure 4] FIG. 4 is a schematic diagram illustrating the operation of the data processing device according to the first embodiment. [Diagram 5] FIG. 5 is a schematic diagram illustrating the operation of the data processing device according to the first embodiment. [Figure 6] 6(a) to 6(e) are schematic views illustrating the operation of the data processing device according to the first embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0007] Hereinafter, embodiments of the present invention will be described with reference to the drawings. The drawings are schematic or conceptual, and the relationships of the lengths of the parts are not necessarily the same as those in reality. Even when the same parts are shown, the dimensions and ratios of the parts may be different depending on the drawing. In this specification and each drawing, elements similar to those described above with reference to the previous drawings are given the same reference numerals and detailed descriptions thereof will be omitted as appropriate.

[0008] (First embodiment) FIG. 1 is a schematic diagram illustrating a data processing device according to the first embodiment. 1, a data processing device 110 according to an embodiment includes an acquisition unit 31 and a processing unit 32. The acquisition unit 31 is capable of acquiring first data Da1. The first data Da1 includes, for example, time-series image data. The acquisition unit 31 is, for example, an input port.

[0009] The processing unit 32 is capable of deriving the first characteristic information Pa1 based on the first data Da1 acquired by the acquisition unit 31. The processing unit 32 is capable of outputting the derived first characteristic information Pa1.

[0010] The data processing device 110 may include, for example, an output unit 35. The processing unit 32 may output the first feature information Pa1 via the output unit 35. The output unit 35 is, for example, an output port.

[0011] The acquisition unit 31 may have an input function and an output function. In this case, the output unit 35 may be omitted. The acquisition unit 31 is, for example, an input / output port.

[0012] The processing unit 32 includes, for example, an electric circuit (electronic circuit) etc. The processing unit 32 may include, for example, a CPU (Central Processing Unit) etc.

[0013] The data processing device 110 may include a storage unit 33. The storage unit 33 may include, for example, at least one of a magnetic recording device and a semiconductor storage device. The storage unit 33 may include, for example, at least one of a ROM (Read Only Memory) and a RAM (Random Access Memory). The data processing device 110 may include a display unit 34 and the like. The display unit 34 may have an input function.

[0014] The data processing device 110 may supply a control signal Sig1 to the target device 40. The target device 40 is an object to be controlled by the data processing device 110.

[0015] Data (information) may be transmitted and received between multiple elements included in the data processing device 110 via wires and / or wirelessly.

[0016] The processing unit 32 can derive an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data Da1, for example. The first feature information Pa1 is based on the derived n-dimensional (n is an integer equal to or greater than 3) multidimensional array. One axis (first axis) of the n-dimensional (n is an integer equal to or greater than 3) multidimensional array relates to time.

[0017] First feature information Pa1 is derived based on a multidimensional array including a time-related axis. First data Da1 including time-series image data is evaluated using the first feature information Pa1. This allows, for example, the time-series image data to be evaluated with high accuracy. For example, smaller temporal changes can be extracted from the time-series image data. For example, smaller spatial changes can be extracted from the time-series image data.

[0018] In one example, the time-series data of the first data Da1 acquired by the acquisition unit 31 may be video data of the state of the target device 40. Depending on the conditions of the target device 40, proper operation may or may not be obtained. For example, in order to obtain information on proper conditions for the target device 40, there are attempts to obtain proper conditions by acquiring various video data under various conditions and analyzing the various video data.

[0019] In one example, the target device 40 is a circuit breaker. In the circuit breaker, an arc generated in a reduced pressure vessel switches on and off. The state of the circuit breaker is evaluated based on video data capturing the state of the arc. In another example, the target device 40 may be a fluid forming device. For example, the flow of a fluid such as a liquid or gas is imaged by a plurality of particles. A fluid that changes over time is evaluated. In another example, the target device 40 may be a welding device. For example, the state of welding is evaluated based on operation data of a welding part during welding. In another example, a flow is evaluated for an object or a person. The data processing device 110 according to the embodiment can be applied to such various examples.

[0020] An example of the operation of the data processing device 110 will now be described. FIG. 2 is a flowchart illustrating the operation of the data processing device according to the first embodiment. 2, in the data processing device 110, the first data Da1 is acquired (step S11). Step S11 is performed by the acquiring unit 31.

[0021] The first data Da1 changes, for example, during one period. The first data Da1 includes, for example, time-series image data. For example, the acquisition unit 31 reads the time-series image data.

[0022] The acquired first data Da1 (time-series image data) is supplied to the processing unit 32. The processing unit 32 performs the following processes on the first data Da1 (time-series image data).

[0023] The processing unit 32 converts, for example, the first data Da1 (time-series image data) into a plurality of data (step S12). For example, the processing unit 32 converts the time-series image data into pixel data per unit time. The pixel data includes, for example, a value related to the brightness of the pixel.

[0024] For example, for the i-th time, the first data Da1 (time-series image data) is converted into data of j×k pixels. "i" is an integer greater than or equal to 1 and less than or equal to "im". "j" is an integer greater than or equal to 1 and less than or equal to "jm". "k" is an integer greater than or equal to 1 and less than or equal to "km". "im", "jm", and "km" are integers greater than or equal to 2 and may be determined based on the first data Da1.

[0025] The processing unit 32 derives a multidimensional array based on the data obtained by the conversion (step S13). For example, the multidimensional array is derived based on a plurality of values, 1 to "im", 1 to "jm", and 1 to "km". The plurality of values, 1 to "im", 1 to "jm", and 1 to "km", correspond to, for example, pixel data (for example, a plurality of values ​​related to the brightness of a pixel). The pixel data (for example, a plurality of values ​​related to the brightness of a pixel) may be, for example, a value L 1,1,1 ~Value L im,jm,km One of the multiple values ​​of the pixel data is the value L i,j,k It is.

[0026] For example, a first axis of the multi-dimensional array relates to time. For example, a second axis of the multi-dimensional array relates to a first coordinate value on a first spatial axis in the space. For example, a third axis of the multi-dimensional array relates to a second coordinate value on a second spatial axis in the space. The second spatial axis intersects the first spatial axis. The second spatial axis may be perpendicular to the first spatial axis. For example, the second axis may be a coordinate value on an X spatial axis in the space. For example, the second axis may be a coordinate on a Y spatial axis in the space. The space may be at least a portion of a space corresponding to the time series image data.

[0027] For example, the time-series image data includes values ​​related to brightness at positions corresponding to the first coordinate value and the second coordinate value.

[0028] In the embodiment, for example, the time series image data may include a value related to a color at a position corresponding to the first coordinate value and the second coordinate value. The value related to the color may include a stimulus value in a chromaticity coordinate. The pixel data (multiple values ​​in a pixel) may include, for example, a value L 1,1,1,1 ~Value L im,jm,km,pm One of the multiple values ​​of the pixel data is the value L i,j,k,p "p" is an integer equal to or greater than 1 and equal to or less than "pm". "pm" is an integer equal to or greater than 2 and may be determined based on the first data Da1. "p" may correspond to, for example, one of the axes of a plurality of stimulus values ​​in the chromaticity coordinates. A value for one of the axes of a plurality of stimulus values ​​in the chromaticity coordinates may be divided into a plurality of divisions from 1 to "pm". When the time-series image data includes values ​​related to color, the n-dimensional multidimensional array includes, for example, a four-dimensional multidimensional array. For example, when "p" is 1, it corresponds to the chromaticity coordinate x in the xy chromaticity diagram. When "p" is 2, it corresponds to the chromaticity coordinate y in the xy chromaticity diagram.

[0029] In the embodiment, for example, processing is performed based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array. The n-dimensional (n is an integer equal to or greater than 3) multidimensional array includes time as one of a plurality of components, whereby a temporal change in the time-series image data is effectively extracted, and the first feature information Pa1 is calculated.

[0030] In the embodiment, the first feature information Pa1 may be derived based on a persistence map. As shown in Fig. 2, the processing unit 32 may derive a persistence map (step S14).

[0031] The persistence diagram corresponds to a distribution state of at least a part of the first data Da1 (e.g., time-series image data). For example, the persistence diagram is derived based on the above multidimensional array. The first feature information Pa1 may include a persistent feature value in the persistence diagram. An example of the persistence diagram will be described later.

[0032] The processing unit 32, for example, determines whether a persistent feature value in the persistence diagram exceeds a reference value (threshold value) (step S15). For example, if a persistent feature value that exceeds the reference value exists, a signal (data) corresponding to "the presence of a persistent feature value that exceeds the reference value" is output (step S16). For example, this signal (data) becomes at least a part of the first feature information Pa1. The first feature information Pa1 may include, for example, information indicating that a peculiar state has been detected. The peculiar state may be, for example, an abnormality.

[0033] For example, if no persistent feature value exceeds the reference value, the method may return to step S11, where additional data may be acquired and processed.

[0034] FIG. 3 is a schematic diagram illustrating the operation of the data processing device according to the first embodiment. 3 illustrates the first data Da1. The first data Da1 includes time-series data (data DA1 to DAim).

[0035] The processing unit 32 converts such first data Da1 into data DC1 of 1st to im-th j×k pixels.

[0036] FIG. 4 is a schematic diagram illustrating the operation of the data processing device according to the first embodiment. FIG. 4 illustrates an example of a persistence diagram PD. For example, a mathematical operation called filtration is performed on a simplicial complex to obtain an augmented sequence of simplicial complexes. The information included in the persistence diagram PD relates to a "set" including, for example, the occurrence of a generator of a homology and the disappearance of a generator of a homology. The occurrence of a generator of a homology corresponds, for example, to a first parameter F1. The disappearance of a generator of a homology corresponds, for example, to a second parameter F2. The horizontal axis of the persistence diagram PD corresponds to the first parameter F1. The vertical axis of the persistence diagram PD corresponds to the second parameter F2. The persistence diagram PD includes information regarding the presence of holes in the figure. The persistence diagram PD also includes scale information including the size of the hole and the shape of the hole.

[0037] For example, when the input is a sequence of points, a persistence diagram PD is obtained by filtering the check complex. An augmented sequence of check complexes can be obtained by, for example, increasing the radius of the input sequence of points as a parameter. When analyzing a one-dimensional ring structure by persistent homology, the first parameter F1 corresponds to the minimum radius at which a space surrounded by multiple spheres is formed. The second parameter F2 corresponds to the minimum radius at which the space surrounded by multiple spheres disappears. The persistence diagram PD obtained in this case corresponds to a first-order persistence diagram PD. A persistence diagram PD obtained by analyzing a q-dimensional ring structure is called a q-order persistence diagram PD. "q" is an integer equal to or greater than 0. When "q" is 0, the ring structure is a connected structure. When "q" is 2, the ring structure is a void structure.

[0038] On the other hand, if the input is image data, which is a collection of pixel values ​​or voxels, then cube complex filtration is applied. An augmented sequence of cube complexes can be obtained, for example, by binarization or level set methods.

[0039] For example, in the case of pixel data, the binarization method may obtain an augmented sequence of pixel images by gradually expanding the 0 (black) region pixel by pixel, while the augmented sequence of pixel images may also be obtained by gradually expanding the 255 (white) region pixel by pixel.

[0040] In the level set method, for example, a threshold value for pixel values ​​is defined. For example, the threshold value is changed from a first value to a second value, and an augmented sequence of pixel images is obtained by augmenting regions with pixel values ​​above (or below) the threshold value. The first value is, for example, one of 0 and 255, and the second value is, for example, the other of 0 and 255.

[0041] 4, in the persistence diagram PD, each of a plurality of values ​​of the acquired first data Da1 (time-series image data) has a combination of a first parameter F1 and a second parameter F2. The combination of the first parameter F1 and the second parameter F2 may be plotted in the persistence diagram PD.

[0042] 4, a group G1 may be recognized in the persistence diagram PD. Meanwhile, a value G2 that does not belong to the group G1 may be obtained in the persistence diagram PD. For example, the group G1 or the value G2 may be extracted.

[0043] For example, as shown in FIG. 4, a characteristic function P1 is defined that indicates a relationship between a first parameter F1 and a second parameter F2. A deviation from the characteristic function P1 is defined as a reference value R1. For example, values ​​included in a group G1 are within a range of the reference value R1 based on the characteristic function P1. For example, a value G2 exceeds the range of the reference value R1 based on the characteristic function P1. The value G2 corresponds to a persistent feature value that exceeds the reference value R1, for example.

[0044] FIG. 5 is a schematic diagram illustrating the operation of the data processing device according to the first embodiment. These figures illustrate a multidimensional array of the first data Da1 (time-series image data). Fig. 5 illustrates a multidimensional array of the first to fifth sample data SP1 to SP5. The samples have different times t. In this example, the first data Da1 (time-series image data) corresponds to a temporal change in an arc in a circuit breaker.

[0045] 6(a) to 6(e) are schematic views illustrating the operation of the data processing device according to the first embodiment. 6(a) to 6(e) correspond to persistence diagrams PD corresponding to the first to fifth sample data SP1 to SP5, respectively. These diagrams include zero-order persistence.

[0046] As shown in FIG. 6(a) and FIG. 6(e), in the first sample data SP1 and the fifth sample data SP5, there is a value G2 that does not belong to the group G1.

[0047] In this example, in the first sample data SP1 and the fifth sample data SP5, the circuit breaker is easy to break. It is considered that easy breaking is related to the behavior of the arc. In the second to fourth sample data SP2 to SP4, the circuit breaker is difficult to break. It is considered that difficult breaking is related to the behavior of the arc. In this way, for example, multiple sample data may be evaluated depending on the presence or absence of the value G2.

[0048] According to the embodiment, for example, time-series image data can be evaluated with high accuracy, and for example, moving image data can be evaluated effectively.

[0049] As described above, the persistence diagram PD may include zero-order persistence. The persistence diagram PD may also include zero-order persistence, first order persistence, and / or second order persistence.

[0050] The persistence diagram PD may correspond to one of the first data Da1 divided into a plurality of periods. For example, the processing unit 32 may divide the time series data included in the first data Da1 into a plurality of periods. The persistence diagram PD may be derived for at least one of the periods.

[0051] In the embodiment, the first feature information Pa1 may include a feature vector obtained by converting the persistent feature value (e.g., value G2). The feature vector is obtained, for example, by dividing the persistent diagram PD into a plurality of grids and converting it into a persistent image in which the difference values ​​between the first parameter F1 and the first parameter F2 in the plurality of grids are arranged. In the case of the above example, the feature vector is, for example, the difference between the first parameter F1 and the first parameter F2 in the vicinity of the value G2.

[0052] A first reference value R1 for a persistent feature value (e.g., value G2) may be defined. The processing unit 32 may output first feature information Pa1 for the first data based on a result of comparing the first reference value R1 with the persistent feature value (e.g., value G2).

[0053] In the embodiment, the processing unit 32 may be capable of extracting characteristic parts from the first data Da1 based on machine learning using the first feature information as an input.

[0054] 1 and 2, the data processing device 110 may be capable of supplying a control signal Sig1 to the target device 40 based on the first characteristic information Pa1 (step S17). In the above example, the target device 40 is a circuit breaker. In this case, the first data Da1 relates to the target device 40.

[0055] Second embodiment The second embodiment relates to a data processing method. The data processing method according to the embodiment includes acquiring first data Da1 including time-series image data (e.g., step S11). The data processing method derives first feature information Pa1 based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the acquired first data Da1 (e.g., step S16). The first axis of the multidimensional array relates to time.

[0056] A second axis of the multi-dimensional array relates to a first coordinate value on a first spatial axis in the space. A third axis of the multi-dimensional array relates to a second coordinate value on a second spatial axis in the space. The second spatial axis intersects with the first spatial axis.

[0057] The time-series image data includes, for example, a value related to brightness at a position corresponding to the first coordinate value and the second coordinate value. The time-series image data may also include a value related to color at a position corresponding to the first coordinate value and the second coordinate value.

[0058] The first feature information Pa1 may include persistent feature values ​​in a persistent map PD, which corresponds to a distribution state of at least a part of the first data Da1, and which is derived based on a multi-dimensional array.

[0059] Third embodiment The third embodiment relates to a data processing program. The data processing program according to the embodiment causes a computer to carry out the above-described method.

[0060] For example, a data processing program according to an embodiment causes a computer to acquire first data Da including time-series image data, and derive first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data Da.

[0061] For example, a first axis of the multi-dimensional array relates to time. A second axis of the multi-dimensional array relates to a first coordinate value on a first spatial axis in space. A third axis of the multi-dimensional array relates to a second coordinate value on a second spatial axis in space. The second spatial axis intersects with the first spatial axis.

[0062] For example, the time-series image data includes a value related to brightness at a position corresponding to the first coordinate value and the second coordinate value. For example, the time-series image data may include a value related to color at a position corresponding to the first coordinate value and the second coordinate value.

[0063] The first feature information Pa1 may include, for example, persistent feature values ​​in a persistent map PD. The persistent map PD corresponds to a distribution state of at least a part of the first data Da1. The persistent map PD is derived based on a multidimensional array.

[0064] The embodiment may include the following configurations (e.g., technical solutions). (Configuration 1) An acquisition unit capable of acquiring first data including time-series image data; A data processing device comprising: a processing unit capable of deriving first feature information based on an n-dimensional (n is an integer greater than or equal to 3) multidimensional array based on the first data acquired by the acquisition unit, wherein a first axis of the multidimensional array relates to time.

[0065] (Configuration 2) a second axis of the multi-dimensional array relating to a first coordinate value on a first spatial axis in space; 2. The data processing apparatus of claim 1, wherein a third axis of the multi-dimensional array relates to a second coordinate value relative to a second spatial axis in the space, the second spatial axis intersecting the first spatial axis.

[0066] (Configuration 3) 3. The data processing device according to configuration 2, wherein the time-series image data includes values ​​related to brightness at positions corresponding to the first coordinate value and the second coordinate value.

[0067] (Configuration 4) 4. The data processing device according to configuration 2 or 3, wherein the time-series image data includes values ​​related to colors at the positions corresponding to the first coordinate values ​​and the second coordinate values.

[0068] (Configuration 5) 5. The data processing device according to any one of configurations 1 to 4, wherein the first feature information includes a persistent feature value in a persistent diagram that corresponds to a distribution state of at least a part of the first data and is derived based on the multidimensional array.

[0069] (Configuration 6) 6. The data processing apparatus of configuration 5, wherein the persistence diagram corresponds to one of the first data divided into a plurality of time periods.

[0070] (Configuration 7) 7. The data processing apparatus of configuration 5 or 6, wherein the persistence diagram includes zero-order persistence.

[0071] (Configuration 8) 8. The data processing device according to any one of configurations 5 to 7, wherein the first feature information includes a feature vector obtained by converting the persistent feature value.

[0072] (Configuration 9) The data processing device according to any one of configurations 1 to 8, wherein the processing unit is capable of outputting first feature information about the first data based on a result of comparing the persistent feature value with a first reference value related to the persistent feature value.

[0073] (Configuration 10) 2. The data processing device according to claim 1, wherein the processing unit is capable of extracting a characteristic portion of the first data based on machine learning using the first feature information as an input.

[0074] (Configuration 11) The data processing device is capable of providing a control signal to a target device based on the first characteristic information; 11. The data processing device according to any one of configurations 1 to 10, wherein the first data relates to the target device.

[0075] (Configuration 12) Acquire first data including time-series image data; A data processing method comprising: deriving first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data, the first axis of the multidimensional array relating to time.

[0076] (Configuration 13) a second axis of the multi-dimensional array relating to a first coordinate value on a first spatial axis in space; 13. The data processing method of claim 12, wherein a third axis of the multidimensional array relates to a second coordinate value relative to a second spatial axis in the space, the second spatial axis intersecting the first spatial axis.

[0077] (Configuration 14) 14. The data processing method according to claim 13, wherein the time-series image data includes values ​​related to brightness at positions corresponding to the first coordinate value and the second coordinate value.

[0078] (Configuration 15) 15. The data processing method according to configuration 13 or 14, wherein the time-series image data includes values ​​related to color at the positions corresponding to the first coordinate value and the second coordinate value.

[0079] (Configuration 16) The data processing method according to any one of configurations 12 to 15, wherein the first feature information includes a persistent feature value in a persistent diagram that corresponds to a distribution state of at least a portion of the first data and is derived based on the multidimensional array.

[0080] (Configuration 17) On the computer, Acquire first data including time-series image data; A data processing program for deriving first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data, the first axis of the multidimensional array relating to time.

[0081] (Configuration 18) a second axis of the multi-dimensional array relating to a first coordinate value on a first spatial axis in space; 18. The data processing program of claim 17, wherein a third axis of the multidimensional array relates to a second coordinate value relative to a second spatial axis in the space, the second spatial axis intersecting the first spatial axis.

[0082] (Configuration 19) 19. The data processing program according to claim 18, wherein the time-series image data includes values ​​related to brightness at positions corresponding to the first coordinate value and the second coordinate value.

[0083] (Configuration 20) 20. The data processing program according to any one of configurations 17 to 19, wherein the first feature information includes a persistent feature value in a persistent diagram that corresponds to a distribution state of at least a portion of the first data and is derived based on the multidimensional array.

[0084] According to the embodiments, it is possible to provide a data processing device, a data processing method, and a data processing program that can improve accuracy.

[0085] The above describes the embodiment of the present invention with reference to specific examples. However, the present invention is not limited to these specific examples. For example, the specific configuration of each element, such as an acquisition unit and a processing unit, included in a data processing device is included in the scope of the present invention as long as a person skilled in the art can implement the present invention in the same way and obtain the same effect by appropriately selecting from the known range.

[0086] Furthermore, any combination of two or more elements of each of the specific examples, within the scope of technical feasibility, is also included within the scope of the present invention as long as it includes the gist of the present invention.

[0087] In addition, all data processing devices, data processing methods, and data processing programs that can be implemented by a person skilled in the art through appropriate design modifications based on the data processing device, data processing method, and data processing program described above as embodiments of the present invention also fall within the scope of the present invention, so long as they include the gist of the present invention.

[0088] In addition, within the scope of the concept of the present invention, a person skilled in the art may conceive of various modifications and alterations, and it will be understood that these modifications and alterations also fall within the scope of the present invention.

[0089] Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, and are included in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0090] 31...acquisition unit, 32...processing unit, 33...storage unit, 34...display unit, 35...output unit, 40...target device, 110...data processing device, DA1-DAim...data, DC1...data, Da1...first data, F1, F2...first and second parameters, G1...group, G2...value, P1...characteristic function, Pa1...first feature information, R1...reference value, SP1-SP5...first to fifth sample data, Sig1...control signal

Claims

1. an acquisition unit capable of acquiring first data including time-series image data; a processing unit capable of deriving first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data acquired by the acquisition unit, wherein a first axis of the multidimensional array relates to time; the first feature information includes a persistent feature value in a persistent diagram that corresponds to a distribution state of at least a portion of the first data and is derived based on the multidimensional array; The data processing device, wherein the processing unit is capable of outputting first feature information about the first data based on a result of comparing the persistent feature value with a first reference value related to the persistent feature value.

2. a second axis of the multi-dimensional array relating to a first coordinate value on a first spatial axis in space; 2. The data processing apparatus of claim 1, wherein a third axis of said multi-dimensional array relates to a second coordinate value relative to a second spatial axis in said space, said second spatial axis intersecting said first spatial axis.

3. The data processing device according to claim 2 , wherein the time-series image data includes values ​​relating to brightness at positions corresponding to the first coordinate value and the second coordinate value.

4. The data processing device according to claim 3 , wherein the time-series image data includes values ​​relating to colors at the positions corresponding to the first coordinate value and the second coordinate value.

5. An acquisition unit capable of acquiring first data including time-series image data; a processing unit capable of deriving first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data acquired by the acquisition unit, wherein a first axis of the multidimensional array relates to time; The data processing device, wherein the processing unit is capable of extracting characteristic portions of the first data based on machine learning using the first feature information as input.

6. An acquisition unit capable of acquiring first data including time-series image data; a processing unit capable of deriving first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data acquired by the acquisition unit, wherein a first axis of the multidimensional array relates to time; A control signal is provided to a target device based on the first characteristic information; The first data relates to the target device.

7. Acquire first data including time-series image data; Derive first feature information based on an n-dimensional (n is an integer equal to or greater than 3) multidimensional array based on the first data, a first axis of the multidimensional array relating to time, the first feature information includes a persistent feature value in a persistent diagram that corresponds to a distribution state of at least a portion of the first data and is derived based on the multidimensional array; A data processing method comprising: outputting first feature information about the first data based on a result of comparing the persistent feature value with a first reference value related to the persistent feature value.

8. On the computer, Acquiring first data including time-series image data; Derive first feature information based on an n-dimensional (n is an integer of 3 or more) multidimensional array based on the first data, a first axis of the multidimensional array being related to time, the first feature information includes a persistent feature value in a persistent diagram that corresponds to a distribution state of at least a portion of the first data and is derived based on the multidimensional array; a first reference value for the persistent feature value and a result of comparing the persistent feature value with each other, the first feature information for the first data being output based on the result of the comparison;

Images