JP5417967B2 - Genetic processing apparatus, genetic processing method and program - Google Patents

Description

  The present invention relates to a genetic processing apparatus, a genetic processing method, and a program.

  A filter string generation method using evolutionary computation such as a genetic algorithm or genetic programming is known. In this method, operations such as crossover, mutation, and selection are repeated a plurality of times for a filter row to generate a new filter row. According to such a filter string generation method using evolutionary computation, a filter string having a complicated structure that is optimal for each case and difficult to obtain analytically is designed with less effort. be able to.

Masayoshi Maebuchi and two others, “Study on Image Filter Design Using Genetic Algorithm” [online], Computer Utilization Education Council, [March 20, 2008 search], Internet <URL: http: //www.ciec. or.jp/event/2003/papers/pdf/E00086.pdf>

  By the way, when each filter component constituting the filter has parameters, adjusting the parameters of such filter components by evolutionary calculation requires the use of a plurality of filter components having different parameters. This results in a wasteful increase in the number of genes produced.

In order to solve the above-described problem, in the first aspect of the present invention, genetic processing is performed from a filter group including at least one filter obtained by combining a plurality of filter components that convert input data and output as output data. A generation unit that generates a new filter and adds it to the filter group, and a fitness calculation unit that calculates a fitness for conversion from the learning input data to the learning target data for each filter included in the filter group, and based of the goodness of fit for the filter, and a selection unit for selecting a filter to leave the filter group, the parameters of the filter components that are included in the new filter, and an adjustment unit that adjusts the genetic process, new filter Are included in the two or more filter parts connected by the connection between the input and output, the adjustment unit As a fixed parameter of the portion of the filter components of the filter components, genetic processing apparatus for adjusting the parameters of the other part of the filter components, as well as to provide a genetic processing method and a program related to the genetic apparatus .
In the second aspect of the present invention, a new filter is generated by genetic processing from a filter group including at least one filter in which a plurality of filter components that convert input data and output as output data are combined. For each filter included in the filter group, a fitness calculation unit that calculates a fitness for conversion from learning input data to learning target data for each filter included in the filter group, and each of the filters A selection unit that selects the filter to be left in the filter group on the basis of the fitness of the filter, and an adjustment unit that adjusts parameters of filter components included in the new filter by genetic processing, The first filter including two or more identical filter components connected by connection between the input and output; A different number of the same filter parts as the first filter are connected to the part corresponding to the two or more filter parts in one filter, and the other part is a second filter having the same structure as the first filter. When included in a filter group, a genetic processing device that deletes one of the first filter and the second filter without selecting one, and a genetic processing method and program related to the genetic processing device are provided .

  It should be noted that the above summary of the invention does not enumerate all the necessary features of the present invention. In addition, a sub-combination of these feature groups can also be an invention.

FIG. 1 shows a configuration of a genetic processing apparatus 100 according to the present embodiment. FIG. 2 shows an example of the filter 20 having a configuration in which the filter components 22 according to this embodiment are combined in series. FIG. 3 shows an example of the filter 20 having a configuration in which the filter component 22 according to the present embodiment is combined in a tree structure. FIG. 4 shows an example of genetic operations performed on the filter 20 having a configuration in which the filter components 22 according to this embodiment are combined in series. FIG. 5 shows an example of a crossover operation performed on the filter 20 having a configuration in which the filter component 22 according to this embodiment is combined in a tree structure. FIG. 6 shows an example of a mutation operation performed on the filter 20 having a configuration in which the filter component 22 according to the present embodiment is combined in a tree structure. FIG. 7 illustrates an example of a similarity calculation method that is an example of a parameter that represents the degree of matching of the filter 20 when the filter 20 according to the present embodiment is an image filter. FIG. 8 shows an example of a gene composed of parameters of the filter component according to the present embodiment. FIG. 9 shows an example of parameter adjustment of the filter component according to this embodiment. FIG. 10 shows an example of filter deletion by the selection unit 150 according to the present embodiment. FIG. 11 shows an example of filter component integration by the integration unit 160 and division of the filter component by the division unit 180 according to the present embodiment. FIG. 12 shows a processing flow of the genetic processing apparatus 100 according to the present embodiment. FIG. 13 shows an example of the processing flow of the genetic processing apparatus 100 in step S1230 of FIG. FIG. 14 shows an example of a hardware configuration of a computer 1900 according to this embodiment.

  Hereinafter, the present invention will be described through embodiments of the invention, but the following embodiments do not limit the invention according to the claims. In addition, not all the combinations of features described in the embodiments are essential for the solving means of the invention.

  FIG. 1 shows a configuration of a genetic processing apparatus 100 according to the present embodiment. The genetic processing apparatus 100 includes a filter storage unit 110, a generation unit 120, a learning data storage unit 130, a fitness calculation unit 140, a selection unit 150, an integration unit 160, an iterative processing unit 170, and a division. Unit 180 and adjustment unit 190. The filter storage unit 110 stores a plurality of different filters that have already been generated.

  The filter may include a plurality of filter components that convert an input data set including one or more input data into an output data set including one or more output data. The filter may have a configuration in which filter components are combined in series. Further, the filter may have a configuration in which filter parts are combined in a tree structure or the like. In addition, the filter may be a program that performs an operation on the data set, an arithmetic expression that represents the content of the operation to be performed on the data set, or the like.

  The data set to be converted by the filter may be a one-dimensional data string, a two-dimensional data group, a three-dimensional data group, or a further multidimensional data group. The one-dimensional data string is, for example, time series data or an array data string. The two-dimensional data group is image data in which a plurality of pixel data and the like are arranged in a two-dimensional space. The three-dimensional data group is volume data or the like in which data values representing color or density are arranged at each grid point in the three-dimensional space. The filter may output a data set having a dimension different from that of the input data set.

  The generation unit 120 generates a new filter by genetic processing from a filter group including at least one filter obtained by combining a plurality of filter components that convert input data and output as output data. The genetic process will be described later with reference to FIGS. The generation unit 120 may generate a new filter from the filter group by genetic processing and add it to the filter group.

  The learning data storage unit 130 stores learning data. The learning data includes learning input data and learning target data. The learning target data is data that is a target of data obtained by the filter (converter) converting the learning input data. In the present embodiment, the learning input data may be a learning input image, and the learning target data may be a learning target image. For example, the learning input image may be an inspection target image obtained by imaging an inspection target including a defect, and the learning target image is an extraction image of the defect that is a target to be extracted from the inspection target image. Also good.

  The goodness-of-fit calculation unit 140 calculates the goodness of fit for the conversion from the learning input data to the learning target data for each filter included in the filter group. Here, the fitness is an index value indicating whether or not the learning input data is suitable for conversion into learning target data. In this embodiment, the higher the value, the more the learning input data is represented by the learning target data. Indicates that it is suitable for conversion to data. In one example, the fitness level calculation unit 140 calculates the similarity level of the learning input image to the learning target image, which will be described later with reference to FIG. 7, as the fitness level.

  The selection unit 150 selects a filter to be left in the filter group based on the degree of fitness for each filter. The integration unit 160 integrates two or more filter components connected by connection between the input and output, and replaces them with an integrated filter component that performs conversion corresponding to conversion by the two or more filter components. Here, the process corresponding to the conversion by the two or more filter components is the same or similar output as the output data obtained by processing the input data by the two or more filter components when input data is given. This is a process for obtaining data. The integration unit 160 may replace two or more filter components with a filter component that performs processing that is a combination of processing by the two or more filter components.

  The iterative processing unit 170 generates a new filter by the generation unit 120, calculates a fitness level by the fitness level calculation unit 140, selects a filter to be left in the next generation filter group by the selection unit 150, and filter components by the integration unit 160 Repeat the integration process.

  The dividing unit 180 divides the filter component in at least one filter into two or more filter components. The iterative processing unit 170 performs division when it is necessary to divide the filter component when two or more filters are crossed between integration of the filter component by the integration unit 160 and generation of a new filter by the generation unit 120. The filter component may be divided by the unit 180. In that case, the dividing unit 180 divides at least one filter into two or more partial filters by dividing the filter component in at least one filter into two or more filter components, and the generation unit 120 has two or more parts. Some of the filters may be at least part of the new filter.

  The adjustment unit 190 adjusts the parameters of the filter components included in the new filter by genetic processing. In one example, the adjustment unit 190 adjusts the parameter of the filter component by genetic processing for the adjustment target filter obtained as a result of the repetition processing by the repetition processing unit 170 among at least one new filter. Here, the adjustment target filter may be one or a plurality of filters selected last by the selection unit 150. The parameters of the filter component include, for example, the reference luminance used in one of the image filters (Dilate filter), the number of pixels in the range used in the median filter, and adoption of maximum and minimum luminance (Adoption). The number of pixels in the range used in the filter, the multiplier of the multiplication filter, the threshold value of the maximum value filter, the threshold value of the minimum value filter, or the threshold value of the binarization filter. In this way, the generation unit 120 does not need to use a plurality of filter components having only different parameters each time the iterative processing is performed by the iterative processing unit 170, thereby reducing the calculation cost.

  FIG. 2 shows an example of the filter 20 having a configuration in which the filter components 22 are combined in series. FIG. 3 shows an example of the filter 20 having a configuration in which the filter component 22 is combined in a tree structure. The filter 20 receives input image data, performs a filter operation process on the received input image data, and outputs output data.

  The filter 20 has a configuration in which a plurality of filter components 22 are combined. As shown in FIG. 2, the filter 20 may have a configuration in which filter components 22 are combined in series. Moreover, the filter 20 may have a configuration in which the filter component 22 is combined in a tree structure, as shown in FIG.

  The filter 20 having the structure in which the filter parts 22 are combined in a tree structure is supplied with input image data to the filter part 22 at the end of the tree structure, and outputs output data from the highest-order filter part 22 in the tree structure. Further, in such a filter 20, the same input data is given to each of a plurality of terminal filter parts 22. Alternatively, in such a filter 20, different input data may be given to each of the plurality of end filter components 22.

  Each of the plurality of filter components 22 may be a program module, an arithmetic expression, or the like. The filter component 22 receives the data output from the filter component 22 arranged in the preceding stage, performs an operation on the received data, and gives it to the filter component 22 arranged in the subsequent stage.

  Each of the plurality of filter components 22 is a unary operation such as a binarization operation, a histogram operation, a smoothing operation, an edge detection operation, a morphology operation and / or an operation in a frequency space (for example, a low-pass filtering operation and a high-pass filtering operation). It may be an image filter that performs. Further, each of the plurality of filter components 22 includes an average operation, a difference operation, and / or a fuzzy operation (for example, logical sum operation, logical product operation, algebraic sum, algebraic product, limit sum, limit product, intense sum and intense product). Binary operations such as these may be performed.

  FIG. 4 shows an example of a genetic operation performed on the filter 20 having a configuration in which the filter components 22 are combined in series. FIG. 5 shows an example of a crossover operation performed on the filter 20 having a configuration in which the filter component 22 is combined in a tree structure. FIG. 6 shows an example of a mutation operation performed on the filter 20 having a configuration in which the filter component 22 is combined in a tree structure.

  The generation unit 120 performs a crossover operation as an example of a genetic operation on the two or more filters 20 to generate two or more new filters 20. As illustrated in FIGS. 4 and 5, the generation unit 120 converts a part of the filter component group 24A of at least one filter 20A that has already been generated into a part of at least a part of another filter 20B that has already been generated. New filters 20E and 20F may be generated by replacing the filter component group 24B. The filter component group 24 is a member in which one or a plurality of filter components 22 are combined.

  The generation unit 120 performs a mutation operation as an example of a genetic operation on one filter 20 to generate a new one filter 20. As shown in FIGS. 4 and 6, the generation unit 120 replaces a part of the filter component group 24C of the already generated filter 20C with another filter component group 24G selected at random, A new filter 20G may be generated.

  The generation unit 120 may leave the current generation filter 20 as the next generation filter 20 as it is. As illustrated in FIG. 4, the generation unit 120 may generate a next-generation filter 20H that directly includes the configuration of the filter component 22 of the filter 20D.

  The selection unit 150 selects one or a plurality of filters 20 by a technique in which a natural selection of living organisms is modeled with respect to the plurality of filters 20 generated by the generation unit 120. The selection unit 150 preferentially selects a filter 20 having a higher matching degree among the plurality of filters 20. The selection unit 150 may select the filter 20 according to a technique such as elite selection and roulette selection based on the degree of fitness of each of the plurality of filters 20. Then, the selection unit 150 stores the filter 20 in the filter storage unit 110 so that the selected filter 20 can survive to the next generation, and deletes the filter 20 that has not been selected from the filter storage unit 110 in order to kill it.

  FIG. 7 shows an example of a similarity calculation method that is an example of a parameter that represents the degree of fitness of the filter 20 when the filter 20 is an image filter. In the present embodiment, the fitness level calculation unit 140 calculates a similarity level indicating how similar the output image generated by converting the input image by the filter 20 and the target image are. This similarity is an example of a parameter representing the degree of matching of the filter 20, and is used as an evaluation value or an index indicating that the larger the value is, the more appropriate the filter 20 that generated the output image is.

  The fitness calculation unit 140 calculates a comparison value by comparing the value of each region of the output image with the value of each region corresponding to the target image, and sets the weight specified by the weight data for the comparison value for each region. Multiply. Then, the degree-of-fit calculation unit 140 calculates a value obtained by averaging or summing the comparison values for each region multiplied by the weight for all regions, and outputs the calculated value as the similarity of the output image.

  As an example, the fitness level calculation unit 140 may calculate the difference or ratio between the luminance value of each pixel of the output image and the luminance value of the corresponding pixel of the target image. Then, as an example, the fitness calculation unit 140 multiplies each calculated difference or ratio for each pixel by the corresponding weight specified by the weight data, and sums the difference or ratio for each pixel multiplied by the weight. On average, the similarity may be calculated.

  For example, when the weight data is represented by a weight image, the fitness level calculation unit 140 may calculate the similarity level by performing an operation represented by the following formula (1). In this case, the sizes of the output image, the target image, and the weight image are the same.

  In the formula (1), f represents the similarity. Ymax represents the maximum luminance value.

  X is a variable indicating the pixel position in the horizontal direction of the image. y is a variable indicating the pixel position in the vertical direction of the image. xmax is a constant indicating the number of pixels in the horizontal direction of the image. ymax is a constant indicating the number of pixels in the vertical direction of the image.

  Iweight (x, y) represents the luminance value of the pixel at the (x, y) position in the weighted image. Itarget (x, y) represents the luminance value of the pixel at the (x, y) position in the target image. Ifilter (x, y) represents the luminance value of the pixel at the (x, y) position in the output image.

  That is, as shown in the numerator part in the curly braces of Expression (1), the fitness calculation unit 140 calculates the luminance value of the weighted image with respect to the absolute value of the difference between the luminance value of the target image and the luminance value of the output image. Is calculated for every pixel in the screen, and a total weighted difference value is calculated by adding the calculated weighted difference values for all pixels. Further, as shown in the denominator part in the curly braces of Expression (1), the fitness level calculation unit 140 calculates a total weight obtained by adding the luminance values of the weighted image for all pixels. Further, the fitness calculation unit 140 normalizes the division value (in the curly braces of the expression (1)) obtained by dividing the total weighted difference value by the total weight by the reciprocal (1 / Ymax) of the maximum value of the luminance value. The value (the second term in equation (1)) is calculated. Then, the fitness calculation unit 140 calculates a value obtained by subtracting the normalized value from 1 as the similarity f. In this way, the fitness level calculation unit 140 can calculate the similarity level by weighting the difference between the output image and the target image for each region.

  FIG. 8 shows an example of a gene used when adjusting the parameters of the filter component. Filter 810 has filter components 812, 814, 816, and 818. The filter component 812 has a parameter Pa, the filter component 814 has parameters (Pb, Pc), the filter component 816 has a parameter Pe, and the filter component 818 has a parameter Pd.

  The values of these parameters Pa, Pb, Pc, Pd, and Pe are arranged in a line, for example, as a gene 820. The adjustment unit 190 can generate a filter with a higher degree of fitness by repeating the same genetic operation as in FIGS. 4 to 6 for the gene thus obtained. Note that the length of the gene is kept constant when performing genetic manipulation.

  FIG. 9 shows an example of filter component parameter adjustment. The filter 910 includes a filter component 912 having a parameter P, a filter component 914, and a filter component 916. The values of the parameters P are all equal to 1. When the new filter includes the same two or more filter parts connected by connection between the input and output, the adjustment unit 190 fixes the parameters of some of the two or more filter parts as fixed, and others The parameters of some of the filter components may be adjusted. In addition, when the new filter includes two or more identical filter components connected by connection between input and output, the adjustment unit 190 fixes parameters other than the one filter component among the two or more filter components as fixed. The parameter of one filter component may be adjusted.

  For example, the adjustment unit 190 adjusts the filter 910 to the filter 920 by adjusting the value of the parameter P of the filter component 916 in the filter 910 to 2. Here, the filter component 912 and the filter component 922 are the same filter component, and the filter component 914 and the filter component 924 are the same filter component. Further, when the filter has n stages of the same filter parts having parameters, the filter parameters of one of the filter parts of the filter are set to n-1 to n + 1 or n-0.5 to n + 0.5 stages. You may adjust in the range. In this way, the adjustment unit 190 can adjust the strength of the filter with higher resolution as compared with the case where the strength of the filter is adjusted by the number of stages of the filter components.

  FIG. 10 shows an example of filter deletion by the selection unit 150. The first filter 1010 has the same filter components 1012A, 1012B, and 1012C. The second filter 1020 has a filter component 1012D identical to one of those filter components. The structure of the first filter 1010 and the structure of the second filter 1020 are the same except for the filter components 1012A, 1012B, 1012C, and 1012D.

  In one example, the selection unit 150 includes a first filter 1010 including two or more identical filter components coupled by connection between input and output, and a first portion corresponding to the two or more filter components in the first filter 1010. When a different number of the same filter parts are connected to the filter, and the second filter 1020 having the same structure as that of the first filter 1010 is included in the filter group, the first filter 1010 and the second filter 1020 are included. Delete one of these without selecting it. In this case, the selection unit 150 may delete the first filter 1010 having a more complicated structure without selecting, for example. Instead of this, the selection unit 150 may delete at least one of the plurality of filters in which the number of the same filter components is within a predetermined range. The selection unit 150 may delete a plurality of filters at the same time.

  FIG. 11 shows an example of filter component integration by the integration unit 160 and division of the filter component by the division unit 180 according to the present embodiment. Within the filter 1110, the same filter components, indicated by black circles, are repeatedly connected as 1114A and 1114B. When the integration unit 160 integrates the same filter components in the filter 1110, the filter 1110 is converted into a filter 1120. Within the filter 1120, the filter components 1114A and 1114B are integrated into the filter component 1114C having a changed strength. Further, the filter component 1112, the filter component 1116, and the filter component 1118 in the filter 1110 are not changed.

  Subsequently, when the dividing unit 180 divides the filter component 1114 </ b> C in the filter 1120, the filter 1120 is converted into a filter 1130. In the filter 1130, the filter component 1114C is divided into filter components 1114D, 1114E, and 1114F whose strengths are changed. Here, the dividing unit 180 may divide all of the filter components integrated by the integrating unit 160 or may divide a part thereof. Subsequently, the adjustment unit 190 may adjust the parameters of the integrated filter component integrated by the integration unit 160. By doing so, the adjustment unit 190 can adjust the parameter of the filter component for the filter having a shorter length.

  FIG. 12 shows a processing flow of the genetic processing apparatus 100. The genetic processing device 100 repeatedly executes the processes in steps S1210 to S1250 and step S1270 a plurality of times (for example, a plurality of generations). Note that the genetic processing device 100 uses a filter storage unit that generates a filter that is randomly generated or has been generated in advance as an initial filter for generating a filter that converts given learning input data into learning target data. 110 is stored.

  In each generation, first, the generation unit 120 performs a genetic operation such as a crossover operation and a mutation operation on a plurality of filters remaining from the previous generation to generate a plurality of new filters (S1210). . In the first generation, the generation unit 120 may generate a plurality of new filters by performing a genetic operation on the plurality of filters generated in advance by the user or the like.

  Subsequently, the fitness level calculation unit 140 filters the learning input data with each of the plurality of filters remaining from the previous generation and the plurality of filters newly generated in step S1210 to generate output data (S1220). ).

  Subsequently, the fitness level calculation unit 140 calculates the fitness level of each output data and the learning target data, and the selection unit 150 selects a plurality of filters corresponding to the output data having a higher fitness level (S1230). ). Note that, in the last generation, the selection unit 150 may select one filter corresponding to one output data having the highest fitness.

  Subsequently, the selection unit 150 causes the one or more filters selected in step S1230 to remain in the next generation, and deletes the filter that generated the output data not selected in step S1230 (S1240). Subsequently, the integration unit 160 integrates the filter components 22 in the filter obtained in step S1260 by replacement (S1250). For example, the integration unit 160 may replace n connected expansion / contraction filters with n-fold expansion / contraction filters. For example, when n blur filters, median filters, average filters, histogram expansion / compression filters, and the like are connected, the integration unit 160 replaces the n filters with a filter that performs n times the intensity. Also good. In this way, the integration unit 160 can reduce the number of filter components 22 by replacing the same filter component 22 with one filter component 22 that provides the same or similar effect with respect to input data, thereby reducing the calculation cost. Decrease.

  The iterative processing unit 170 determines whether the above processing has been executed up to the Nth generation (N is an integer of 2 or more, for example, several tens or several hundreds) (S1260). When the above processing has not been executed up to the Nth generation, the dividing unit 180 divides the integrated filter component in at least one filter into a set of two or more filter components (S1270). The dividing unit 180 may divide the integrated filter component into a set of two or more filter components that perform the same conversion as the integrated filter component. For example, the dividing unit 180 is the same having an intensity of a times and b times, where n = a + b, where n = a + b is an integrated filter component such as a blur filter, a median filter, an average filter, and a histogram expansion / compression filter having an intensity of n times. It may be replaced with synthesis of filter parts that perform conversion.

  In addition, when the generation unit 120 divides the filter component 22 for the purpose of causing the generation unit 120 to cross the plurality of filters 20, the generation unit 120 may instruct the division unit 180 to divide the filter component 22. Good. Accordingly, the dividing unit 180 can perform division suitable for the generating unit 120 to perform various genetic operations.

  When the above processing is executed up to the Nth generation, the genetic processing device 100 exits the flow (S1260).

  Subsequently, the adjustment unit 190 adjusts the parameters of the filter components in the filter obtained when the genetic processing apparatus 100 exits the flow (S1280). Note that step S1280 may be executed when the filter obtained when exiting the flow is actually applied.

  FIG. 13 shows an example of the processing flow of the genetic processing apparatus 100 in step S1230 of FIG. The genetic processing device 100 executes the following processing in step S1230 shown in FIG.

  First, the fitness level calculation unit 140 executes the following step S1320 for each of the plurality of output data generated by each of the plurality of filters stored in the filter storage unit 110 (S1310, S1330). In step S1320, the fitness level calculation unit 140 calculates the fitness level between the output data and the learning target data. When the fitness level calculation unit 140 finishes processing for all of the plurality of output data, the fitness level calculation unit 140 advances the processing to step S1340 (S1330).

  Subsequently, the selection unit 150 selects output data closer to the learning target data from among the plurality of output data based on the degree of matching of each of the plurality of output data (S1340). As an example, the selection unit 150 may preferentially select output data having a higher matching degree from among a plurality of output data converted by a plurality of filters. For example, the selection unit 150 may select output data having a higher fitness than the standard fitness. For example, the selection unit 150 may select output data in a range in which the fitness is determined in advance from the top. In addition, the selection unit 150 may select output data randomly under a condition in which output data having a higher degree of matching is selected with a higher probability.

  Subsequently, the selection unit 150 selects the filter that generates the output data selected by the selection unit 150 as a filter that converts the input data for learning into data that matches the target data for learning (S1350). When the process of step S1350 is completed, the genetic processing apparatus 100 ends the flow.

  FIG. 14 shows an example of a hardware configuration of a computer 1900 according to the embodiment of the present invention. A computer 1900 according to this embodiment is connected to a CPU peripheral unit having a CPU 2000, a RAM 2020, a graphic controller 2075, and a display device 2080 that are connected to each other by a host controller 2082, and to the host controller 2082 by an input / output controller 2084. An input / output unit having a communication interface 2030, a hard disk drive 2040, and a DVD drive 2060, and a legacy input / output unit having a ROM 2010, a flexible disk drive 2050, and an input / output chip 2070 connected to the input / output controller 2084. Prepare.

  The host controller 2082 connects the RAM 2020 to the CPU 2000 and the graphic controller 2075 that access the RAM 2020 at a high transfer rate. The CPU 2000 operates based on programs stored in the ROM 2010 and the RAM 2020 and controls each unit. The graphic controller 2075 acquires image data generated by the CPU 2000 or the like on a frame buffer provided in the RAM 2020 and displays it on the display device 2080. Instead of this, the graphic controller 2075 may include a frame buffer for storing image data generated by the CPU 2000 or the like.

  The input / output controller 2084 connects the host controller 2082 to the communication interface 2030, the hard disk drive 2040, and the DVD drive 2060, which are relatively high-speed input / output devices. The communication interface 2030 communicates with other devices via a network. The hard disk drive 2040 stores programs and data used by the CPU 2000 in the computer 1900. The DVD drive 2060 reads a program or data from the DVD 2095 and provides it to the hard disk drive 2040 via the RAM 2020.

  The input / output controller 2084 is connected to the ROM 2010, the flexible disk drive 2050, and the relatively low-speed input / output device of the input / output chip 2070. The ROM 2010 stores a boot program that the computer 1900 executes at startup and / or a program that depends on the hardware of the computer 1900. The flexible disk drive 2050 reads a program or data from the flexible disk 2090 and provides it to the hard disk drive 2040 via the RAM 2020. The input / output chip 2070 connects the flexible disk drive 2050 to the input / output controller 2084 and inputs / outputs various input / output devices via, for example, a parallel port, a serial port, a keyboard port, a mouse port, and the like. Connect to controller 2084.

  A program provided to the hard disk drive 2040 via the RAM 2020 is stored in a recording medium such as the flexible disk 2090, the DVD 2095, or an IC card and provided by the user. The program is read from the recording medium, installed in the hard disk drive 2040 in the computer 1900 via the RAM 2020, and executed by the CPU 2000.

  A program that is installed in the computer 1900 and causes the computer 1900 to function as the genetic processing apparatus 100 according to the first embodiment includes a filter storage module, a generation module, a learning data storage module, a fitness calculation module, a selection module, an integration module, It includes a repetitive processing module, a division module, and an adjustment module. These programs or modules work on the CPU 2000 or the like to make the computer 1900 into a filter storage unit 110, a generation unit 120, a learning data storage unit 130, a fitness calculation unit 140, a selection unit 150, an integration unit 160, and an iterative processing unit. 170, the dividing unit 180, and the adjusting unit 190 function.

  The information processing described in these programs is read by the computer 1900, so that the filter storage unit 110, the generation unit 120, and the learning unit which are specific means in which the software and the various hardware resources described above cooperate with each other. It functions as a data storage unit 130, a fitness calculation unit 140, a selection unit 150, an integration unit 160, an iterative processing unit 170, a division unit 180, and an adjustment unit 190. And the specific genetic processing apparatus 100 according to the intended use is constructed | assembled by implement | achieving the calculation or processing of the information according to the intended use of the computer 1900 in 1st Embodiment by these specific means.

  In an example, when communication is performed between the computer 1900 and an external device or the like, the CPU 2000 executes a communication program loaded on the RAM 2020, and based on the processing contents described in the communication program, the communication interface A communication process is instructed to 2030. Under the control of the CPU 2000, the communication interface 2030 reads transmission data stored in a transmission buffer area or the like provided on a storage device such as the RAM 2020, the hard disk drive 2040, the flexible disk 2090, or the DVD 2095, and transmits it to the network. Alternatively, the reception data received from the network is written into a reception buffer area or the like provided on the storage device. As described above, the communication interface 2030 may transfer transmission / reception data to / from the storage device by a DMA (direct memory access) method. Instead, the CPU 2000 transfers the storage device or the communication interface 2030 as a transfer source. The transmission / reception data may be transferred by reading the data from the data and writing the data to the communication interface 2030 or the storage device of the transfer destination.

  In addition, the CPU 2000 DMAs all or necessary portions from among files or databases stored in an external storage device such as the hard disk drive 2040, the DVD drive 2060 (DVD 2095), and the flexible disk drive 2050 (flexible disk 2090). The data is read into the RAM 2020 by transfer or the like, and various processes are performed on the data on the RAM 2020. Then, CPU 2000 writes the processed data back to the external storage device by DMA transfer or the like. In such processing, since the RAM 2020 can be regarded as temporarily holding the contents of the external storage device, in the present embodiment, the RAM 2020 and the external storage device are collectively referred to as a memory, a storage unit, or a storage device. Various types of information such as various programs, data, tables, and databases in the present embodiment are stored on such a storage device and are subjected to information processing. Note that the CPU 2000 can also store a part of the RAM 2020 in the cache memory and perform reading and writing on the cache memory. Even in such a form, the cache memory bears a part of the function of the RAM 2020. Therefore, in the present embodiment, the cache memory is also included in the RAM 2020, the memory, and / or the storage device unless otherwise indicated. To do.

  In addition, the CPU 2000 performs various operations, such as various operations, information processing, condition determination, information search / replacement, etc., described in the present embodiment, specified for the data read from the RAM 2020 by the instruction sequence of the program. Is written back to the RAM 2020. For example, when performing the condition determination, the CPU 2000 determines whether the various variables shown in the present embodiment satisfy the conditions such as large, small, above, below, equal, etc., compared to other variables or constants. When the condition is satisfied (or not satisfied), the program branches to a different instruction sequence or calls a subroutine.

  Further, the CPU 2000 can search for information stored in a file or database in the storage device. For example, in the case where a plurality of entries in which the attribute value of the second attribute is associated with the attribute value of the first attribute are stored in the storage device, the CPU 2000 displays the plurality of entries stored in the storage device. The entry that matches the condition in which the attribute value of the first attribute is specified is retrieved, and the attribute value of the second attribute that is stored in the entry is read, thereby associating with the first attribute that satisfies the predetermined condition The attribute value of the specified second attribute can be obtained.

  The program or module shown above may be stored in an external recording medium. As the recording medium, in addition to the flexible disk 2090 and the DVD 2095, an optical recording medium such as a CD, a magneto-optical recording medium such as an MO, a tape medium, a semiconductor memory such as an IC card, and the like can be used. Further, a storage device such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet may be used as a recording medium, and the program may be provided to the computer 1900 via the network.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. It will be apparent to those skilled in the art that various modifications or improvements can be added to the above-described embodiment. It is apparent from the scope of the claims that the embodiments added with such changes or improvements can be included in the technical scope of the present invention.

  The order of execution of each process such as operations, procedures, steps, and stages in the apparatus, system, program, and method shown in the claims, the description, and the drawings is particularly “before” or “prior to”. It should be noted that the output can be realized in any order unless the output of the previous process is used in the subsequent process. Regarding the operation flow in the claims, the description, and the drawings, even if it is described using “first”, “next”, etc. for convenience, it means that it is essential to carry out in this order. It is not a thing.

  20 filters, 22 filter parts, 24 filter parts group, 100 genetic processor, 110 filter storage part, 120 generation part, 130 learning data storage part, 140 fitness calculation part, 150 selection part, 160 integration part, 170 iteration Processing unit, 180 division unit, 190 adjustment unit, 810 filter, 812 filter component, 814 filter component, 816 filter component, 818 filter component, 820 gene, 910 filter, 912 filter component, 914 filter component, 916 filter component, 920 filter 1010 1st filter, 1012 filter component, 1020 2nd filter, 1110 filter, 1112 filter component, 1114 filter component, 1116 filter component, 1118 filter component, 1120 filter 1130 filter, 1900 computer, 2000 CPU, 2010 ROM, 2020 RAM, 2030 communication interface, 2040 hard disk drive, 2050 flexible disk drive, 2060 DVD drive, 2070 input / output chip, 2075 graphic controller, 2080 display device, 2082 Host controller, 2084 I / O controller, 2090 flexible disk, 2095 DVD

Claims (15)

A generation unit that generates a new filter by genetic processing from a filter group including at least one filter that combines a plurality of filter components that convert input data and output as output data; and
For each filter included in the filter group, a fitness calculation unit that calculates a fitness for conversion from learning input data to learning target data; and
A selection unit for selecting the filter to be left in the filter group based on the degree of matching for each of the filters;
An adjustment unit for adjusting parameters of filter components included in the new filter by genetic processing;
Equipped with a,
When the new filter includes the same two or more filter parts connected by connection between the input and output, the adjustment unit fixes parameters of some of the two or more filter parts, A genetic processing device that adjusts parameters of some other filter components . A generation unit that generates a new filter by genetic processing from a filter group including at least one filter that combines a plurality of filter components that convert input data and output as output data; and
For each filter included in the filter group, a fitness calculation unit that calculates a fitness for conversion from learning input data to learning target data; and
A selection unit for selecting the filter to be left in the filter group based on the degree of matching for each of the filters;
An adjustment unit for adjusting parameters of filter components included in the new filter by genetic processing;
Equipped with a,
The selection unit is the same as the first filter in a portion corresponding to the two or more filter components in the first filter, and a first filter including the same two or more filter components connected by connection between input and output. When a different number of filter parts are connected and the other group includes a second filter having the same structure as that of the first filter, the filter group includes one of the first filter and the second filter. Genetic processing device to delete without selecting . And further comprising: an iterative processing unit that repeatedly generates the new filter by the generating unit, calculates the fitness by the fitness calculating unit, and selects the filter to be left in the next generation filter group by the selecting unit. Item 3. The genetic processing device according to item 1 or 2.   The said adjustment part adjusts the parameter of the said filter component by a genetic process about the adjustment object filter obtained as a result of the repetition process by the said repetition process part among the said at least 1 new filter. Genetic processing equipment. When the new filter includes the same two or more filter parts connected by connection between the input and output, the adjustment unit fixes parameters other than one filter part among the two or more filter parts, The genetic processing apparatus according to claim 1 , wherein a parameter of the one filter component is adjusted. The selection unit is the same as the first filter in a portion corresponding to the two or more filter components in the first filter, and a first filter including the same two or more filter components connected by connection between input and output. When a different number of filter parts are connected and the other group includes a second filter having the same structure as that of the first filter, the filter group includes one of the first filter and the second filter. The genetic processing device according to claim 1 or 5 , wherein the genetic processing device is deleted without selection. An integration unit that integrates two or more filter components connected by connection between the input and output and replaces the filter components with an integrated filter component that performs conversion corresponding to the conversion by the two or more filter components;
The genetic processing device according to any one of claims 1 to 6 , wherein the adjustment unit adjusts parameters of the integrated filter component integrated by the integration unit. The genetic processing device according to claim 7 , further comprising a dividing unit that divides a filter component in the at least one filter into two or more filter components. The genetic processing device according to claim 8 , wherein the division unit divides the integrated filter component in the at least one filter into a set of two or more filter components. The genetic processing device according to claim 9 , wherein the dividing unit divides the integrated filter component into a group of the two or more filter components that perform the same conversion as the integrated filter component together. The dividing unit divides the at least one filter into two or more partial filters by dividing the filter part in the at least one filter into the two or more filter parts,
The genetic processing device according to any one of claims 8 to 10 , wherein the generation unit uses a partial filter of the two or more partial filters as at least a part of the new filter. A step of generating a new filter by genetic processing from a filter group including at least one filter obtained by combining a plurality of filter components that convert input data and output as output data by the generation unit, and adding the filter to the filter group ; ,
Calculating a fitness for conversion from learning input data to learning target data for each filter included in the filter group by a fitness calculator;
Selecting the filters to be left in the filter group based on the fitness for each of the filters by a selection unit;
Adjusting the parameters of the filter components included in the new filter by genetic processing by the adjustment unit;
Equipped with a,
When the new filter includes two or more identical filter components connected by connection between input and output, the adjustment unit fixes parameters of some filter components among the two or more filter components, A genetic processing method for adjusting parameters of some other filter components . A step of generating a new filter by genetic processing from a filter group including at least one filter obtained by combining a plurality of filter components that convert input data and output as output data by the generation unit, and adding the filter to the filter group ; ,
Calculating a fitness for conversion from learning input data to learning target data for each filter included in the filter group by a fitness calculator;
Selecting the filters to be left in the filter group based on the fitness for each of the filters by a selection unit;
Adjusting the parameters of the filter components included in the new filter by genetic processing by the adjustment unit;
Equipped with a,
The first filter including the same two or more filter components connected by the connection between the input and output by the selection unit, and the portion corresponding to the two or more filter components in the first filter is the same as the first filter When a different number of filter parts are connected and the other group includes a second filter having the same structure as that of the first filter, one of the first filter and the second filter is included. Genetic processing method to delete without selecting . A program for causing a computer to function as a genetic processing device,
The computer,
A generation unit that generates a new filter by genetic processing from a filter group including at least one filter that combines a plurality of filter components that convert input data and output as output data; and
For each filter included in the filter group, a fitness calculation unit that calculates a fitness for conversion from learning input data to learning target data; and
A selection unit for selecting the filter to be left in the filter group based on the degree of matching for each of the filters;
An adjustment unit for adjusting parameters of filter components included in the new filter by genetic processing;
To function ,
When the new filter includes the same two or more filter parts connected by connection between the input and output, the adjustment unit fixes parameters of some of the two or more filter parts, A program that adjusts the parameters of some other filter components . A program for causing a computer to function as a genetic processing device,
The computer,
A generation unit that generates a new filter by genetic processing from a filter group including at least one filter that combines a plurality of filter components that convert input data and output as output data; and
For each filter included in the filter group, a fitness calculation unit that calculates a fitness for conversion from learning input data to learning target data; and
A selection unit for selecting the filter to be left in the filter group based on the degree of matching for each of the filters;
An adjustment unit for adjusting parameters of filter components included in the new filter by genetic processing;
To function ,
The selection unit is the same as the first filter in a portion corresponding to the two or more filter components in the first filter, and a first filter including the same two or more filter components connected by connection between input and output. When a different number of filter parts are connected and the other group includes a second filter having the same structure as that of the first filter, the filter group includes one of the first filter and the second filter. A program to delete without selecting .

Images