JP7532087B2 - Apparatus, method, and program - Google Patents

Description

The present invention relates to data processing technology for processing image data and the like.

Tone transformation functions, which correct or transform the tone of an image, and geometric transformation functions, which correct or transform the shape of an image, can be approximated by global functions or local functions whose variables are the coordinate positions of the image, and many of these are expressed as rational polynomials such as polynomials and projective transformations. However, even when approximation is possible with local functions, it is difficult to properly define the application domain of the local function and switch the local function at the domain boundary.

Therefore, in data processing such as tone correction and shape transformation, it is common to evaluate local function values at grid points at a certain pixel interval, store the values in an LUT (lookup table), and then use the function values of the grid points read from the LUT as interpolation values or transformation values.

The data stored in the LUT does not need to be calculated in real time, so it is generated by calculation with as high precision as possible using high-order polynomials and various functions that require complex calculations. In contrast, the interpolation process of the data read from the LUT must be performed in real time, so linear or cubic interpolation is generally used, as described in Patent Document 1. Patent Document 2 also discloses a technology related to quadratic interpolation, which has intermediate performance and calculation scale between the linear and cubic interpolation described above.

However, when trying to perform interpolation using a high-order polynomial, it is necessary to perform multiple multiplications and additions, and the coefficients of the polynomial must be calculated in real time, resulting in a large-scale circuit configuration. For example, in video equipment, the pixel rate and the operating frequency of the arithmetic circuit are roughly the same, and to perform calculations for each pixel, the same number of multipliers as the number of multiplications appearing in the arithmetic formula are required. In particular, in recent years, with the emergence of high-resolution video such as 8K x 4K, in situations where the pixel rate is higher than the operating frequency of the arithmetic circuit, it is necessary to drive the arithmetic circuit in parallel, further increasing the circuit scale.

To achieve high-precision data conversion using highly nonlinear functions, in other words high-order functions, using table data and interpolation processing, it becomes necessary to narrow the grid spacing of the table data or to use high-order interpolation processing. However, the narrower the grid spacing, the larger the memory requirement, and the higher the order of the interpolation processing, the larger the calculation scale. This poses the problem that the circuit scale required to increase the interpolation precision becomes larger, resulting in higher costs.

Therefore, in Patent Document 3, the applicant disclosed a technology related to quadratic interpolation with non-uniform grid spacing, which enables data processing using high-order functions to be realized with high accuracy and a small amount of calculation, thereby enabling miniaturization of circuit scale and reduction in cost.

JP 2011-81477 A Special Publication No. 9-511079 Patent application No. 2019-026495

However, the quadratic interpolation process used in Patent Document 3 has the problem that the output interpolation curve does not pass through the set points, making calculation of the set values complicated.

In view of the above problems, one embodiment of the present invention aims to realize data conversion using higher-order functions that can easily calculate setting values with a simple configuration.

One embodiment of the present invention is an apparatus characterized by comprising: a determination means for determining an interpolation range based on input data; a first reference data generation means for generating first reference value data corresponding to a first interpolation range based on first table data in a first table holding a plurality of setting value data in which the relationship between the input data and output data is discretely expressed at a plurality of grid intervals; a second reference data generation means for generating second reference value data corresponding to a second interpolation range adjacent to the first interpolation range based on the first table data and the first reference value data; a first calculation means for calculating quadratic function coefficients for drawing a quadratic approximation curve that passes through two midpoints obtained for each of two neighboring reference value data out of three reference value data corresponding to each interpolation range and is tangent to two line segments obtained for the two neighboring reference value data; and a second calculation means for calculating output data corresponding to the input data by a quadratic interpolation calculation using the quadratic function coefficients.

According to one embodiment of the present invention, it is possible to realize data conversion using a high-order function that allows the setting value to be easily calculated with a simple configuration.

FIG. 1 is a block diagram showing a configuration of an apparatus according to a first embodiment. FIG. 1 is an explanatory diagram of a quadratic interpolation process when the grid interval DIS is constant; FIG. 1 is an explanatory diagram of a quadratic interpolation process when the grid interval DIS is not constant. FIG. 1 is a block diagram showing a detailed configuration of a coefficient calculation unit according to a first embodiment. 1 is a flowchart of a process executed by a reference data generating unit according to the first embodiment; FIG. 13 is a block diagram showing a configuration of a data conversion device according to a second embodiment. FIG. 11 is an explanatory diagram of another example of the quadratic interpolation process in the first embodiment; FIG. 1 is an explanatory diagram of a process executed by a reference data generating unit in the first embodiment; 11 is a flowchart of a process executed by a data processing device according to a second embodiment.

The present invention will now be described in detail based on preferred embodiments thereof with reference to the accompanying drawings. Note that the configurations shown in the following embodiments are merely examples, and the present invention is not limited to the configurations shown in the drawings.

[First embodiment]
<Configuration of Data Processing Device>
1(a) and 1(b) are diagrams showing an example of the configuration of an apparatus in this embodiment. Fig. 1(a) is a functional block diagram showing each functional unit corresponding to each process performed in a data conversion apparatus 100 in this embodiment. Fig. 1(b) is a block diagram showing the hardware configuration of a data processing apparatus 110 that realizes the data conversion apparatus 100 shown in Fig. 1(a).

The data processing device 110 is specifically a personal computer or the like, and as shown in FIG. 1(b), has a CPU 111, a RAM 112, a ROM 113, a communication interface (hereinafter abbreviated as I/F) 114, a storage I/F 115, and a display I/F 116. The data processing device 110 is also connected to a storage medium 120 such as a hard disk or semiconductor memory via the storage I/F 115, and to a display device 121 such as a liquid crystal display via the display I/F 116. The data processing device 110 is also connected to an imaging device 130 such as a camera and various networks (hereinafter abbreviated as NW) 131 via the communication I/F 114. The data processing device is also called an information processing device.

ROM 113 stores various programs including the data processing program according to this embodiment. The programs stored in ROM 113 are deployed in RAM 112 and executed by CPU 111. In addition to being used to deploy programs, RAM 112 is also used as a working area for CPU 111, and temporarily stores input image data, image data in the middle of processing, etc.

The communication I/F 114 communicates with, for example, the imaging device 130 or the network 131. For example, when the communication I/F 114 receives image data or the like sent from the imaging device 130 or the network 131, the image data is temporarily stored in the RAM 112 under the control of the CPU 111. The image data is then subjected to various image processing and the like as necessary by the CPU 111. Thereafter, based on the image data, an image is displayed on the screen of the display device 121 via the display I/F 116, or the image data is stored in the storage medium 120 via the storage I/F 115.

The CPU 111 performs various processes and controls in the data processing device 110, and also executes the data processing program according to this embodiment to perform various image processes including the data conversion process described below. In this embodiment, the processing of each functional unit of the data conversion device 100 shown in FIG. 1(a) is described as being realized by software processing in which the CPU 111 executes a program. However, of course, the processing of each functional unit may also be realized by a hardware configuration such as a corresponding circuit.

<Overview of secondary interpolation processing>
The functional configuration of the data conversion device 100 and the theory of secondary interpolation processing, which is the basis of the data interpolation processing performed in the data conversion device 100, will be described below with reference to Fig. 1(a) and Fig. 2 to Fig. 4. Fig. 2 is a diagram for explaining the concept of secondary interpolation processing. Fig. 3 is a diagram for explaining the concept of secondary interpolation processing with uneven grid intervals employed in this embodiment. Fig. 4 is a block diagram showing an example of the functional configuration of a coefficient calculation unit in this embodiment.

In order to execute data interpolation processing, the data conversion device 100 determines interpolation range information corresponding to the input data 10 in the range determination unit 102 and outputs it as a reference address 12. The coefficient calculation unit 103 calculates quadratic function coefficients 14 based on the LUT corresponding to the reference address 12, that is, data (referred to as reference value table data) 13 held in the reference value table stored in the reference data storage unit 101. In the coefficient calculation unit 103, first, based on the reference value table data 13, the reference data update unit 201 (see FIG. 4) generates updated reference data 21, which is new reference data. The updated reference data 21 is discrete data at lattice points at every pixel interval. The generation of reference data will be described later. Based on this updated reference data 21, the coefficient data calculation unit 204 calculates quadratic function coefficients 14.

The input data 10 is image data, data such as the brightness and coordinates of the image. The quadratic function coefficients 14 are data determined by a tone conversion function that corrects (specifically converts) the tone of the image, and a geometric conversion function that corrects (specifically deforms) the shape of the image. Based on the quadratic function coefficients 14 and the input data 10, the interpolation calculation unit 104 calculates and outputs output data 15, which is a quadratic approximation curve.

Hereinafter, the spacing between discrete grid points in the update reference data 21 will be referred to as the grid spacing DIS. Details of the grid spacing DIS will be described later, but for example, the narrower the grid spacing, the higher the accuracy of data conversion in the interpolation process, but the larger the amount of table data that must be stored. On the other hand, the wider the grid spacing, the lower the accuracy of data conversion, but the smaller the amount of table data that must be stored.

Fig. 2 is a conceptual diagram of the secondary interpolation process, showing an example in which the grid interval DIS is constant. _Y1 , _Y2 , _Y3 , and _Y4 in Fig. 2 are discrete data in the updated reference data 21, and the grid interval DIS, which is the interval between these _Y1 , _Y2 , _Y3 , and _Y4 , is a constant interval. In addition, since the interpolation range A and the interpolation range B are each a range from the midpoint between adjacent reference data to the next midpoint, the widths of the interpolation range A and the interpolation range B are the same as the grid interval DIS.

Here, for example, when the input data 10 is _X1 , the nearest update reference data 21 corresponding to this data _X1 is _Y2 , and the two neighboring (before and after) update reference data 21 with _Y2 as the center are _Y1 and _Y3 . The interpolation range is determined by the range determination unit 102, and in the example of Fig. 2, the range from the midpoint _M1 between _Y1 and _Y2 to the midpoint _M2 between _Y2 and _Y3 is calculated as the interpolation range A. The three update reference data 21 used in determining this interpolation range A are Y(-1) = _Y1 , Y(0) = _Y2 , and Y(1) = _Y3 , respectively.

In addition, the interpolation calculation unit 104 calculates a line segment, which will be described later, obtains a quadratic approximation curve based on the line segment and midpoints _M1 and _M2 , and performs an interpolation process using the quadratic approximation curve to generate interpolated output data _Y0 as output data 15.

The theoretical formula for the interpolation process executed by the interpolation calculation unit 104 is expressed by formulas (1) to (4). Note that x in formula (1) is the relative position of the input data 10 within the interpolation range normalized by the grid interval DIS, and takes a value between 0 and 1. Also, Y in formulas (1) to (4) is the output data 15.

The interpolation calculation unit 104 performs an interpolation process based on the quadratic approximation curve expressed by the formula (1). In the interpolation range A in FIG. 2, the quadratic approximation curve passes through the midpoint _M1 between _Y1 and _Y2 and the midpoint _M2 between _Y2 and _Y3 , and _is tangent to the line segment ( _Y1 - _Y2 ) connecting _Y1 and _Y2 and the line segment ( _Y2 - _Y3 ) connecting _Y2 and _Y3 . The quadratic approximation curve in the interpolation range B in FIG. 2 is obtained in the same manner as in the case of the interpolation range A described above. The quadratic approximation curve in the interpolation range B passes through the midpoint _M2 between Y2 and _Y3 and the midpoint _M3 between _Y3 and _Y4 , and is tangent to the line segment ( _Y2 - _Y3 ) connecting _Y2 and _Y3 and the line segment ( _Y3 - _Y4 ) connecting _Y3 and _Y4 .

That is, the midpoint _M2 is the same point at the boundary between the adjacent interpolation range A and the interpolation range B, and the line segment ( _Y2 - _Y3 ) is the same line segment. In other words, since the midpoint _M2 and the line segment ( _Y2 - _Y3 ) are both commonly used at the boundary between the adjacent interpolation range A and the interpolation range B, the quadratic approximation curve at the boundary between the two interpolation ranges is always smoothly connected. Note that this quadratic approximation curve is determined by a series of reference data stored as table data, without passing through a reference data generation unit, in terms of the conversion characteristics from input data to output data. Therefore, the data conversion device 100 can realize data interpolation suitable for gamma conversion or the like that requires a predetermined conversion characteristic.

<Outline of secondary interpolation process when grid interval DIS is not constant>
One of the features of this embodiment is that it takes into consideration the case where the lattice intervals in the table data are not constant, that is, the lattice intervals are uneven. Below, we will explain the case where the lattice interval to be applied changes depending on the input data.

Such uneven grid spacing is suitable for tone conversion functions with strong nonlinearity. For example, by applying a narrow grid spacing in low tone regions with strong nonlinearity and a wide grid spacing in high tone regions with strong linearity, it is possible to reduce the amount of table data that must be stored while maintaining high tone conversion accuracy throughout the entire range from low to high tones. However, in this case, the grid spacing and the interpolation spacing may differ at the boundary between adjacent interpolation ranges, so measures are required to smoothly connect the quadratic approximation curves of both interpolation ranges. In the quadratic interpolation in this embodiment when the grid spacing is uneven, it is possible to smoothly connect the quadratic approximation curves of both interpolation ranges when the grid spacing and the interpolation spacing differ at the boundary between adjacent interpolation ranges.

<Reference data generation>
Since only a portion of the necessary updated reference data 21 is stored as table data in the reference data storage unit 101, the missing reference data is generated by the reference data update unit 201. Specifically, due to the characteristics of the quadratic interpolation employed in this embodiment, in order to smoothly connect the quadratic approximation curves at the boundary between different interpolation ranges, that is, at the midpoint _M2 in Figure 3, two pieces of reference data, _Y1 ' and _Y2 ', are generated so as to satisfy the following expressions (5) to (9).

Formula (5) means that the midpoints of the input data coincide with each other, and formula (6) means that the midpoints of the output data coincide with each other. Therefore, by satisfying formulas (5) and (6), the coordinates of the midpoint _M2 , which is the boundary between the interpolation range A and the interpolation range B, coincide with each other, so that the quadratic approximation curve is continuously and smoothly connected even when the two interpolation ranges are straddled.

Equation (7) means that the slopes of the line segments connecting adjacent reference data are the same. Therefore, by satisfying equation (7), the slopes of the quadratic approximation curves at the coordinates of midpoint _M2 , which is the boundary between interpolation range A and interpolation range B, can be made to match.

Fig. 3 is a diagram for explaining the concept of processing by the reference data update unit 201 in this embodiment. Of the reference data shown in Fig. 3, data _Y1 , _Y2 , and _Y3 represented by squares and data _Y3 ' represented by triangles are stored as table data in the reference data storage unit 101. On the other hand, data _Y1 ' and _Y2 ' represented by circles in the figure are generated by the reference data update unit 201 according to equations (10) and (11), which are derived from equations (5) to (9).

In this embodiment, by satisfying formulas (10) and (11), the slopes of the quadratic approximation curves at the coordinates of the midpoint _M2 that is the boundary between the interpolation range A and the interpolation range B can be made to match.

In addition, the value of each lattice interval in the above description is preferably a power of 2 in order to reduce the computation load when selecting table data and when calculating the relative position x, as well as the relational equations of formula (10) and formula (11), but is not limited to this. Furthermore, the lattice interval DIS_B shown in FIG. 3 is twice as large as DIS_A, but the relationship between the lattice intervals is not limited to this, and may be, for example, 3 times, 4 times, 0.5 times, 0.25 times, or the like, as long as it is a positive real number multiple. In addition, an example is shown in which Y ₁ ' and Y ₂ ' are generated based on Y ₂ and Y ₃ using formula (10) and formula (11), but this is not limited to this, and Y 2 and Y ₃ may be generated based on Y ₁ ' and Y ₂ '. In this case, a calculation formula solved for Y ₂ and Y ₃ based on formulas ( ₅ ) to (9) may be used.

Furthermore, the way of holding the table data is not limited to the above. For example, FIG. 7 is a diagram for explaining another example of the concept of processing in the reference data update unit 201 according to this embodiment. Of the reference data shown in FIG. 7, data Y ₁ and Y ₂ represented by squares in the figure, and data Y ₂ ' and Y ₃ ' represented by triangles in the figure are stored in the reference data storage unit 101 as table data. On the other hand, data Y ₃ and Y ₁ ' represented by circles in the figure are generated by the reference data update unit 201 according to the calculation formulas of formulas (12) and (13) derived based on formulas (5) to (9).

The feature of this embodiment is that the reference value table data 13 required by the coefficient calculation unit 103 and stored in the reference data storage unit 101 is generated by the reference data generation unit 106 based on the setting value table data 16 stored in the setting value data storage unit 105.

The set value data storage unit 105 stores the set value provided by the higher-level control means, which is the data that is actually to be passed through in the output data 15 resulting from the quadratic interpolation calculation of the data conversion device 100 in this embodiment.

Figure 8 is a diagram for explaining the concept of the reference data generation process in this embodiment. This is an example of having four interpolation ranges for a certain grid interval, and illustrates two grid interval regions, grid interval DIS_A and grid interval DIS_B. Note that for the grid interval DIS_B, the interpolation ranges B1, B2, B3, and B4 are shown, but for the grid interval DIS_A, only the interpolation range A4 is shown. (The interpolation ranges A1, A2, and A3 are omitted.)

The set value data storage unit 105 stores the set value data M1 to M5 shown in FIG. 8 as part of the set value table data 16, and these M1 to M5 are the points through which the output of the data conversion device 100, that is, the curve of the quadratic approximation curve, actually passes. Also, Y ₄ , Y ₅ , Y ₆ , Y ₃ ', Y ₄ ', Y ₅ ', and Y ₆ ' in FIG. 8 are parts of the reference value table data 13 that must be stored in the reference data storage unit 101. The position (horizontal axis) of the set value table data 16 is placed at the midpoint of the positions (horizontal axis) of each point of the reference value table data 13. This is because the quadratic approximation curve (that is, the output data 15) output as a result of the quadratic interpolation calculation employed in this embodiment passes through the midpoints of each data stored in the above-mentioned reference value table data 13. That is, a point of the new set value table data 16 is placed at the midpoint, and reference value table data is generated so as to pass through the point, thereby realizing the quadratic approximation curve that passes through the set point, which is the object of this embodiment.

<Processing Executed by Reference Data Generator 106>
5 is a flowchart of a series of processes executed by the reference data generating unit 106. First, at the start of the process, that is, the first time reference data is generated, in step S501, reference data is generated based on one of the setting value table data 16 and a predetermined reference point. Specifically, reference value data _Y4 ', _Y5 ', and _Y6 ' are generated by the calculations of equations (14) to (16), which are extrapolation or interpolation calculations of the setting value data _M5 and the reference point P in FIG. 8. Note that this reference point is, for example, the maximum gradation point of the input data 10.

Here, DIS_B is the grid interval of the desired reference value data, and in the cases of _Y4 ', _Y5 ', and _Y6 ', it is DIS_B, but in the case of another grid interval, that value is used. As described above, _Y4 ' and _Y6 ' are extrapolated points of the set value data _M5 and the reference point P, and _Y5 ' is an interpolated point of the set value data _M5 and the reference point P, so _Y4 ' to _Y6 ' lie on a straight line. In other words, the interpolation range B4, which is the interpolation range that uses these three reference points, has output characteristics that are linear rather than a quadratic approximation curve.

Next, in step S502, it is determined whether or not it is a boundary surface of the lattice interval. If the result of this determination is true, the process proceeds to step S503, whereas if the result of this determination is false, the process proceeds to step S504. To explain in more detail, for example, in the case of Fig. 8, it is determined whether or not the data up to _Y1 ' has been referenced. In the first iteration of step S502, only the data up to _Y4 ' has been referenced, so the process proceeds to step S504.

In step S504, it is determined whether reference value data has been generated for all of the interpolation ranges. If the result of the determination in this step is true, the process proceeds to step S506, whereas if the result of the determination is false, the process proceeds to step S505. In the first iteration of step S504, reference value data has not yet been generated for all of the interpolation ranges, so the process proceeds to step S505.

In step S505, the reference value data Y ₃ ' is generated by calculating equation (17), which is an extrapolation calculation of the reference value data Y ₄ ' calculated in step S501 and the new set value data M ₄ .

As shown in FIG. 8, equation (17) for calculating Y ₃ ' is an extrapolation calculation using Y ₄ ' and M ₄ with the same lattice spacing DIS_B.

The above processing steps are repeated, and when calculations have been completed up to _Y1 ', the process proceeds to step S503 due to the conditional branch in step S502. In step S503, processing is performed on the boundary surface with different lattice spacing. Specifically, the reference value data generated in this step are _Y6 and _Y5 with a lattice spacing of DIS_A, and the reference value data _Y5 and _Y6 are generated by the calculation of equations (18) and (19), which are interpolation calculations for _Y2 ' and _Y1 ', which are used as temporary reference data to calculate these data.

As shown in Fig. 8, the lattice spacing between _Y2 ' and _Y1 ' is DIS_B, and the lattice spacing between _Y5 and _Y6 is DIS_A, and equations (18) and (19) are interpolation calculations for different lattice spacings. Note that the lattice spacing will change to DIS_A from the next time onwards, so this value is saved. Then, the process moves to step S506 again, which is the processing for the same lattice spacing.

The above process is repeated until reference value data for all interpolation ranges is generated, and when the process is completed, the conditional branch of step S504 is exited and the reference value data generated by step S506 is stored in the reference data storage unit 101.

In this manner, in this embodiment, the output data 15 can be a quadratic approximation curve passing through the setting value table data 16 set from the outside. This eliminates the need for an external setter to perform extra calculations, making it possible to reduce the effort required for setting.

In the first embodiment, the reference point, i.e., the maximum point of the input gradation, is in the interpolation range B4, but the position of the reference point may be in the interpolation range B2 or B3. In that case, processing is performed using the same procedure as when the maximum point of the input gradation is in the interpolation range B4. Furthermore, the same procedure can be used when the position of the reference point is in the interpolation range B1, which is the boundary surface of the lattice spacing, and processing of that boundary surface is performed first.

The reference point is not limited to the maximum point of the input gradation, and other points may be used. For example, the minimum point of the input gradation may be used as the reference point, in which case the procedure already described is performed starting from the low gradation side. If the reference point is any one point of the input gradation, two steps may be performed: a step of determining the low gradation side from that point, and a step of determining the high gradation side. Furthermore, the reference point may be a point outside the range of the input gradation. For example, if the reference point is a point that exceeds the maximum point of the input gradation, it may be determined starting from the high gradation side using the same procedure as described above. On the other hand, if the reference point is a point that is smaller than the minimum point of the input gradation, it may be determined starting from the low gradation side using the same procedure as described above.

In the above embodiment, the flow is exemplified in which the reference value data is stored after the generation of the reference value data for all the interpolation ranges is completed, but the timing for storing the reference value data is not limited to this, and the reference value data to be stored may be stored in the reference data storage unit 101 as appropriate. Furthermore, of the initial reference value data _Y4 ', _Y5 ', and _Y6 ' described in this embodiment, only _Y4 ' is used in the next processing step. Therefore, only _Y4 ' may be generated by the reference data generation unit 106. In that case, however, the calculation for generating _Y5 ' and _Y6 ' needs to be executed by the coefficient calculation unit 103.

Second Embodiment
A feature of this embodiment is that the reference data generating unit 106 and the reference data saving unit 101 operate only when a calculation start signal is input. In the following, differences from the previously described embodiment will be mainly described, and descriptions of contents similar to the previously described embodiment will be omitted as appropriate.

Figure 6 is a functional block diagram showing the functional units corresponding to the processes performed by the data conversion device 100 in this embodiment. Figure 9 is a flowchart of a series of processes executed by the data conversion device 100 in this embodiment.

Computation start signal 11 is added to the data conversion device 100 in this embodiment compared to the first embodiment. The reference data generation unit 106 and the reference data storage unit 101 operate when this computation start signal 11 is input. That is, in step S901, the CPU 111 determines whether or not the computation start signal 11 has been received.

When the calculation start signal 11 is received, in step S902, the reference data generating unit 106 executes processing. Note that the processing content of step S902 is the same as that of the first embodiment, specifically, the processing content of step S500 (see FIG. 5).

If the calculation start signal 11 is not received, the reference data generation unit 106 does not operate, and the reference value table data 13 stored in the reference data storage unit 101 is not updated.

Next, the process executed based on the input of input data 10 will be described. In step S903, the CPU 111 determines whether input data 10 has been input.

When input data 10 is input (YES in S903), processing is performed by the range determination unit 102 in step S904, processing is performed by the coefficient calculation unit 103 in step S905, and processing is performed by the interpolation calculation unit 104 in step S906. The respective processing contents are the same as the processing contents described in the first embodiment. In particular, the reference value table data 13 read by the coefficient calculation unit 103 is the reference value table data 13 generated in the above-mentioned step S902.

If the input data 10 is not input (NO in S903), the range determination unit 102, the coefficient calculation unit 103, and the interpolation calculation unit 104 do not operate.

In this way, in the data conversion device 100 of this embodiment, the reference data generation unit 106 and the reference data storage unit 101 mainly operate only when the calculation start signal 11 is input, so that the power consumption can be reduced.

The calculation start signal 11 may be input at any time other than the period during which the input data 10 is being input. For example, in the case of an image processing device, the calculation start signal 11 may be input at any time other than the image input period. Examples of such a period include when the data in the reference data storage unit 101 is updated, or during a non-image input period (V blanking) between frames or a non-image input period (H blanking) between horizontal scans during which the input data 10 is temporarily not being input. As a result, the reference value table data 13 is determined before the input data 10 is input, and is only read from the reference data storage unit 101 while the input data 10 is being input, thereby improving the processing speed of the data conversion device 100.

[Other embodiments]
The present invention can also be realized by a process in which a program for realizing one or more functions of each of the above-described embodiments is supplied to a system or device via a network or a storage medium, and one or more processors in a computer of the system or device read and execute the program. The present invention can also be realized by a circuit (e.g., ASIC) that realizes one or more functions.

The above-mentioned embodiments are merely illustrative examples of how the present invention may be implemented, and should not be construed as limiting the technical scope of the present invention.

In other words, the present invention can be implemented in various forms without departing from its technical concept or main features. Although each embodiment describes a case where there is only one input data, it can also be applied to a case where interpolation processing is performed on three input data such as RGB. In this case, if the RGB interpolation curve is the same in the second embodiment, it is possible to reduce the circuit size by sharing all or any of the set value data storage unit, reference data generation unit, and reference data storage unit for the three input data. Furthermore, since the shared circuit only needs to complete the calculation during the non-image input period, it is possible to further reduce the circuit size by performing calculations in multiple cycles.

In the above embodiment, an example with multiple lattice intervals, i.e., an example with unequal lattice intervals, was described, but the above embodiment can also be applied to cases with only a single lattice interval. In that case, processing at the boundary between different lattice intervals is omitted.

10 Input data 13 Reference value table data 14 Quadratic function coefficient 15 Output data 100 Data conversion device 101 Reference data storage unit 102 Range determination unit 103 Coefficient calculation unit 104 Interpolation calculation unit 105 Set value data storage unit 106 Reference data generation unit

Claims (13)

A determination means for determining an interpolation range based on input data;
a first reference data generating means for generating first reference value data corresponding to a first interpolation range based on first table data in a first table holding a plurality of set value data in which the relationship between the input data and the output data is discretely expressed at a plurality of grid intervals;
a second reference data generating means for generating second reference value data corresponding to a second interpolation range adjacent to the first interpolation range, based on the first table data and the first reference value data;
a first calculation means for calculating quadratic function coefficients for drawing a quadratic approximation curve passing through two midpoints obtained for two adjacent reference value data out of three reference value data corresponding to each interpolation range and tangent to two line segments obtained for the two adjacent reference value data;
a second calculation means for calculating output data corresponding to the input data by a quadratic interpolation calculation using the quadratic function coefficients;
An apparatus comprising: a setting value data storage means for storing the plurality of setting value data as the first table data in the first table;
a reference data storage means for storing reference value data including the first reference value data and the second reference value data as second table data in a second table different from the first table;
2. The apparatus of claim 1 further comprising: the second reference data generating means generates second reference value data in which, when two lattice intervals corresponding to two adjacent interpolation ranges are different, a midpoint among the midpoints in the two adjacent interpolation ranges that is close to a boundary between the two interpolation ranges is common to the two adjacent interpolation ranges;
3. The apparatus of claim 2 . the second reference data generating means generates second reference value data in which, when two lattice intervals corresponding to two adjacent interpolation ranges are different from each other, a slope of a line segment, among the line segments in the two adjacent interpolation ranges, that is close to a boundary between the two interpolation ranges is common to the two adjacent interpolation ranges;
4. Apparatus according to claim 2 or 3. the first reference data generating means generates the first reference value data based on a predetermined reference point and first setting value data included in the first table data;
the second reference data generating means generates the second reference value data based on second setting value data included in the first table data and the first reference value data.
5. Apparatus according to any one of claims 2 to 4. when a calculation start signal given from a higher-level control means is input, the first reference data generating means generates the first reference value data, and the first calculation means calculates the quadratic function coefficients.
6. Apparatus according to any one of claims 2 to 5, characterized in that The operation start signal is input at a timing when the input data is not being input.
7. The apparatus according to claim 6 . The calculation start signal is input at a timing when the second table data stored in the reference data storage means is updated.
7. The apparatus according to claim 6 . When there are a plurality of pieces of input data, the reference data storage means is shared among the plurality of pieces of input data.
9. Apparatus according to any one of claims 2 to 8, characterized in that 10. The apparatus according to claim 2, wherein the ratio of the values of the plurality of lattice intervals is a power of two. The reference value data stored in the reference data storage means includes reference value data with a fine grid interval and reference data with a sparse grid interval.
11. Apparatus according to any one of claims 2 to 10, characterized in that determining an interpolation range based on input data;
generating first reference value data corresponding to a first interpolation range based on first table data in a first table that holds a plurality of set value data in which the relationship between the input data and the output data is discretely expressed at a plurality of grid intervals;
generating second reference value data corresponding to a second interpolation range adjacent to the first interpolation range based on the first table data and the first reference value data;
a step of determining quadratic function coefficients for drawing a quadratic approximation curve that passes through two midpoints obtained for two adjacent reference value data out of three reference value data corresponding to each interpolation range and is tangent to two line segments obtained for the two adjacent reference value data;
obtaining output data corresponding to the input data by a quadratic interpolation calculation using the quadratic function coefficients;
The method according to claim 1, further comprising: A program for causing a computer to execute the method according to claim 12.

Images