JP6783707B2 - Data number determination device, data number determination method and data number determination program - Google Patents

Description

The present invention relates to a data number determination device, a data number determination method, and a data number determination program.

Advances in network technology and sensing technology have made it possible to acquire diverse and enormous amounts of data at each sampling time. As one of the data that can be acquired for each sampling time, there is multidimensional time series data (hereinafter referred to as "spatial information data") including position information in space. The spatial information data includes, for example, time-series data including the position information of the mobile terminal as a sensing device and the statistical information of the user of the mobile terminal, the vehicle position information of the car navigation system as the sensing device, and the vehicle user. It is time-series data including statistical information, time-series data including position information of a ticket gate passed by a traffic IC card user and statistical information of a traffic IC card user.

Spatial information data is expected to be used for marketing, infrastructure development, urban development, disaster support, etc. When using in these fields, the meaning of the spatial information data is examined by analyzing the spatial information data with high accuracy according to the usage pattern. The data analyzer acquires and stores a large amount of spatial information data in order to analyze the spatial information data with high accuracy.

An increase in the amount of acquired spatial information data leads to an increase in the cost required for accumulating the spatial information data. Therefore, the accumulated spatial information data needs to be efficiently compressed. A data compression device, which is a device for compressing data, can utilize the correlation in the time direction in the spatial information data in order to efficiently compress the stored spatial information data.

Spatial information data may be represented by an image. For example, spatial information data may be represented by an image representing a map in which bar graphs are distributed. The data compressor models the temporal correlation of spatial information data based on an audio or image coding standard. The data compression device uses the modeled temporal correlation to compress spatial information data (see Non-Patent Document 1).

"H.265 / HEVC Textbook", Impress Japan Co., Ltd., pp. 23

However, the time-direction correlation used in audio or image frame compression techniques may not apply to the time-direction correlation of spatial information data. Therefore, the data compression device cannot efficiently compress the audio or image frame data by simply diverting the audio or image coding compression technique to the compression of spatial information data. Also, data compressors cannot take the initiative in future events.

In order for the data compressor to efficiently compress the data of audio or image frames and for the data compressor to take the lead in future events, the number of frames referred to by the data compressor when compressing the data is required. Need to be decided. That is, the conventional device needs to determine the number of data referenced by the data compression device. However, the conventional apparatus has a problem that the number of data to be referred to when compressing the data cannot be determined.

In view of the above circumstances, an object of the present invention is to provide a data number determination device, a data number determination method, and a data number determination program capable of determining the number of data to be referred to when compressing data. There is.

One aspect of the present invention is a separation that separates each of a plurality of past data included in the time series data, which is referred to for predicting the prediction target data to be predicted, into upper bit and lower bit data. a processing unit, and the number of historical data used for the processing, a prediction unit generating a prediction data for each combination of the respective data of the data and the lower bit of the upper bits after separation, the time correlation with the past data The process of calculating the difference between the predicted target data and the predicted data for at least the upper bit data and the lower bit data after separation is performed for each candidate for the number of past data used in the process. It is a data number determination device including a difference calculation unit and a data number determination unit that determines the number of past data based on the difference obtained for each of the number candidates.
In one aspect of the present invention, the upper bit and the lower bit are separated at a position defined by a boundary bit.
In one aspect of the present invention, the boundary bit is a bit that minimizes the prediction error, which is the sum of the differences between the separated signals obtained when the past data is separated by the candidate of the position to be separated.

One aspect of the present invention is the data number determination device, wherein the data number determination unit minimizes the sum of the differences between the data of the upper bit and the data of the lower bit after the separation. Determine the number of data.

One aspect of the present invention is the above-mentioned data number determination device, further including a separation position determination unit for determining a position for separating the past data, and the separation position determination unit is a candidate for a position to be separated and the past The separation position where the sum of the differences between the separated signals obtained when the data is separated is minimized is determined.

One aspect of the present invention is a data number determination method executed by a data number determination device that determines the number of data, in order to predict prediction target data that is included in time series data and is a prediction target. referenced, and separating process step of separating the plurality of historical data to the upper bit and the lower bit data, and the number of historical data used for the processing, and the data and the lower bit of the high order bit data respectively after separation The upper bit data and the lower bit data after at least separating the difference between the prediction step for generating the prediction data for each combination of the above, the prediction target data having a time correlation with the past data, and the prediction data, respectively. Data for determining the number of past data based on the difference calculation step for each candidate for the number of past data used in the process and the difference obtained for each candidate for the number of past data. It is a data number determination method having a number determination step.

One aspect of the present invention is to convert a plurality of past data, which are included in the time series data and are referred to for predicting the prediction target data to be predicted, into upper bit and lower bit data, respectively. a separation process step of separating, and the number of historical data used for the processing, a prediction step of generating a prediction data for each combination of the respective data of the data and the lower bit of the upper bits after separation, said past data Candidates for the number of past data used in the processing for calculating the difference between the prediction target data having a time correlation and the prediction data for at least the upper bit data and the lower bit data after separation. It is a data number determination program for executing the difference calculation step performed for each, and the data number determination step for determining the number of past data based on the difference obtained for each of the number of candidates.

According to the present invention, it is possible to determine the number of data to be referred to when compressing data.

It is a figure which shows the example of the structure of the forecasting apparatus in 1st Embodiment. It is a figure which shows the example of the forecast processing and the forecast processing in 1st Embodiment. It is a figure which shows the example of the virtual forecast processing in 1st Embodiment. It is a figure which shows the example of the frame group including the reference frame in 1st Embodiment. It is a flowchart which shows the example of the operation of the forecasting apparatus in 1st Embodiment. It is a figure which shows the example of the structure of the forecasting apparatus in 2nd Embodiment. It is a flowchart which shows the example of the operation of the forecasting apparatus in 2nd Embodiment. It is a flowchart which shows the example of the operation of the forecasting apparatus in 2nd Embodiment. It is a figure which shows the example of the structure of the forecasting apparatus in 3rd Embodiment. It is a figure which shows the example of the forecast processing based on a class in 3rd Embodiment. It is a figure which shows the example of the forecast processing based on a class in 3rd Embodiment. It is a flowchart which shows the example of the operation of the forecasting apparatus in 3rd Embodiment. It is a flowchart which shows the example of the operation of the forecasting apparatus in 3rd Embodiment.

Embodiments of the present invention will be described in detail with reference to the drawings.
(First Embodiment)
FIG. 1 is a diagram showing an example of the configuration of the forecast device 1. The forecast device 1 is an information processing device that predicts future data based on past data that constitutes time series data. The forecast device 1 notifies another device such as a data compression device of a forecast signal representing future data.

The forecast device 1 includes an input signal storage unit 10, a degree update unit 11, a prediction coefficient calculation unit 12, a virtual forecast signal generation unit 13, a virtual forecast error calculation unit 14, and a virtual forecast error minimization determination unit 15. A virtual forecast error minimum value storage unit 16, an optimum order storage unit 17, a prediction coefficient determination unit 18, and a forecast signal generation unit 19 are provided.

The forecast device 1 includes a degree update unit 11, a prediction coefficient calculation unit 12, a virtual forecast signal generation unit 13, a virtual forecast error calculation unit 14, a virtual forecast error minimization determination unit 15, and a virtual forecast error minimum value storage. The unit 16 is provided as a data number determination device 2 which is an apparatus for determining the number of data. The forecast device 1 includes at least one of an input signal storage unit 10, an optimum order storage unit 17, a prediction coefficient determination unit 18, and a forecast signal generation unit 19 as functional units of the data number determination device 2. May be good. The number of data is, for example, the number of frames of audio or image. In the following, the number of data is, for example, the number of frames of an image.

A part or all of each functional unit of the forecast device 1 is realized by, for example, a processor such as a CPU (Central Processing Unit) executing a program stored in the storage unit. The storage unit has, for example, a non-volatile recording medium (non-temporary recording medium) such as a magnetic hard disk device or a semiconductor storage device. A part or all of each functional unit may be realized by using hardware such as LSI (Large Scale Integration) or ASIC (Application Specific Integrated Circuit).

The input signal storage unit 10 stores a plurality of frames as time-series data. For example, the input signal storage unit 10 stores a plurality of frames as spatial information data. Each frame is divided into N blocks for each frame. The shape of the block is rectangular. The data number determination device 2 executes the prediction process and the forecast process for each block.

FIG. 2 is a diagram showing an example of forecast processing and forecast processing. In FIG. 2, the frame at the current time among the plurality of frames 100 is the frame 100-t. Hereinafter, the target frame for which the pixel value is predicted is referred to as a “prediction target frame”. The time interval between the frames 100 is a unit time length such as 1 minute.
[Prediction processing]
The prediction process is a process in which the data number determination device 2 generates a pixel value of the prediction target frame for each block based on the prediction target frame and the frame 100 past the prediction target frame. In FIG. 2, the data number determination device 2 generates a pixel value of the frame 100-t for each block based on the reference result in the frame of the frame 100-t and the past frame 100- (t-1) and the like. To do.

The order update unit 11 shown in FIG. 1 updates the forecast order p, which is the order used in the prediction process and the forecast process, in the order of p = 1, ..., P. P is a predetermined order. The prediction coefficient calculation unit 12 reads the frame 100- (t-1) immediately before the frame 100-t from the input signal storage unit 10. The prediction coefficient calculation unit 12 reads one or more frames 100 determined by the forecast order p updated by the order update unit 11 from the input signal storage unit 10 as reference frames.

The prediction coefficient calculation unit 12 executes the prediction process for each block of the frame 100- (t-1) immediately before the frame 100-t based on the linear model shown in the equation (1). In the following, for the sake of simplicity, the frame data is represented as a one-dimensional signal.

Here, f (x, t) represents the pixel value of the pixel at the position x in the frame 100-t. x represents a value in the range of 0 ≦ x ≦ X-1. X represents a predetermined value for the pixel value. t represents a value in the range of 0 ≦ t ≦ T-1. T represents a predetermined value for time. P in the equation (1) represents a prediction order which is a degree used in the prediction processing. The prediction order p represents the number of frames 100 referred to during the prediction processing. The predicted order is equal to the predicted order.

In the following, the symbol "^" written above the character in the mathematical formula is written immediately before the character. ^ F _p (x, t) represents a frame 100-t which is a prediction target frame. a ₀ (t), a ₁ (t), ..., _Ap (t) represent prediction coefficients, which are coefficients used in the prediction processing, respectively. In the _{following, a 0 (t), a} 1 (t), ..., a p a _(t), referred to as _a p (t). The i-th block in the frame 100 satisfies the equation (2).

The prediction coefficient determination unit 18 shown in FIG. 1 calculates a p-th order prediction coefficient that minimizes the prediction error when the frame 100-t is used as the prediction target frame. Similarly, the prediction coefficient calculation unit 12 calculates the p-th order prediction coefficient that minimizes the prediction error when the frame 100- (t-1) is set as the prediction target frame. Minimizing the prediction error means, for example, making the prediction error less than or equal to the threshold value. Prediction coefficient calculation unit 12, the case of a prediction target frame frame 100- (t-1), so as to minimize the prediction error E _p represented by the equation (3), defining a prediction coefficient.

Prediction coefficient determining unit 18, based on the order p prediction parameters that minimize the prediction error E _p in the case where the prediction target frame frame 100-t, the equation (4), (5) and (6) To calculate. Similarly, the prediction coefficient calculation unit 12, a p-th order prediction parameters that minimize the prediction error _{E p} in the case where the prediction target frame frame 100- (t-1), formula (4), (5) And formula (6).

The prediction coefficient _ap (t) obtained as the solution of the equation (4) represents a p-th order prediction coefficient that minimizes the prediction error when the frame 100-t is used as the prediction target frame. For example, when p = 1, the equation (4) is expressed as the equation (7).

When p = 1, the prediction coefficient _ap (t) is expressed as the solution of the equation (7) as in the equation (8).

Here, Δ is expressed as in the equation (9).

[Forecast processing]
Forecast processing is the pixel value of a forecasted frame, which is a future frame after the predicted frame, based on a past frame before the predicted frame, without being based on a future frame after the predicted frame. Is a process to generate for each block. In FIG. 2, the data number determination device 2 is based on the pixel values of the past frame 100 before the frame 100-t, not based on the future frame 100- (t + 1), and is based on the future frame 100- (t + 1). Pixel values are generated for each block.

The forecast signal generation unit 19 shown in FIG. 1 uses a forecast signal representing the result of the forecast processing of frame 100- (t + 1), which is a frame to be predicted in the forecast processing, as a solution of the equation (4). It is generated as shown in equation (10) based on _ap (t). Hereinafter, the forecast processing for the frame 100-t immediately before the forecast target frame is referred to as "virtual forecast processing". The virtual forecast signal generation unit 13 similarly generates a virtual forecast signal representing the result of the virtual forecast processing of the frame 100-t.

The forecast signal generation unit 19 calculates the forecast error of the frame 100− (t + 1) as in the equation (11). Similarly, the virtual forecast error calculation unit 14 calculates the virtual forecast error of the frame 100-t.

FIG. 3 is a diagram showing an example of virtual forecast processing. In FIG. 3, the data number determination device 2 generates the pixel value of the frame 100-t for each block without referring to the pixel value of the frame 100-t in the frame.

The virtual forecast signal generation unit 13 reads one or more reference frames determined by the forecast order p and the prediction coefficient calculated by the prediction coefficient calculation unit 12. The virtual forecast signal generation unit 13 executes virtual forecast processing for the frame 100-t based on the prediction coefficient and the reference frame. The virtual forecast signal generation unit 13 generates a virtual forecast signal representing the result of executing the virtual forecast process for the frame 100-t.

[Adaptation selection of forecast order]
A method for setting the forecast order p (method for determining the number of data), which is the order used for the forecast processing of the equation (10), will be described. The prediction coefficient calculation unit 12 acquires the forecast order p from the order update unit 11. The prediction coefficient calculation unit 12 calculates the p-th order prediction coefficient _ap (t-1) that minimizes the prediction error when the frame 100- (t-1) is used as the prediction target frame.

The virtual forecast signal generation unit 13 uses a forecast coefficient that minimizes the prediction error in the frame 100-t immediately before the frame to be forecast as the forecast coefficient used for the forecast processing of the frame 100- (t + 1). The prediction error in the prediction process decreases as the prediction order in the prediction process increases. However, the relationship between the forecast error in the forecast processing and the forecast order p in the forecast processing is not clear.

The virtual forecast signal generation unit 13 executes the forecast processing for the frame 100-t immediately before the forecast target frame instead of executing the forecast processing for the frame 100- (t + 1) which is the forecast target frame in the forecast processing. That is, the virtual forecast signal generation unit 13 executes the virtual forecast process for the frame 100-t. The virtual forecast signal generation unit 13 generates a virtual forecast signal representing the result of executing the virtual forecast process for the frame 100-t.

The virtual forecast error calculation unit 14 calculates a forecast error (hereinafter, referred to as “virtual forecast error”) between the frame 100-t obtained from the input signal storage unit 10 and the virtual forecast signal representing the frame 100-t. The virtual forecast error calculation unit 14 calculates the p-th order virtual forecast error when the frame 100-t is set as the forecast target frame in the virtual forecast processing based on the p-th order prediction coefficient _ap (t-1). ..

The virtual forecast error minimization determination unit 15 compares the magnitude of the calculated virtual forecast error with the provisional minimum value of the virtual forecast error. The virtual forecast error minimization determination unit 15 determines the forecast order p ^* that minimizes the virtual forecast error as shown in equation (12) by comparing the magnitude of the virtual forecast error for each forecast order p.

When the calculated virtual forecast error is smaller than the provisional minimum value of the virtual forecast error, the virtual forecast error minimization determination unit 15 sets the virtual forecast error smaller than the provisional minimum value of the virtual forecast error to the provisional minimum value of the virtual forecast error. As a value, it is recorded in the virtual forecast error minimum value storage unit 16. When the calculated virtual forecast error is smaller than the provisional minimum value of the virtual forecast error, the virtual forecast error minimization determination unit 15 records the forecast order p in the optimum order storage unit 17. In this way, the virtual forecast error minimization determination unit 15 estimates the optimum forecast order p ^* based on the virtual forecast error of the frame 100-t.

The virtual forecast error minimum value storage unit 16 stores a provisional minimum value of the virtual forecast error. The optimum order storage unit 17 stores the optimum forecast order p ^* . The forecast signal generation unit 19 executes forecast processing for frame 100− (t + 1) based on the forecast order p ^* and the formula (10) determined by the formula (12).

[Use of periodic correlation]
The virtual forecast signal generation unit 13 may refer to each frame 100 in the cycle including the current time in the virtual forecast processing. The virtual forecast signal generation unit 13 may refer to each frame 100 (non-neighborhood frame) in a cycle before the cycle including the current time in the virtual forecast processing.

The virtual forecast signal generation unit 13 refers to non-neighborhood frames in the virtual forecast processing based on the periodicity of the time series data. The forecast processing that refers to each frame 100 and the frame in the vicinity of each frame 100 from one cycle before the current time of frame 100-t to M cycle before is expressed by the equation (13).

Here, D represents a cycle such as one day, one week, one month, and one year. Virtual Warning signal generator 13, based on equation (13) refers to the _(2p m +1) frames including and neighboring the previous M cycle of the frame 100-t frame 100- (t-M × D) .. p _m satisfies the expression (14).

The virtual forecast signal generation unit 13 uses a forecast coefficient that minimizes the prediction error in the frame 100-t immediately before the frame to be forecast as the forecast coefficient used for the forecast processing of the frame 100- (t + 1). The virtual forecast error calculation unit 14 calculates the prediction error in the frame 100-t as in the equation (15).

Here, ^ f _p (x, t, _ap (t)) is expressed as in the equation (16).

[Adaptive selection of forecast order (use of periodic correlation)]
A method for setting the forecast order p (method for determining the number of data), which is the order used for the forecast processing based on the equation (13), will be described. The order update unit 11, which is a functional unit shown in FIG. 1, updates the forecast order p, which is the order used for the forecast processing, in the order of p = 1, ..., P. The prediction coefficient calculation unit 12 calculates the p-th order prediction coefficient _ap (t-1) that minimizes the prediction error when the frame 100- (t-1) is used as the prediction target frame. The virtual forecast signal generation unit 13 executes virtual forecast processing for the frame 100-t.

The virtual forecast error calculation unit 14 calculates the p-th order virtual forecast error when the frame 100-t is set as the forecast target frame based on the p-th order prediction coefficient _ap (t-1). The virtual forecast error minimization determination unit 15 determines the forecast order p ^* that minimizes the virtual forecast error as shown in equation (17) by comparing the magnitude of the virtual forecast error for each forecast order p.

The prediction coefficient determining unit 18 sets the forecast order p ^* stored in the optimum order storage unit 17 as the forecast order used in the forecast processing of the frame 100− (t + 1) which is the forecast target frame. Prediction coefficient determining unit 18, based on the reference frame determined as the frame 100-t by forecast orders ^{p *,} so as to minimize the prediction error _{E p} of represented by the equation (3), the frame 100-t Calculate the prediction coefficient of.

The forecast signal generation unit 19 generates a forecast signal of frame 100− (t + 1), which is a forecast target frame, based on one or more reference frames determined by the forecast order p ^* and a prediction coefficient. The forecast signal generation unit 19 executes forecast processing as in equation (13) based on the forecast order p ^* determined by equation (17).

FIG. 4 is a diagram showing an example of a frame group including a reference frame. The time series data includes a plurality of frames 100 having periodic time correlations with each other. The frame group 200 including the reference frame is composed of, for example, each frame 100 continuous with the frame 100- (t + 1) on the same day. Frame groups 200 on the same day may have periodic time correlations with each other.

The frame group 201 including the reference frame may be composed of, for example, each frame 100 having the same time as the time of frame 100- (t + 1) on different days. Frame groups 201 at the same time on different days may have periodic time correlations with each other. The frame group 202 including the reference frame may be composed of, for example, each frame 100 having a time different from the time of the frame 100- (t + 1) on a different day. Frame groups 202 at different times on different days may have periodic time correlations with each other.

The forecast signal generation unit 19 is based on at least one of past data in a cycle including the time of frame 100- (t + 1) as future data and past data in a cycle not including the time of frame 100- (t + 1). To generate a frame 100- (t + 1). In FIG. 4, the cycle including the time of frame 100- (t + 1) is "the day". In FIG. 4, the cycle not including the time of frame 100- (t + 1) is "the day before" or "the day before before".

Hereinafter, the block including the pixel value to be multiplied by the forecast coefficient on the right side of the equation (13) is referred to as a “forecast reference block”. Therefore, the forecast reference block is a unit block in the reference frame in the forecast processing. Hereinafter, the block including the pixel value to be multiplied by the prediction coefficient on the right side of the equation (16) is referred to as a “prediction reference block”. Therefore, the prediction reference block is a unit block in the reference frame in the prediction processing.

Virtual Warning signal generator 13, to limit the combination of forecasts reference block in the process of determining the forecast orders p ^*, may reduce the amount of calculation processing for determining a forecast orders p ^*. For example, the virtual forecast signal generation unit 13 calculates the similarity between the virtual forecast block, which is a unit block in the frame 100-t, and the forecast reference block for the frame 100-t, which is the forecast target frame in the virtual forecast processing. The virtual forecast signal generation unit 13 sorts the forecast reference blocks in descending order of similarity. That is, the virtual forecast signal generation unit 13 rearranges the forecast reference blocks in descending order of similarity. The order update unit 11 can limit the reference blocks for each forecast order p to one set by giving priority to the forecast reference blocks having a high degree of similarity and changing the forecast order p. The similarity of the predictive reference block is expressed, for example, by equation (18).

As described above, the prediction coefficient calculation unit 12 and the virtual forecast signal generation unit 13 refer to the past data when generating the prediction data based on the similarity between the unit block in the prediction target data and the unit block in the past data. Limit the number of unit blocks in. As a result, the prediction coefficient calculation unit 12 and the virtual forecast signal generation unit 13 can reduce the amount of calculation in the process of determining the forecast order p ^* .

[Exception handling]
The virtual forecast signal generation unit 13 does not execute the sort processing of the reference block when the pixel values of all the pixels in the prediction block are zero values as in the equation (19) (first exception processing).

When the pixel values of all the pixels in the prediction block are zero values, the virtual forecast signal generation unit 13 sets the selection order of the reference blocks in a predetermined order. Further, the prediction coefficient determination unit 18 determines the prediction coefficient to a predetermined coefficient (second exception handling). For example, the forecast order p of frame 100- (t + 1) may be the same as the forecast order defined for frame 100-t one unit time before the time of frame 100- (t + 1). The virtual forecast signal generation unit 13 may set the weight of each reference frame to an equal value.

When the similarity between the prediction block and the prediction reference block is less than the threshold value, the virtual forecast signal generation unit 13 sets the prediction coefficient of the prediction reference block whose similarity with the prediction block is less than the threshold value to zero (third). Exception handling). The threshold is, for example, the same value as the minimum similarity.

Next, an example of the operation of the forecast device 1 will be described.
FIG. 5 is a flowchart showing an example of the operation of the forecast device 1. The input signal storage unit 10 reads a plurality of frames 100 as time-series data (step S101). The order update unit 11 updates the forecast order p in the order of forecast order candidates p = 1, ..., P (step S102). By executing steps S102 to S109, the data number determination device 2 determines the forecast order p.

The prediction coefficient calculation unit 12 reads the frame 100- (t-1), which is the frame immediately before the frame 100-t. The prediction coefficient calculation unit 12 reads one or more frames 100 determined by the forecast order p updated by the order update unit 11 as reference frames. The prediction coefficient calculation unit 12 calculates the p-th order prediction coefficient that minimizes the prediction error when the frame 100- (t-1) is set as the prediction target frame (step S103).

The virtual forecast signal generation unit 13 reads one or more reference frames determined by the forecast order p and the prediction coefficient calculated by the prediction coefficient calculation unit 12. The virtual forecast signal generation unit 13 generates a virtual forecast signal representing the frame 100-t as a result of executing the virtual forecast process for the frame 100-t based on the prediction coefficient and the reference frame (step S104).

The virtual forecast error calculation unit 14 calculates a virtual forecast error, which is a forecast error between the frame 100-t obtained from the input signal storage unit 10 and the virtual forecast signal representing the frame 100-t (step S105). The virtual forecast error minimization determination unit 15 compares the magnitude of the calculated virtual forecast error with the provisional minimum value of the virtual forecast error (step S106). When the virtual forecast error is equal to or greater than the provisional minimum value of the virtual forecast error (step S106: NO), the virtual forecast error minimization determination unit 15 proceeds to step S109.

When the virtual forecast error is smaller than the provisional minimum value of the virtual forecast error (step S106: YES), the virtual forecast error minimization determination unit 15 sets the smaller virtual forecast error as the provisional minimum value of the virtual forecast error (step). S107). The optimum order order storage unit 17 stores the forecast order p associated with the provisional minimum value of the virtual forecast error as the optimum forecast order p (step S108).

The prediction coefficient determining unit 18 sets the forecast order p ^* stored in the optimum order storage unit 17 as the forecast order used in the forecast processing of the frame 100− (t + 1) which is the forecast target frame (step S110). The prediction coefficient determining unit 18 calculates the prediction coefficient of the frame 100-t based on the frame 100-t and the reference frame determined by the forecast order p ^* (step S111). The forecast signal generation unit 19 generates a forecast signal for frame 100− (t + 1), which is a forecast target frame, based on one or more reference frames determined by the forecast order p ^* and a forecast coefficient (step S112).

As described above, the time series data does not include future data. Time series data includes past data. The time series data further includes forecast target data in addition to past data. That is, the time series data includes the prediction target data separately from the past data. In the embodiment, the past data and the prediction target data are continuous on the time axis, and the prediction target data and the future data are continuous on the time axis, but in each case, the mutual data is not continuous on the time axis. You may. For example, the forecaster 1 has a frame of 100-t (past data) based on a past frame (past data) composed of 100- (t-2), 100- (t-3), ..., 100- (t-4). Prediction target data) may be virtually forecasted. For example, the forecast device 1 is 100- (t + 10) based on a past frame (past data) composed of 100- (t-1), 100- (t-2), ..., 100- (t-3). You may forecast the frame (future data). For example, the forecast device 1 has a frame of 100- (t + 1) based on a past frame (past data) consisting of 100- (t-10), 100- (t-20), and 100- (t-30). Future data) may be forecasted. Further, the forecast device 1 may forecast future data based on time series data composed of past data having different time intervals. The data number determination device 2 of the first embodiment includes a virtual forecast signal generation unit 13 as a prediction unit, a virtual forecast error calculation unit 14 as a difference determination unit, and a virtual forecast error minimization determination unit as a data number determination unit. It is provided with 15. The virtual forecast signal generation unit 13 as a prediction unit predicts the prediction target data included in the time series data based on a different number of past data among the plurality of past data included in the time series data. Generate data for each number. The virtual forecast error calculation unit 14 as the difference determination unit determines the difference between the prediction target data and the prediction data having a time correlation with the past data for each number. The virtual forecast error minimization determination unit 15 as the data number determination unit determines the number of past data (p ^* ) based on the difference determined for each number. As a result, the data number determination device 2 of the first embodiment can determine the number of data to be referred to when compressing the data. Further, the data number determination device 2 of the first embodiment can predict future data at a specific spatiotemporal position based on the determined number of data.

The forecast device 1 of the first embodiment further includes a forecast signal generation unit 19 as a signal generation unit. The forecast device 1 of the first embodiment may include a forecast signal generation unit 19 as a part of the data number determination device 2. Warning signal generator 19, the number ^{(p *)} on the basis of only one small number ^(p * -1) of the frame 100 as historical data and one frame 100-t than the time-series data A frame 100- (t + 1) as future data having a time correlation is generated. As a result, the data number determination device 2 of the first embodiment can determine the number of frames referred to when the frames are compressed.

The forecasting device 1 of the first embodiment predicts data that does not exist in the time series data based on a predetermined number of data having a correlation with the data existing in the time series data and data existing in the time series data. .. In this case, the data number determination device 2 of the first embodiment can appropriately determine the predetermined number of data used by the forecast device 1 for the forecast processing.

The virtual forecast signal generation unit 13 as a prediction unit generates prediction data of the prediction target data for each number of past data based on the prediction coefficient of the past data immediately before the prediction target data. The virtual forecast error minimization determination unit 15 as the data number determination unit determines the number of past data that minimizes the difference. The forecast signal generation unit 19 as a signal generation unit is based on the number of past data (p ^* -1) that is one less than the number of past data (p ^* ) that minimizes the difference and the data to be predicted. Generate future data. As a result, the data number determination device 2 of the first embodiment can efficiently determine the number of frames referred to when the frames are compressed.

Since the forecast device 1 of the first embodiment sets an appropriate forecast order p in the forecast processing based on the linear model regarding the block in the frame for the time series spatial information data, it is possible to improve the forecast accuracy. Become. The data compression device of the first embodiment can efficiently compress the spatial information data by the prediction model using the correlation in the time direction of the spatial information data which is the time series data.

(Second Embodiment)
The second embodiment is different from the first embodiment in that the optimum forecast order is determined by extending it to the bit units constituting the frame. Hereinafter, only the differences from the first embodiment will be described.
FIG. 6 is a diagram showing an example of the configuration of the forecast device 1a. The forecast device 1a includes an input signal storage unit 10, a degree update unit 11, a prediction coefficient calculation unit 12a, a virtual forecast signal generation unit 13a, a virtual forecast error calculation unit 14a, and a virtual forecast error minimization determination unit 15a. , Virtual forecast error minimum value storage unit 16, prediction coefficient determination unit 18a, forecast signal generation unit 19a, bit separation processing unit 20, virtual forecast error minimization determination unit 21, boundary bit update unit 22, and virtual. It includes a forecast error minimum value storage unit 23, an optimum boundary bit storage unit 24, an optimum order / optimum boundary bit storage unit 25, and a bit separation processing unit 26.

The forecast device 1a includes a degree update unit 11, a prediction coefficient calculation unit 12a, a virtual forecast signal generation unit 13a, a virtual forecast error calculation unit 14a, a virtual forecast error minimization determination unit 15a, and a virtual forecast error minimum value storage. Unit 16, prediction coefficient determination unit 18a, forecast signal generation unit 19a, bit separation processing unit 20, virtual forecast error minimization determination unit 21, boundary bit update unit 22, and virtual forecast error minimum value storage unit 23. And the optimum boundary bit storage unit 24 are provided as a data number determination device 2a, which is an apparatus for determining the number of data. The forecast device 1a includes at least one of an input signal storage unit 10, a prediction coefficient determination unit 18a, a forecast signal generation unit 19a, an optimum order / optimum boundary bit storage unit 25, and a bit separation processing unit 26. It may be provided as a functional unit of the data number determination device 2a.

A part or all of each functional unit of the forecast device 1a is realized by, for example, a processor such as a CPU executing a program stored in the storage unit. The storage unit has, for example, a non-volatile recording medium (non-temporary recording medium) such as a magnetic hard disk device or a semiconductor storage device. A part or all of each functional unit may be realized by using hardware such as LSI or ASIC.

The bit separation processing unit 20 reads a plurality of frames 100 stored in the input signal storage unit 10 as time-series data, and separates the read time-series data at positions determined by the boundary bits by the boundary bit update unit 22. As a result, the upper bit signal and the lower bit signal are separated. The boundary bit represents a bit position that serves as a reference for separating the high-order bit signal and the low-order bit signal. For example, the bit separation processing unit 20 separates the portion from the beginning of the frame 100 to the position defined by the boundary bit as the upper bit signal, and the portion other than the upper bit signal as the lower bit signal.

The prediction coefficient calculation unit 12a includes a high-order bit prediction coefficient calculation unit 121 and a low-order bit prediction coefficient calculation unit 122. The high-order bit prediction coefficient calculation unit 121 reads a high-order bit signal separated from one or more frames 100 determined by the forecast order p and the boundary bit as a high-order bit reference frame. The high-order bit prediction coefficient calculation unit 121 calculates the first p-th order prediction coefficient that minimizes the prediction error when predicting the immediately preceding frame 100- (t-1) using the read high-order bit reference frame. .. The low-order bit prediction coefficient calculation unit 122 reads a low-order bit signal separated from one or more frames 100 determined by the forecast order p and the boundary bit as a low-order bit reference frame. The low-order bit prediction coefficient calculation unit 122 calculates a second p-th order prediction coefficient that minimizes the prediction error when predicting the immediately preceding frame 100- (t-1) using the read low-order bit reference frame. ..

The virtual forecast signal generation unit 13a includes an upper bit virtual forecast signal generation unit 131 and a lower bit virtual forecast signal generation unit 132. The high-order bit virtual forecast signal generation unit 131 reads the high-order bit signal separated from one or more frames 100 determined by the forecast order p and the boundary bit, and the first p-th order prediction coefficient, and reads the first p-th order prediction coefficient to the frame 100-t. A first virtual forecast signal representing the result of the virtual forecast processing is generated. The lower bit virtual forecast signal generation unit 132 reads the lower bit signal separated from one or more frames 100 determined by the forecast order p and the boundary bit, and the second p-th order prediction coefficient, and reads the lower bit signal and the second p-th order prediction coefficient. A second virtual forecast signal representing the result of the virtual forecast processing is generated.

The virtual forecast error calculation unit 14a includes a high-order bit virtual forecast error calculation unit 141 and a low-order bit virtual forecast error calculation unit 142. The high-order bit virtual forecast error calculation unit 141 calculates the first virtual forecast error, which is the forecast error between the high-order bit signal of the frame 100-t and the first virtual forecast signal. The low-order bit virtual forecast error calculation unit 142 calculates the second virtual forecast error, which is the forecast error between the low-order bit signal of the frame 100-t and the second virtual forecast signal.

The virtual forecast error minimization determination unit 21 includes a virtual forecast error sum obtained based on the first virtual forecast error and the second virtual forecast error, and a virtual stored in the virtual forecast error minimum value storage unit 23. Compare the magnitude of the forecast error with the provisional minimum.
The boundary bit update unit 22 sequentially updates the boundary bits with B1 = 1, ..., B. The virtual forecast error minimum value storage unit 23 stores the provisional minimum value of the virtual forecast error. The optimum boundary bit storage unit 24 stores the optimum boundary bit. The virtual forecast error minimization determination unit 15a has a provisional minimum value stored in the virtual forecast error minimum value storage unit 23 and a provisional minimum value of the virtual forecast error stored in the virtual forecast error minimum value storage unit 16. Compare sizes.

The optimum order / optimum boundary bit storage unit 25 stores the optimum forecast order p and the optimum boundary bit.
The bit separation processing unit 26 sets the optimum forecast order p and the optimum boundary bit stored in the optimum order / optimum boundary bit storage unit 25 as the forecast order and the boundary bit of the forecast processing, and sets the frame 100-t and the frame 100-t. The reference frame determined by the forecast order is read, and the read frame is separated at the position defined by the boundary bit to separate the high-order bit signal and the low-order bit signal.

The prediction coefficient determination unit 18a includes an upper bit prediction coefficient determination unit 181 and a lower bit prediction coefficient determination unit 182. The high-order bit prediction coefficient determination unit 181 reads a high-order bit signal separated from one or more frames 100 determined by the forecast order p and the boundary bit as a high-order bit reference frame. Upper bit prediction coefficient determining unit 181 uses the upper bits reference frames read, calculates a first prediction coefficients for the frame 100-t so as to minimize the prediction error E _p. The low-order bit prediction coefficient determination unit 182 reads a low-order bit signal separated from one or more frames 100 determined by the prediction order p and the boundary bit as a low-order bit reference frame. Lower bit prediction coefficient determining unit 182, by using a lower bit reference frames read, calculates a second prediction coefficients of the frame 100-t so as to minimize the prediction error E _p.

The forecast signal generation unit 19a includes an upper bit forecast signal generation unit 191, a lower bit forecast signal generation unit 192, and a bit synthesis processing unit 193. The high-order bit forecast signal generation unit 191 is based on the high-order bit signal separated from one or more frames 100 determined by the forecast order p and the boundary bit, and the first prediction coefficient, and the frame 100-, which is a forecast target frame. Generate the first forecast signal of (t + 1). The lower bit forecast signal generation unit 192 is based on the lower bit signal separated from one or more frames 100 determined by the forecast order p and the boundary bit and the second prediction coefficient, and the frame 100-, which is a forecast target frame. Generate the second forecast signal of (t + 1). The bit synthesis processing unit 193 integrates the first forecast signal and the second forecast signal to generate a forecast signal.

Hereinafter, the processing of the forecast device 1a in the second embodiment will be described in detail.
[Adaptive forecast processing based on bit separation]
Forecast processing based on separation in the bit depth direction will be described. Let f (x, t) be a signal having a bit depth of B [bit / sample]. The bit separation processing unit 20 uses the f (x, t) signal as the upper bit signal f ^(B ₁ ^{+ 1: B)} (x, t) and the lower bit signal f ^{(1: B} ₁ ⁾ (x, t). Separate into. Significant bit signal _{^{f (B 1 +1: B)}} (x, t) _is the upper _{B-B 1} bits from _B 1 +1 bit to the most significant bit. Lower bit signal _{^{f (1: B 1) (}} x, t) is the lower _{B 1} bits from the least significant bit to _{B 1} bit.

The bit synthesis processing unit 193 generates a forecast signal representing the result of the forecast processing of the frame 100- (t + 1), which is the forecast target frame in the forecast processing, as in the equation (20).

A _p ^(B ₁ ^{+ 1: B)} (t) and a _p ^{(1: B} ₁ ⁾ (t) in Eq. 20 are coefficients obtained as solutions of Eqs. (4) to (6). That is, a _p ^(B ₁ ^{+ 1: B)} (t) and a _p ^{(1: B} ₁ ⁾ (t) are calculated through the minimization of the prediction error of the frame 100-t.
The forecast signal to f ^(B ₁ + 1, ^B) (x, t + 1, a _p ^(B ₁ + 1, ^B) (t)) for the high-order bit signal is within the range [2 ^B ₁ ^{+ 1} , 2 ^B -1]. It is a value that should exist in. In addition, "~ f" means that it is described above "~" f.

If the forecast signal for the high-order bit signal exceeds the equivalence range, the forecast signal is clipped so that it falls within the equivalence range. That ^forecast signals below ₂ ^{B 1 +1} is set to ² _B ^{1 +1,} forecast signal exceeding the _{2 B} 1 -1 is a ² _B 1 -1. Warning signal ^~f for the lower bit signal ^{_{(1: B 1) (x}} , t + 1, a p (1: B 1) (t)) is a value to be present in range _^[0,2 B 1 -1] in is there. If the forecast signal for the lower bit signal exceeds the equivalence range, the forecast signal is clipped so that it falls within the equivalence range. Forecast values less than 0 are set to ^0, forecast signal exceeding the _{2 B} 1 -1 is a ² _B 1 -1.

[Adaptive bit separation]
The method of setting the boundary bit, which is the boundary between the high-order bit signal and the low-order bit signal, will be described. The prediction coefficient _ap ^(B ₁ ^{+ 1: B)} (t) for the high-order bit signal set to minimize the prediction error in the frame 100-t and the prediction coefficient _ap ^{(1: B} ₁ ⁾ (t ⁾ for the low-order bit. ) Is used to perform forecast processing for frame 100- (t + 1). At this time, the forecast error for the frame 100− (t + 1) is expressed by the equation (21).

The bit separation processing unit 26 sets the optimum value of the boundary bit B ₁ based on the equation (22).

Prediction coefficient calculation unit 12a, the candidate value 1 of _{B 1,} ..., the prediction coefficients for B, and minimizes the prediction error of the frame ^{_{100-t a p (1:}} B 1) (t) and _a ^{p (B} ₁ ^{+ 1: B)} Calculate (t). Further, the virtual forecast error calculation unit 14a uses the prediction coefficients a _p ^{(1: B} ₁ ⁾ (t) and a _p ^(B ₁ ^{+ 1: B)} (t) to frame 100, which is a forecast target frame in the forecast processing. -Calculate the forecast error when performing forecast processing for (t + 1). Finally, selecting a candidate value of B ₁ that minimizes the forecast error. Note that B ₁ = B means that bit separation is not performed. When bit separation is not performed, the forecast signal generation unit 19a performs forecast processing based on the above [forecast processing].

When optimizing the boundary bit and the forecast order at the same time, the bit separation processing unit 26 sets the boundary bit and the forecast order based on the equation (23).

Here, B ₁ ^* and p ^* are selected from B × P candidates. P is the maximum value of the forecast order and is given from the outside. The above is the case where the bits are separated into two ranges of the upper bit signal and the lower bit signal. Let us consider a case where this is generalized and separated into the m range. The possible values of m are m = 0, ..., B. The jth boundary bit when separating into the m range is expressed as B _{m, j} . Here, j = 1, ..., M. In addition, the boundary bits for separation into the m range are collectively expressed as B _m = {B _{j, m} | j = 0, ..., M}. In addition, in order to make a unified notation, B _{m, 0} = 0 is set.

When separating into the m range, the prediction coefficient _ap ^(B _{j-1, m} ^{: B} _{j, M} ⁾ (t) for the jth range set to minimize the prediction error for the frame 100-t is used. Consider performing forecast processing for frame 100- (t + 1) by using it.
When separating into m ranges, the coefficients for each range set to minimize the prediction error for the frame 100-t are collectively expressed as equation (24).

At this time, the forecast error for the frame 100− (t + 1) is expressed by the equation (25).

When separating into the m range, the optimum value of the boundary bit B _m is set based on the equation (26).

The optimum value of the number of ranges to be separated is set based on the equation (27).

Therefore, the optimum number of ranges is m ^* , and the optimum boundary bit at that time is B ^* _m ^* .

When the boundary bit and the forecast order are optimized at the same time, the boundary bit and the forecast order are set for each range number based on the equation (28).

The optimum value of the number of ranges to be separated is set based on the equation (29).

Therefore, the optimum number of ranges is m ^* , the optimum boundary bit at that time is B ^* _m ^* , and the optimum forecast order at that time is p ^* _m ^* .

Next, an example of the operation of the forecast device 1a will be described.
FIG. 7 is a flowchart showing an example of the operation of the forecast device 1a.
The input signal storage unit 10 reads a plurality of frames 100 as time-series data (step S201). The order update unit 11 updates the forecast order p in the order of forecast order candidates p = 1, ..., P (step S202). By executing steps S203 to S209, the data number determination device 2a determines the optimum forecast order p. The boundary bit update unit 22 updates the boundary bit B1 in the order of boundary bit candidates B1 = 1, ..., B (step S203). For example, the boundary bit update unit 22 updates the boundary bits in order. The data number determination device 2a determines the optimum boundary bit by executing the optimum boundary bit update process (step S204). The specific processing of the optimum boundary bit update processing will be described later.

When the optimum boundary bit update process is executed for all the boundary bits (step S205), the virtual forecast error minimization determination unit 15a combines the virtual forecast error sum obtained based on the boundary bits with the virtual forecast error minimum value storage unit. The size is compared with the provisional minimum value stored in 16 (step S206). When the sum of virtual forecast errors is equal to or greater than the provisional minimum value (step S206: NO), the virtual forecast error minimization determination unit 15a proceeds to step S209. For example, the virtual forecast error minimization determination unit 15a instructs the order update unit 11 to update the forecast order.

When the sum of virtual forecast errors is smaller than the provisional minimum value (step S206: YES), the virtual forecast error minimization determination unit 15a sets the smaller virtual forecast error as the provisional minimum value of the virtual forecast error and stores the virtual forecast error minimum value. The value stored in unit 16 is updated (step S207). Further, the virtual forecast error minimization determination unit 15a updates the forecast order stored in the optimum order / optimum boundary bit storage unit 25 with the current forecast order p, and the optimum boundary bit for the updated forecast order p ^*. Is recorded in the optimum order / optimum boundary bit storage unit 25 (step S208).

When the processes of steps S203 to S208 are executed in all the forecast orders (step S209), the bit separation processing unit 26 has the forecast order p ^* and the boundary bits stored in the optimum order / optimum boundary bit storage unit 25. Is set to the forecast order and the boundary bit used in the forecast processing of the frame 100- (t + 1) which is the forecast target frame (step S210).

The bit separation processing unit 26 separates the reference frame determined by the set forecast order into a high-order bit signal and a low-order bit signal by the set boundary bit (step S211). The high-order bit prediction coefficient determination unit 181 calculates the first prediction coefficient of the frame 100-t based on the high-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^* (step S212). The high-order bit forecast signal generation unit 191 generates the first forecast signal of the frame 100- (t + 1), which is the forecast target frame, based on the high-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^*. (Step S213).

The low-order bit prediction coefficient determining unit 182 calculates the second prediction coefficient of the frame 100-t based on the low-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^* (step S214). The low-order bit forecast signal generation unit 192 generates a second forecast signal of frame 100- (t + 1), which is a forecast target frame, based on the low-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^*. (Step S215). After that, the bit synthesis processing unit 193 combines the first forecast signal and the second forecast signal to generate a forecast signal for frame 100− (t + 1), which is a forecast target frame (step S216).

FIG. 8 is a flowchart showing an example of the operation of the forecast device 1a. Note that FIG. 8 describes the optimum boundary bit update process performed by the forecast device 1a.
The bit separation processing unit 20 reads a plurality of frames 100 stored in the input signal storage unit 10 as time-series data, and separates the read time-series data into a high-order bit signal and a low-order bit signal by the boundary bits. (Step S2041). The high-order bit prediction coefficient calculation unit 121 minimizes the prediction error when the frame 100- (t-1) is set as the prediction target frame based on the high-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^*. The first p-th order prediction coefficient to be converted is calculated (step S2042). The high-order bit virtual forecast signal generation unit 131 reads the high-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^* and the first p-th order prediction coefficient, and performs the virtual forecast processing of the frame 100-t. A first virtual forecast signal representing the result is generated (step S2043).

The high-order bit virtual forecast error calculation unit 141 reads the frame 100-t and the generated first virtual forecast signal, and calculates the first virtual forecast error of the frame 100-t (step S2044). The low-order bit prediction coefficient calculation unit 122 minimizes the prediction error when the frame 100- (t-1) is set as the prediction target frame based on the low-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^*. The second p-th order prediction coefficient to be converted is calculated (step S2045). The low-order bit virtual forecast signal generation unit 132 reads the low-order bit signal of the reference frame determined by the frame 100-t and the forecast order p ^* and the second p-th order prediction coefficient, and performs the virtual forecast processing of the frame 100-t. A second virtual forecast signal representing the result is generated (step S2046). The low-order bit virtual forecast error calculation unit 142 reads the frame 100-t and the generated second virtual forecast signal, and calculates the second virtual forecast error of the frame 100-t (step S2047).

The virtual forecast error minimization determination unit 21 reads the first virtual forecast error and the second virtual forecast error, and calculates the virtual forecast error sum (step S2048). After that, the virtual forecast error minimization determination unit 21 compares the calculated virtual forecast error sum with the provisional minimum value of the virtual forecast error stored in the virtual forecast error minimum value storage unit 23 (step S2049). When the virtual forecast error sum is smaller than the provisional minimum value (step S2049: YES), the virtual forecast error minimization determination unit 21 is the calculated virtual forecast error sum and is stored in the virtual forecast error minimum value storage unit 23. The provisional minimum value of the virtual forecast error is updated (step S2050). Further, the virtual forecast error minimization determination unit 21 updates the optimum boundary bit stored in the optimum boundary bit storage unit 24 to the current boundary bit (step S2051).
On the other hand, when the sum of the virtual forecast errors is larger than the provisional minimum value of the virtual forecast error (step S2049: NO), the virtual forecast error minimization determination unit 21 instructs the boundary bit update unit 22 to update the boundary bits. ..

As described above, the data number determination device of the second embodiment includes the bit separation processing unit 20 as the separation processing unit, the virtual forecast signal generation unit 13a as the prediction unit, and the virtual forecast error calculation unit as the difference calculation unit. It includes a 14a and a virtual forecast error minimization determination unit 15a as a data number determination unit. The bit separation processing unit 20 separates each of the plurality of past data included in the time series data and referred to when the data is compressed into a plurality of data. The virtual forecast signal generation unit 13a generates forecast data for each combination of the number of past data used for processing and the data after separation. The virtual forecast error calculation unit 14a performs a process of calculating the difference between the forecast target data having a time correlation with the past data and the predicted data for each of the separated data for each candidate for the number of past data used for the process. To do. The virtual forecast error minimization determination unit 15a determines the number of past data based on the difference obtained for each of the number of candidates. As a result, in the forecast processing for the block-based linear model, the optimum boundary bit and forecast order can be set, and the forecast accuracy can be improved.

In addition, the breakdown of the 8 bits, which is the processing unit, is often a bit based on data that is completely unrelated to the high-order bit signal 3 bits and the low-order bit signal 5 bits, and only the processing unit is considered when making a prediction or forecast. If this is done, the predicted residual and the amount of calculation may increase. Based on the above circumstances, the data number determination device of the second embodiment bit-separates at the optimum bit boundary to generate a forecast signal. As a result, it is possible to prevent an increase in the predicted residual and the amount of calculation.

As described above, in the data number determination device of the second embodiment, the virtual forecast error minimization determination unit 15a as the data number determination unit determines the number of past data in which the sum of the differences between the separated data is minimized. decide. Thereby, the data number determination device of the second embodiment can efficiently determine the number of frames referred to when compressing the frames.

As described above, the data number determination device of the second embodiment further includes a virtual forecast error minimization determination unit 21 as a separation position determination unit that determines a position for separating past data. The virtual forecast error minimization determination unit 21 determines the separation position where the sum of the differences between the separated signals obtained when the past data is separated by the candidates for the separation position is minimized. Thereby, the data number determination device of the second embodiment can efficiently determine the optimum separation position for determining the number of frames referred to when compressing the frames.

(Third Embodiment)
The third embodiment is different from the first embodiment and the second embodiment in that the optimum forecast order is determined by extending it to a class composed of a plurality of frames. Further, the third embodiment is different from the first embodiment and the second embodiment in that the (third exception handling) in the [exception handling] is not executed. Hereinafter, only the differences from the first embodiment and the second embodiment will be described.

FIG. 9 is a diagram showing an example of the configuration of the forecast device 1b. The forecast device 1b includes an input signal storage unit 10b, a degree update unit 11, a prediction coefficient calculation unit 12b, a virtual forecast signal generation unit 13b, a virtual forecast error calculation unit 14b, and a virtual forecast error minimization determination unit 15b. , Virtual forecast error minimum value storage unit 16, prediction coefficient determination unit 18b, forecast signal generation unit 19b, class separation processing unit 28, virtual forecast error minimization determination unit 29, reference pattern update unit 30, and virtual. It includes a forecast error minimum value storage unit 31, an optimum reference pattern storage unit 32, and an optimum order / optimum reference pattern storage unit 33.

The forecast device 1b includes an input signal storage unit 10b, a degree update unit 11, a prediction coefficient calculation unit 12b, a virtual forecast signal generation unit 13b, a virtual forecast error calculation unit 14b, and a virtual forecast error minimization determination unit 15b. , Virtual forecast error minimum value storage unit 16, class separation processing unit 28, virtual forecast error minimization determination unit 29, reference pattern update unit 30, virtual forecast error minimum value storage unit 31, and optimum reference pattern storage unit. 32 is provided as a data number determination device 2b, which is an apparatus for determining the number of data. The forecast device 1b may include at least one of a prediction coefficient determination unit 18b, a forecast signal generation unit 19b, and an optimum order / optimum reference pattern storage unit 33 as functional units of the data number determination device 2b. Further, the prediction coefficient calculation unit 12b, the virtual forecast signal generation unit 13b, the virtual forecast error calculation unit 14b, the virtual forecast error minimization determination unit 29, and the reference pattern update unit 30 are for determining the number of data. It may be configured as a class determining device which is a device for determining a class.

A part or all of each functional unit of the forecast device 1b is realized by, for example, a processor such as a CPU executing a program stored in the storage unit. The storage unit has, for example, a non-volatile recording medium (non-temporary recording medium) such as a magnetic hard disk device or a semiconductor storage device. A part or all of each functional unit may be realized by using hardware such as LSI or ASIC.

The class separation processing unit 28 separates the input frame into a plurality of (for example, C) classes according to the attributes. Examples of attributes include gender, age group, etc. in the case of human flow data. In the following description, attributes are called classes. That is, the class separation processing unit 28 separates the input frame into classes according to the attributes of the input frame.

The input signal storage unit 10b stores a plurality of frames as time-series data. For example, the input signal storage unit 10b stores a plurality of frames for each class in association with the class ID. The class ID is an ID for identifying the class.
The reference pattern update unit 30 sequentially updates the reference pattern. The reference pattern represents a class ID and a reference frame. That is, the reference pattern update unit 30 sequentially updates the class and the reference frame corresponding to the class.

The prediction coefficient calculation unit 12b reads the frame 100- (t-1) immediately before the frame 100-t corresponding to the class ID updated by the reference pattern update unit 30 from the input signal storage unit 10b. The prediction coefficient calculation unit 12b reads one or more frames 100 determined by the forecast order p updated by the order update unit 11 from the input signal storage unit 10b as reference frames. The prediction coefficient calculation unit 12b calculates the p-th order prediction coefficient that minimizes the prediction error when predicting the immediately preceding frame 100- (t-1) using the reference frame.

The virtual forecast signal generation unit 13b reads one or more reference frames determined by the forecast order p and the reference pattern and the p-th order prediction coefficient calculated by the prediction coefficient calculation unit 12b, and performs virtual forecast processing of the frame 100-t. Generates a virtual forecast signal that represents the result of.
The virtual forecast error calculation unit 14b calculates a virtual forecast error, which is a forecast error between the frame 100-t obtained from the input signal storage unit 10b and the virtual forecast signal representing the frame 100-t.

The virtual forecast error minimization determination unit 29 determines the magnitude of the virtual forecast error calculated by the virtual forecast error calculation unit 14b and the provisional minimum value of the virtual forecast error stored in the virtual forecast error minimum value storage unit 31. Compare. The virtual forecast error minimization determination unit 29 determines the optimum reference pattern that minimizes the virtual forecast error by comparing the magnitude of the virtual forecast error for each reference pattern.

The virtual forecast error minimum value storage unit 31 stores a provisional minimum value of the virtual forecast error. The optimum reference pattern storage unit 32 stores the optimum reference pattern. The virtual forecast error minimization determination unit 15b has a provisional minimum value of the virtual forecast error stored in the virtual forecast error minimum value storage unit 31 and a provisional virtual forecast error stored in the virtual forecast error minimum value storage unit 16. Compare the size with the minimum value. The optimum order / optimum reference pattern storage unit 33 stores the optimum forecast order p ^* and the optimum reference pattern.

The prediction coefficient determining unit 18b sets the forecast order p ^* and the reference pattern stored in the optimum order / optimum reference pattern storage unit 33 as the forecast order and the reference pattern of the forecast processing, and sets the frame 100-t and the forecast order p. ^* and on the basis of the reference frame determined by the reference pattern, and calculates prediction coefficients of the frame 100-t so as to minimize the prediction error E _p.
The forecast signal generation unit 19b is a frame 100- (t + 1) which is a forecast target frame based on one or more reference frames determined by the forecast order p ^* and the reference pattern and the prediction coefficient calculated by the prediction coefficient determination unit 18b. ) Forecast signal is generated.

10 and 11 are diagrams showing an example of class-based forecast processing. In FIGS. 10 and 11, the frame at the current time among the plurality of frames 100 is the frame 100-t. The time interval between the frames 100 is a unit time length such as 1 minute. As shown in FIGS. 10 and 11, the forecast device 1b in the third embodiment classifies the frame 100 based on the attributes and performs forecast processing for each class.

In FIGS. 10 and 11, three classes of Class1 (for example, a woman in her thirties), Class2 (for example, a woman in her forties), and Class3 (for example, a woman in her fifties) are illustrated as Classes. In FIG. 10, the forecasting device 1b performs the forecast processing of Class2 with a frame 100-t of the same class, a frame 100- (t + 1-D), a frame 100-t of another class, and a frame 100- (t + 1-). The case where D) and is used as a reference frame is shown. FIG. 11 shows a case where the forecast device 1b uses more frames 100 as reference frames than in FIG. 10 for the forecast processing of Class2.

Hereinafter, the processing of the forecast device 1b in the third embodiment will be described in detail.
[Inter-class prediction]
The class separation processing unit 28 decomposes the spatial information data f (x, t) for each attribute. When decomposed for each attribute, f (x, t) is expressed as the sum of each attribute as shown in equation (30).

The virtual forecast signal generation unit 13b generates a forecast signal of the frame 100-t belonging to the classes to c as shown in the equation (31). In addition, "~ c" means that "~" is described above c.

Here, i-th reference block, class _IDc i (t) and the time position _l i (t), is defined. The vector c _p (t) (= c ₁ (t), ..., C _p (t)) is a class list of reference blocks. The vector l _p (t) (= l ₁ (t), ..., L _p (t)) is a frame list of reference blocks.

Scalar _c i (t), the possible range of _l i (t) _{is, c i (t) = 0} , ..., C-1, l i (t) = 0, ..., L-1, A, A ± 1, ..., A ± L, ..., MA, MA ± 1, ..., MA ± L. Here, C is the number of classes. L is an existing interval of the reference frame in the vicinity. A is a time interval corresponding to a signal cycle, M is a parameter given from the outside, and means that a frame before the first cycle to the M cycle before and a frame close thereto are designated as reference candidates. Examples of cycles include one day, one week, one month, one year, and so on. The specific method of constructing the list will be described later.

The vector a _p (t) (= a ₀ (t), a ₁ (t), ..., a _p (t)) is obtained as the solution of equation (1). Prediction coefficient calculation unit 12b derives the solution by replacing f (x, t-i) in the equation to _{g (x, t-l i} ). In other words, the prediction coefficient _ap (t) refers to p prediction reference blocks defined by a plurality of lists (vector c _p (t) and vector l _p (t)), and predicts the frame 100-t. It is set based on the norm that minimizes the error. This prediction coefficient _ap (t) is used as the prediction coefficient of the frame 100-t as shown in the equation (31).

Next, for the purpose of generalizing the forecast processing, the case of forecasting the first to τ frames (frame 100- (t + to τ)) using the coefficient _ap (τ) that minimizes the prediction error of the τ frame. think of. In addition, "~ τ" means that "~" is described above "τ". This is the case where the forecast processing is applied to the first to τ frames, while the forecast coefficient _ap (τ) is generated based on the forecast processing of the τ frame. In this case, the virtual forecast error calculation unit 14b calculates the forecast error based on the equation (32).

[Adaptation selection of forecast order]
A method for setting the forecast order p (method for determining the number of data), which is the order used for the forecast processing of the formula (32), will be described. In the third embodiment, the method of estimating the optimum value of the predicted order is composed of two steps. In the first step, the optimum reference pattern is identified for the forecast processing of order p. First, the prediction coefficient calculation unit 12b sets the p-th order prediction coefficient _ap (t-1) for each reference pattern to minimize the prediction error when the frame 100- (t-1) is the prediction target. calculate. Next, the virtual forecast error calculation unit 14 uses the prediction coefficient _ap (t-1) to obtain the p-th order forecast error when the frame 100-t is the forecast target. Further, the virtual forecast error minimization determination unit 29 compares the forecast errors of each reference pattern and identifies the reference pattern that minimizes the forecast error. That is, the process of the first step is a process of selecting a reference pattern that minimizes the virtual forecast error for the frame 100-t, and can be expressed as in the equation (33).

In the second step, the optimum order is identified using the reference pattern obtained in the first step. By using the reference pattern obtained in the first step, the virtual prediction error at the order p can be minimized. Here, p = 1, ..., P. It is assumed that P is given from the outside. Therefore, the virtual forecast error minimization determination unit 15b compares the minimum virtual forecast errors at each order, and identifies the order p that minimizes the virtual forecast error as the forecast order p ^* . That is, the process of the second step is a process of identifying the forecast order that minimizes the virtual forecast error for the frame 100-t, and can be expressed as in the equation (34).

The forecast device 1b performs forecast processing based on the formula (31) for each class using the forecast order p ^* determined by the above formula. In the above-mentioned optimization of the forecast order p ^* , the amount of calculation can be reduced by limiting the combination of reference patterns. For example, there are the following methods.
1. For p = 1, identify the class ID and reference frame in the primary virtual forecast. In this case, equation (33) becomes like equation (35).

2. for p = 2, ..., P
3. 3. p-1 Under the condition of using the p-1 class ID and the reference frame determined by the following virtual forecast, the lowest pth class ID and the reference frame are determined based on the equation (36).

4. Vector ^{_{c p * (t-1)}} , l p * (t-1) the p-1 component from a first vector ^{_{c p * (t-1)}} , inherit * vector _l ^p (t-1) , C _p (t-1), l _p (t-1), and the scalar c _p ^* (t-1) and the scalar l _p ^* (t-1) are set as the first p element.
According to the above method, the reference pattern for each forecast order can be limited to one set.

Next, an example of the operation of the forecast device 1b will be described.
FIG. 12 is a flowchart showing an example of the operation of the forecast device 1b.
The class separation processing unit 28 reads a plurality of frames 100 as time series data (step S301). The class separation processing unit 28 separates a plurality of frames 100, which are input time series data, into a plurality of classes according to attributes (step S302). The class separation processing unit 28 records the separated frame 100 in 10b in association with the class ID.

The order update unit 11 updates the forecast order p in the order of forecast order candidates p = 1, ..., P (step S303). By executing steps S304 to S310, the data number determination device 2b determines the optimum forecast order p. The reference pattern update unit 30 updates the reference patterns in order (step S304). For example, the reference pattern update unit 30 updates the class ID in order. The data number determination device 2b determines the optimum optimum pattern by executing the optimum reference pattern update process (step S305). The specific processing of the optimum reference pattern update processing will be described later.

When the optimum reference pattern update process is executed for all the reference patterns (step S306), the virtual forecast error minimization determination unit 15b stores the virtual forecast error obtained in the reference pattern and the virtual forecast error minimum value storage unit 16. The magnitude is compared with the provisional minimum value set (step S307). When the virtual forecast error is equal to or greater than the provisional minimum value (step S307: NO), the virtual forecast error minimization determination unit 15b proceeds to step S310. For example, the virtual forecast error minimization determination unit 15b instructs the order update unit 11 to update the forecast order.

On the other hand, when the virtual forecast error is smaller than the provisional minimum value of the virtual forecast error (step S307: YES), the virtual forecast error minimization determination unit 15b sets the smaller virtual forecast error as the provisional minimum value of the virtual forecast error and virtualizes it. The value stored in the forecast error minimum value storage unit 16 is updated (step S308). Further, the virtual forecast error minimization determination unit 15b updates the forecast order stored in the optimum order / optimum reference pattern storage unit 33 with the current forecast order p, and the optimum reference pattern for the updated forecast order p ^*. Is recorded in the optimum order / optimum reference pattern storage unit 33 (step S309).

When the processes of steps S304 to S309 are executed in all the forecast orders (step S310), the prediction coefficient determination unit 18b receives the forecast order p ^* and the reference pattern stored in the optimum order / optimum reference pattern storage unit 33. Is set to the forecast order and the reference pattern used in the forecast processing of the frame 100- (t + 1) which is the forecast target frame (step S311). The prediction coefficient determining unit 18b calculates the prediction coefficient of the frame 100-t based on the frame 100-t, the forecast order p ^*, and the reference frame determined by the reference pattern (step S312). The forecast signal generation unit 19b generates a forecast signal for frame 100- (t + 1), which is a forecast target frame, based on one or more reference frames determined by the forecast order p ^* and the reference pattern and the forecast coefficient (step S313). ).

FIG. 13 is a flowchart showing an example of the operation of the forecast device 1b. Note that FIG. 13 describes the optimum reference pattern update process performed by the forecast device 1b.
The prediction coefficient calculation unit 12b reads one or more frames 100 determined by the forecast order p and the reference pattern updated by the order update unit 11 as reference frames. The prediction coefficient calculation unit 12b calculates the p-th order prediction coefficient that minimizes the prediction error when the frame 100- (t-1) is set as the prediction target frame (step S3051). The virtual forecast signal generation unit 13b reads one or more reference frames determined by the forecast order p and the reference pattern and the p-th order prediction coefficient calculated by the prediction coefficient calculation unit 12b, and performs virtual forecast processing of the frame 100-t. A virtual forecast signal representing the result of is generated (step S3052). The virtual forecast error calculation unit 14b reads the frame 100-t and the generated virtual forecast signal, and calculates the virtual forecast error of the frame 100-t (step S3053).

The virtual forecast error minimization determination unit 29 determines the magnitude of the virtual forecast error calculated by the virtual forecast error calculation unit 14b and the provisional minimum value of the virtual forecast error stored in the virtual forecast error minimum value storage unit 31. Compare (step S3054). When the virtual forecast error is larger than the provisional minimum value (step S3054: NO), the virtual forecast error minimization determination unit 29 instructs the reference pattern update unit 30 to update the reference pattern.

On the other hand, when the virtual forecast error is smaller than the provisional minimum value (step S3054: YES), the virtual forecast error minimization determination unit 29 sets the smaller virtual forecast error as the provisional minimum value of the virtual forecast error and sets the virtual forecast error minimum value. The value stored in the storage unit 31 is updated (step S3055). Further, the virtual forecast error minimization determination unit 29 updates the reference pattern stored in the optimum reference pattern storage unit 32 with the current reference pattern as the optimum reference pattern (step S3056).

As described above, the class determination device of the third embodiment includes a virtual forecast signal generation unit 13b as a prediction unit, a virtual forecast error calculation unit 14b as a difference calculation unit, and a virtual forecast error minimization as a class determination unit. A determination unit 29 is provided. The virtual forecast signal generation unit 13b is a class different from the class to which the prediction target data referred to for predicting the prediction target data, which is included in the time series data classified into the class based on the attribute, belongs. Forecast data is generated for each class using the past data of at least one class classified in. The virtual forecast error calculation unit 14b performs a process of calculating the difference between the forecast target data having a time correlation with the past data and the forecast data for each class in which the forecast data is generated. The virtual forecast error minimization determination unit 29 determines the class that minimizes the difference obtained for each class as a class for determining the number of a plurality of past data referred to for predicting the prediction target data. .. This makes it possible to determine the appropriate reference class. Therefore, in the forecast processing for the block-based linear model that refers to other classes, it is possible to set an appropriate reference class, a reference frame in each reference class, and a forecast order, and it is possible to improve the prediction accuracy. ..

At least a part of the data number determination device, the forecast device, and the class determination device in the above-described embodiment may be realized by a computer. In that case, the program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by the computer system and executed. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, or a storage device such as a hard disk built in a computer system. Further, a "computer-readable recording medium" is a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, and dynamically holds the program for a short period of time. It may also include a program that holds a program for a certain period of time, such as a volatile memory inside a computer system that serves as a server or a client in that case. Further, the above program may be for realizing a part of the above-mentioned functions, and may be further realized for realizing the above-mentioned functions in combination with a program already recorded in the computer system. It may be realized by using a programmable logic device such as FPGA (Field Programmable Gate Array).

Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs and the like within a range that does not deviate from the gist of the present invention.

The present invention is applicable to a data number determining device for determining the number of data and a forecasting device for predicting future data.

1, 1a, 1b ... Forecast device, 2, 2a, 2b ... Data number determination device, 10 ... Input signal storage unit, 11 ... Order update unit, 12, 12a, 12b ... Prediction coefficient calculation unit, 121 ... Higher bit prediction coefficient Calculation unit, 122 ... Lower bit prediction coefficient calculation unit, 13, 13a, 13b ... Virtual forecast signal generation unit, 131 ... Upper bit virtual forecast signal generation unit, 132 ... Lower bit virtual forecast signal generation unit, 14, 14a, 14b ... Virtual forecast error calculation unit, 141 ... Upper bit virtual forecast error calculation unit, 142 ... Lower bit virtual forecast error calculation unit, 15, 15a, 15b ... Virtual forecast error minimization determination unit, 16 ... Virtual forecast error minimum value storage unit, 17 ... Optimal order storage unit, 18, 18a, 18b ... Prediction coefficient determination unit, 181 ... Upper bit prediction coefficient determination unit, 182 ... Lower bit prediction coefficient determination unit, 19, 19a, 19b ... Forecast signal generation unit, 191 ... Upper Bit forecast signal generation unit, 192 ... Lower bit forecast signal generation unit, 20 ... Bit separation processing unit, 21 ... Virtual forecast error minimization determination unit, 22 ... Boundary bit update unit, 23 ... Virtual forecast error minimum value storage unit, 24 ... Optimal boundary bit storage unit, 25 ... Optimal order / optimum boundary bit storage unit, 26 ... Bit separation processing unit, 28 ... Class separation processing unit, 29 ... Virtual forecast error minimization determination unit, 30 ... Reference pattern update unit, 31 ... Virtual forecast error minimum value storage unit, 32 ... Optimal reference pattern storage unit, 33 ... Optimal order / optimum reference pattern storage unit, 100 ... Frame, 200 ... Frame group, 201 ... Frame group, 202 ... Frame group

Claims (7)

A separation processing unit that separates a plurality of past data, which are included in the time series data and are referred to for predicting the prediction target data to be predicted, into upper bit and lower bit data, respectively.
And the number of historical data used for the processing, a prediction unit generating a prediction data for each combination of the respective data of the data and the lower bit of the upper bits after separation,
The past used in the process is to calculate the difference between the prediction target data having a time correlation with the past data and the prediction data for at least the separated upper bit data and the lower bit data. The difference calculation unit for each candidate for the number of data,
A data number determination unit that determines the number of past data based on the difference obtained for each of the number candidates.
A data number determination device including. The data number determination device according to claim 1, wherein the upper bit and the lower bit are separated at a position defined by a boundary bit. The data according to claim 2, wherein the boundary bit is a bit that minimizes a prediction error, which is the sum of the differences between the separated signals obtained when the past data is separated by a candidate for a position to be separated. Number determination device. The data number determination unit determines the number of past data in which the sum of the differences between the data of the upper bit and the data of the lower bit after the separation is minimized, any one of claims 1 to 3. The data number determination device described in 1. Further, a separation position determining unit for determining a position for separating the past data is provided.
Any of claims 1 to 4, wherein the separation position determining unit determines a separation position that minimizes the sum of the differences between the separated signals obtained when the past data is separated by candidates for the position to be separated. The data number determination device according to one item . It is a data number determination method executed by the data number determination device that determines the number of data.
A separation processing step that separates a plurality of past data, which are included in the time series data and are referred to for predicting the prediction target data to be predicted, into upper bit and lower bit data, respectively.
And the number of historical data used for the processing, a prediction step of generating a prediction data for each combination of the respective data of the data and the lower bit of the upper bits after separation,
The past used in the process is to calculate the difference between the prediction target data having a time correlation with the past data and the prediction data for at least the separated upper bit data and the lower bit data. The difference calculation step performed for each candidate for the number of data,
A data number determination step for determining the number of past data based on the difference obtained for each of the number candidates, and
A method for determining the number of data having. On the computer
A separation processing step that separates a plurality of past data, which are included in the time series data and are referred to for predicting the prediction target data to be predicted, into upper bit and lower bit data, respectively.
And the number of historical data used for the processing, a prediction step of generating a prediction data for each combination of the respective data of the data and the lower bit of the upper bits after separation,
The past used in the process is to calculate the difference between the prediction target data having a time correlation with the past data and the prediction data for at least the separated upper bit data and the lower bit data. The difference calculation step performed for each candidate for the number of data,
A data number determination step for determining the number of past data based on the difference obtained for each of the number candidates, and
Data number determination program to execute.

Images