JP2005184518A - Image processor, image processing method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain an image processor having an improved capability of rapid and consecutive photography by reducing a calculation load in an image processing section, enabling display of a photographed image using an ordinary regeneration method without any trouble. <P>SOLUTION: The image processor includes an imaging section for outputting an image signal of a partial image corresponding to a plurality of arbitrary lines or an arbitrary area in a screen; an image processing section for generating the partial image data after performing image processing on the above partial image signal; an orthogonal transformation means for performing orthogonal transformation onto the partial image data excluding a high-frequency portion which becomes short in the image reduction, based on the vertical and horizontal reduction rates of the partial image to the above screen; a resolution conversion means for modifying the image size of the partial image by adding a value smaller than a noise level to the high-frequency side of the above post-orthogonal transformation data; and a compression data generation means for generating a compression data from the above post-resolution conversion data. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus such as an electronic camera, and more particularly to an image processing apparatus capable of further speeding up image processing when obtaining image data for one screen based on partial image data obtained by thinning readout or the like. .

  Conventionally, electronic cameras that perform image processing on image signals read from an image sensor and generate captured image data of a subject image are known, and electronic cameras with a continuous shooting mode for still images are also widely used. It has become. In the continuous shooting mode of this electronic camera, the readout time of the image signal is reduced by partially reading out the image signal from any multiple lines or any area of the image sensor by thinning readout, pixel combination readout, partial readout, etc. There are many cases.

  In the case of the above thinning-out readout, etc., if the captured image data is generated as it is in the case of all pixel readout, the state in which the captured image is crushed in either the vertical direction or the horizontal direction, or the size of the captured image is reduced. It will be in a state. Therefore, when thinning readout or the like is performed with an electronic camera, it is common to perform image processing or compression coding processing after interpolating the read image signal to restore the image size of one screen.

Japanese Patent Application Laid-Open No. 2004-228561 discloses a technique for performing resolution conversion of image data by adding 0 value to the high frequency side of the orthogonal transform coefficient after orthogonal transform of the image data.
JP 7-129759 A

  Here, in order to improve the continuous shooting performance of the electronic camera, it is advantageous to perform image processing in a state where the number of pixels of the image is smaller because the amount of calculation of the image processing unit is reduced. However, when image processing is performed after interpolating the image signal as described above, the calculation amount of the image processing unit is the same as that when reading all pixels. That is, even if the readout time of the image signal is reduced by the above-described thinning readout or the like, the number of pixels is increased by the subsequent interpolation processing, and thus the continuous shooting performance may be limited by the processing speed of the image processing unit.

  Further, according to Patent Document 1, since the captured image data in the case of thinning-out readout or the like is enlarged in a state where the captured image is crushed in either the vertical direction or the horizontal direction, the above problem is still not solved. . Furthermore, Patent Document 1 discloses only resolution conversion based on discrete cosine transform, and does not disclose a specific method for resolution conversion by other orthogonal transform (for example, discrete wavelet transform).

  The present invention is to solve the above-described problems of the prior art, and its purpose is to reduce continuous load performance by reducing the calculation load of the image processing unit when an image signal is read out from an arbitrary plurality of lines on one screen. An object of the present invention is to realize an image processing apparatus that can improve and display a photographed image without any problem by a normal reproduction method.

  According to the first aspect of the present invention, an image pickup unit that outputs an image signal of a partial image corresponding to an arbitrary plurality of lines or an arbitrary area of one screen, and partial image data are generated by performing image processing on the image signal of the partial image. An image processing unit that performs orthogonal transformation on the partial image data excluding a high-frequency region that is insufficient for image reduction based on a vertical / horizontal reduction ratio of the partial image with respect to the one screen, and the orthogonal transformation Resolution conversion means for changing the image size of the partial image by adding a value smaller than the noise level to the high frequency side with respect to the later data, and compressed data generation means for generating compressed data from the data after the resolution conversion And an image processing apparatus.

According to a second aspect of the present invention, in the first aspect of the invention, the resolution converting means applies the small value to a high frequency region portion that is insufficient for image reduction based on a vertical / horizontal reduction ratio of the partial image with respect to the one screen. In addition, the image size of at least one of the horizontal direction and the vertical direction is changed.
According to a third aspect of the invention, in the first or second aspect of the invention, the imaging unit outputs an image signal of the partial image by any one of thinning readout, pixel combination readout, and partial readout. To do.

According to a fourth aspect of the present invention, in the first to third aspects of the present invention, the orthogonal transform unit performs orthogonal transform on the partial image data by discrete cosine transform or discrete wavelet transform.
According to a fifth aspect of the present invention, there is provided a step of performing image processing on an image signal of a partial image corresponding to an arbitrary plurality of lines or an arbitrary area of one screen to generate partial image data, and the vertical and horizontal directions of the partial image for the one screen A step of orthogonally transforming the partial image data excluding a high-frequency region that is insufficient for image reduction based on the reduction ratio of the image, and a value smaller than the noise level for the data after the orthogonal transformation on the high-frequency region side An image processing method comprising: a step of adding and changing an image size of the partial image; and a step of generating compressed data from the data after the resolution conversion.

  According to a sixth aspect of the present invention, a step of performing image processing on an image signal of a partial image corresponding to an arbitrary plurality of lines or an arbitrary area of one screen to generate partial image data, and the vertical and horizontal directions of the partial image with respect to the one screen. A step of orthogonally transforming the partial image data excluding a high-frequency region that is insufficient for image reduction based on the reduction ratio of the image, and a value smaller than the noise level for the data after the orthogonal transformation on the high-frequency region side A program for causing a computer to execute a step of adding and changing the image size of the partial image and a step of generating compressed data from the data after the resolution conversion.

  In the present invention, partial image data is generated from an image signal of a partial image corresponding to an arbitrary plurality of lines or an arbitrary area of one screen, and after performing orthogonal transform, the image size is enlarged by resolution conversion. That is, according to the present invention, since it is only necessary to perform image processing for the image signal of the partial image, it is possible to remarkably reduce the calculation load of the image processing unit, and without increasing the calculation speed of the image processing unit such as an electronic camera. The copying performance can be improved. Further, since the partial image data after resolution conversion is in a state almost the same as when image processing or the like is performed after interpolation, the final compressed data can be displayed without any problem by a normal reproduction method.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
(Description of the first embodiment)
The first embodiment is an example in which the image processing apparatus is an electronic camera and photographic image data is recorded in a JPEG (Joint Photographic Experts Group) format. The first embodiment corresponds to the image processing apparatus according to claims 1 to 4, the image processing method according to claim 5, and the program according to claim 6.

  The electronic camera according to the first embodiment performs orthogonal transform that orthogonally transforms partial image data for each processing unit of a size excluding a high-frequency region portion that is insufficient for image reduction based on the vertical and horizontal reduction ratios of the partial image for one screen. And resolution conversion means for changing the image size of the partial image by adding a value smaller than the noise level to the high-frequency region side with respect to the partial image data after the orthogonal transformation.

(Configuration of the first embodiment)
The configuration of the electronic camera of the first embodiment is shown in FIG. The electronic camera 100 according to the first embodiment includes an image sensor 1, an analog processing circuit 2, an A / D conversion circuit 3, a CPU 4, an ASIC 5, a buffer memory 6 that temporarily saves image data, and the image data is final. A recording medium 7 to be stored, an image sensor driving circuit 8 for generating a drive pulse of the image sensor 1, and a system bus 9.

  On the light receiving surface of the image sensor 1, light receiving elements (photodiodes) that perform photoelectric conversion are two-dimensionally arranged. In the region where the light receiving elements of the image sensor 1 are arranged, vertical readout lines are provided for each column of the light receiving elements. A horizontal readout line is provided at the output end of these vertical readout lines, and the output of the horizontal readout line is connected to the analog processing circuit 2. The area where the light receiving element of the image sensor 1 is arranged is divided into an effective pixel area for generating an image signal for image data and an optical black area for generating a black level reference signal.

  In the effective pixel area, a Bayer color filter in which color filters that transmit red (R), green (G), and blue (B) light are regularly arranged is arranged. It is converted into R, G, B image signals. In the Bayer array, R and G filters are alternately arranged in odd-numbered columns, G and B filters are alternately arranged in even-numbered columns, and the G filters have a checkered pattern as a whole.

  Further, the image pickup device 1 outputs “image quality mode” for outputting image signals of all pixels constituting one screen by sequential scanning, and image signals of partial images corresponding to any plurality of lines or any area of one screen. It is driven in either “high-speed mode” that selectively outputs. This “high-speed mode” is realized by any one of thinning readout, pixel combination readout, and partial readout of the image sensor 1.

  Here, the thinning readout is a method of extracting a partial image by thinning out an arbitrary line of the image sensor 1 and reading out an image signal. FIG. 2 is a diagram illustrating an example of the thinning readout method, and the image sensor sequentially outputs two lines of image signals for every four lines in the vertical direction. That is, the image sensor first outputs the image signal (R, G) of the first line, and then outputs the image signal (G, B) of the second line. Subsequently, the image sensor outputs the image signal of the fifth line after completing the output of the image signal of the second line, and then outputs the image signal of the sixth line. Thereafter, the same processing is repeated. In the example of FIG. 2, the number of readout lines of the image sensor 1 is half of the whole. Note that FIG. 2 is merely an example of thinning-out reading, and the number of lines to be thinned-out and the reading interval can be changed as appropriate.

  Pixel combined readout is a method of extracting partial images by adding and reading out charges from different pixels of the image sensor. FIG. 3 is a diagram illustrating an example of the pixel combination readout method. The image sensor 1 adds and outputs two lines of image signals for every four lines in the vertical direction. That is, the image sensor first adds and outputs the image signals (R, G) of the first line and the third line, and then adds the image signals (G, B) of the second line and the fourth line. And then repeat the same process. In the example of FIG. 3, the number of readout lines of the image sensor 1 is halved, and the charge amount is doubled by the charge addition at the time of transfer, so that a corresponding sensitivity increasing effect can be obtained. it can. Note that FIG. 3 is merely an example of combined pixel readout, and the number of lines to which charges are added, the readout interval, and the like can be changed as appropriate.

Furthermore, the partial reading is a method of cutting out an arbitrary area of the effective pixel area and reading out an image signal of the area as shown in FIG. 4, for example.
The analog processing circuit 2 includes a CDS circuit that reduces noise components such as an image signal by correlated double sampling, a gain circuit that amplifies the output of the image signal after noise removal, and a specific part of the waveform of the input signal. And a clamp circuit for clamping to a voltage level. The A / D conversion circuit 3 performs A / D conversion on the image signal output from the analog processing circuit 2 and outputs a digital image signal.

The CPU 4 controls the operation of each part of the electronic camera 100 via the system bus 9 and executes arithmetic processing necessary for focusing control, exposure control, etc. of the electronic camera 100. Further, the CPU 4 controls the reading mode of the image sensor 1 through the image sensor driving circuit 7 and performs various controls in the compression processing described later.
The ASIC 5 includes an image processing circuit 10 and a compression processing circuit 11.

The image processing circuit 10 performs various image processing such as contour compensation, gamma correction, and white balance adjustment on the digital image signal of the partial image output from the A / D conversion circuit 3 to generate partial image data. Note that the partial image data at this stage is in a state in which at least one of vertical and horizontal is reduced with respect to the image size of one screen.
The compression processing circuit 11 executes the above-described partial image data compression processing. As shown in FIG. 5, the compression processing circuit 11 includes a blocking unit 20, a discrete cosine transform unit (DCT: Discrete Cosine Transform) 21, a quantization unit 22, a resolution conversion unit 23, and a variable length coding unit. 24.

(Operation of the first embodiment)
Hereinafter, a series of operations of the electronic camera of the first embodiment will be described with reference to the flowchart of FIG.
Step S101: With the shutter release, the CPU 4 drives the image sensor 1 via the image sensor drive circuit 8. The image sensor 1 photoelectrically converts the subject image to generate an image signal, and outputs the image signal of the partial image by, for example, thinning-out reading in FIG.

Step S102: The image signal of the partial image is subjected to predetermined analog processing by the analog processing circuit 2 and then digitized by the A / D conversion circuit 3.
Step S103: The image signal of the partial image output from the A / D conversion circuit 3 is input to the ASIC 5. Then, the image processing circuit 10 in the ASIC 5 performs image processing on the image signal of the partial image to generate partial image data, and outputs the partial image data to the compression processing circuit 11.

  Note that the partial image data at this stage is in a state in which at least one of the vertical and horizontal sizes is reduced with respect to the image size of one screen, and if the normal JPEG compression encoding is performed, the reproduced image is The image size of the partial image remains unchanged. For example, in the case of thinning readout in FIG. 2, the partial image data remains in a state where the image size is halved only in the vertical direction.

Step S104: The blocking unit 20 of the compression processing circuit 11 divides the entire partial image data into processing units called blocks. Each block is independently subjected to a series of processes described later.
Here, the block forming unit 20 divides the partial image data into blocks having a size excluding a high-frequency region portion that is insufficient for image reduction, corresponding to the vertical and horizontal reduction ratios of the partial image for one screen. That is, when processing the partial image data, the blocking unit 20 does not divide the partial image data in a block (hereinafter referred to as “basic block”) when dividing the image data of one screen.

For example, in the case of thinning-out reading in FIG. 2, the partial image data is divided as follows. If the basic block is 8 × 8 pixels, the partial image in FIG. 2 has the vertical image size halved, so the blocking unit 20 converts the basic block (8 × 8 pixel) block vertically. The direction is halved and blocks are formed by 4 × 8 pixel blocks (see FIG. 7).
Step S105: The DCT unit 21 performs discrete cosine transform on the partial image data in units of blocks to convert it into DCT coefficients. Data output from the DCT unit 21 is a DCT coefficient having a size corresponding to the block.

Step S106: The quantization unit 22 quantizes the DCT coefficient using the quantization matrix corresponding to the above block.
Step S107: The resolution conversion unit 23 performs resolution conversion of the quantized DCT coefficient, and expands the data so as to correspond to the basic block. Specifically, the resolution conversion unit 23 adds 0 value to the high frequency side of the DCT coefficient by an amount that is insufficient due to image reduction, and outputs a DCT coefficient corresponding to the size of the basic block.

For example, in the case of thinning readout in FIG. 2, since the vertical image size of the partial image is halved, the resolution conversion unit 23 adds 0 values for four columns in the vertical direction to perform DCT. The coefficient matrix is changed to 8 × 8 (see FIG. 7). As a result, the size of the entire image output from the resolution conversion unit 23 is restored to the image size for one screen.
Step S108: The variable length coding unit 24 performs variable length coding on the DCT coefficients output from the resolution conversion unit 23 so as to shorten the average code length, and outputs compressed data. This compressed data is finally stored in the recording medium 7. Thus, a series of operations of the electronic camera according to the first embodiment is completed.

(Effect of 1st Embodiment)
In the first embodiment, the image processing unit generates partial image data, performs DCT without interpolating the partial image data, and then adds a 0 value to the high frequency side of the DCT coefficient to set the image size to 1. Restore to correspond to the screen. Therefore, since the image processing unit only needs to perform image processing for the image signal of the partial image, the calculation load can be significantly reduced, and continuous shooting performance can be improved without increasing the calculation speed of the image processing unit. it can. In the case of partial reading, a digital zoom image obtained by cutting out a part of the screen can be obtained at high speed.

Further, the finally output compressed data is enlarged corresponding to the vertical and horizontal reduction ratios of the partial image for one screen, and is in a state almost the same as when image processing or the like is performed after interpolation. The compressed data can be displayed without any problem by normal JPEG decompression processing.
(Description of Second Embodiment)
The second embodiment is an example in which the image processing apparatus is an electronic camera and the captured image data is recorded in the JPEG2000 format. The second embodiment can be applied when the aspect ratio and the aspect ratio of the image size of the partial image for one screen are 1: 2 ⁿ (where n ≦ 0). The second embodiment corresponds to the image processing apparatus according to claims 1 to 4, the image processing method according to claim 5, and the program according to claim 6.

  The electronic camera according to the second embodiment adjusts the number of horizontal and vertical subband divisions based on the vertical and horizontal reduction ratios of the partial image for one screen, and removes the high-frequency region portion that is insufficient for image reduction. Orthogonal transform means for wavelet transforming the partial image data, and resolution transform means for changing the image size of the partial image by adding a value smaller than the noise level to the high frequency region side for the partial image data after the orthogonal transform, have.

(Configuration of Second Embodiment)
FIG. 8 is a diagram illustrating a configuration of the compression processing circuit 11 according to the second embodiment. Since the configuration other than the compression processing circuit 11 is the same as that of the first embodiment, description of other configurations is omitted.
The compression processing circuit 11 of the second embodiment includes a DC level shift unit 30, a color conversion unit 31, a discrete wavelet transform unit (DWT) 32, a quantization unit 33, a resolution conversion unit 34, A bit modeling unit 35, an arithmetic coding unit 36, and a bit stream generation unit 37 are included.

(Operation of Second Embodiment)
Hereinafter, a series of operations of the electronic camera of the second embodiment will be described with reference to the flowchart of FIG. Note that steps S201 to S203 correspond to steps S101 to S103 of the first embodiment, and a description thereof will be omitted.
Step S204: When the signal value is a positive value, the DC level shift unit 30 performs level shift by subtracting the dynamic range by half.

Step S205: The color conversion unit 31 performs color coordinate conversion on the input RGB partial image data to convert it into a color component such as YCbCr.
Step S206: The DWT unit 32 performs discrete wavelet transform for each color component, and frequency-decomposes each component into subband components. At this time, the DWT unit 32 adjusts the number of subband divisions in the horizontal direction and the vertical direction in accordance with the vertical and horizontal reduction ratios of the partial image for one screen. As a result, the partial image data is subjected to discrete wavelet transform, excluding the high-frequency region that is insufficient for image reduction.

  For example, in the case of thinning readout shown in FIG. 2, the image size of the partial image is halved only in the vertical direction, so the DWT unit 32 does not perform vertical subband division at the decomposition level 1. The sub-band is divided only in the horizontal direction. At the decomposition level 2, the DWT unit 32 performs subband division in the vertical direction and the horizontal direction with respect to the subband 1LL (see FIG. 10).

Step S207: The quantization unit 33 quantizes the wavelet transform coefficient as necessary.
Step S208: The resolution conversion unit 34 adds a zero value to the high frequency region side of the wavelet transform coefficient by the amount that is insufficient due to image reduction, thereby generating a missing high frequency region subband.

  For example, in the case of the above example of step S206, since the partial image data halved in the vertical direction is divided into subbands only in the horizontal direction at the decomposition level 1, the subbands 1LH and 1HH in the vertical direction are missing. Therefore, the resolution conversion unit 34 adds 0 values to the high frequency region side to generate subbands 1LH and 1HH (see FIG. 10). As a result, the size of the entire image output from the resolution conversion unit 34 is restored to the image size for one screen.

  Step S209: The bit modeling unit 35 divides the wavelet transform coefficient into code blocks (for example, 64 × 64) within each subband. The transform coefficient in each code block is divided into a sign bit and an absolute value, and then the absolute value is distributed to a natural binary bit plane, and three encoding passes are sequentially performed from the most significant bit plane. Executed. At this time, the bit modeling unit 35 controls the compression rate by terminating the encoding in an arbitrary encoding pass of an arbitrary bit plane.

Step S210: The arithmetic encoding unit 36 acquires encoded data for each code block from the bit modeling unit 35, and performs arithmetic encoding on the encoded data using an MQ coder that is a binary arithmetic code. Apply.
Step S211: The bit stream generation unit 37 generates a bit stream by rearranging the encoded data in order of priority, and adds a header or the like to generate final compressed data. This compressed data is finally stored in the recording medium 7. Thus, a series of operations of the electronic camera of the second embodiment is completed.

(Effect of 2nd Embodiment)
In the second embodiment, the image processing unit generates partial image data, performs DWT without interpolating the partial image data, and then adds a 0 value to the high-frequency region side of the wavelet transform coefficient to set the image size. Restore to correspond to one screen. In the second embodiment, in addition to the same effects as those of the first embodiment, it is possible to obtain an excellent reproduced image with less block noise and mosquito noise than the first embodiment.

(Correspondence between Claims and Embodiments)
Here, the correspondence between the claims and the embodiment is shown. Note that the correspondence shown below is an interpretation given for reference only, and does not limit the technical scope of the present invention.
The imaging device 1 corresponds to the “imaging unit”. The “image processing unit” corresponds to the image processing circuit 10 of the ASIC 5. The functions of the image processing circuit 11 of the ASIC 5 correspond to “orthogonal transformation means”, “resolution transformation means”, and “compressed data generation means”, respectively. Further, the first embodiment is a case of orthogonal transformation by discrete cosine transformation, and the second embodiment is a case of orthogonal transformation by discrete wavelet transformation. The program according to claim 6 is executed by the CPU 4.

(Supplementary items of the embodiment)
As mentioned above, although this invention has been demonstrated by said embodiment, the technical scope of this invention is not limited to the said embodiment. For example, in the above-described embodiment, the example in which the orthogonal transformation is performed on the partial image data by the discrete cosine transformation or the discrete wavelet transformation has been described. However, the present invention also applies when the partial image data is orthogonally transformed by another known orthogonal transformation method. Can be applied.

  Further, the image processing apparatus is not limited to the electronic camera as in the above embodiment, and the image processing apparatus may be a scanner or the like. For example, in the case of a film scanner, the imaging unit includes a line sensor such as a CCD extending in the main scanning direction and a document reading mechanism. The line sensor reads the transmitted light of the document (film) at predetermined intervals in the sub-scanning direction, and outputs an image signal of the partial image thinned out in the sub-scanning direction.

Furthermore, in the resolution conversion of the above-described embodiment, an example in which 0 value is added to the high frequency side has been shown. However, if the value is smaller than the noise level, the resolution conversion unit adds a value other than 0 value to convert the resolution. May be performed. In the above embodiment, the case where the image is reduced mainly in the vertical direction has been described. However, the present invention can be similarly applied to the case where the image is reduced in the horizontal direction.
In the electronic camera of the above embodiment, an example in which the imaging unit is configured by an imaging element of an XY address method (CMOS, LBCAST, etc.) is shown, but the imaging unit is configured by an imaging device of a charge sequential transfer system (CCD, etc.). May be.

  Furthermore, in the second embodiment, for the sake of convenience, an example in which discrete wavelet transform is performed on the entire partial image has been described. However, the partial image data is divided into tiles of an arbitrary size, and the above processing is performed in units of tile components. Also good.

  When obtaining image data for one screen based on partial image data obtained by thinning readout or the like, the image processing can be further speeded up without increasing the calculation speed of the image processing unit.

It is a block diagram of the electronic camera of 1st Embodiment. It is a figure which shows an example of the thinning-out reading system. It is a figure which shows one example of a pixel combination read-out system. It is a figure which shows one example of a partial reading system. It is a figure which shows the structure of the compression processing circuit of 1st Embodiment. It is a flowchart of a series of operation | movement of the electronic camera of 1st Embodiment. It is a schematic diagram of resolution conversion of a 1st embodiment. It is a figure which shows the structure of the compression processing circuit of 2nd Embodiment. It is a flowchart of a series of operation | movement of the electronic camera of 2nd Embodiment. It is an outline figure of resolution conversion of a 2nd embodiment.

Explanation of symbols

1 Image sensor 2 Analog processing circuit 3 A / D conversion circuit 4 CPU
5 ASIC
6 Buffer memory 7 Recording medium 8 Image sensor drive circuit 9 System bus 10 Image processing circuit 11 Compression processing circuit 20 Blocking unit 21 Discrete cosine transform unit 22 Quantization unit 23 Resolution conversion unit 24 Variable length coding unit 30 DC level shift unit 31 Color conversion unit 32 Discrete wavelet conversion unit 33 Quantization unit 34 Resolution conversion unit 35 Bit modeling unit 36 Arithmetic coding unit 37 Bit stream generation unit 100 Electronic camera

Claims (6)

An imaging unit that outputs an image signal of a partial image corresponding to an arbitrary plurality of lines or an arbitrary area of one screen;
An image processing unit that performs image processing on an image signal of the partial image to generate partial image data;
An orthogonal transform unit that orthogonally transforms the partial image data based on the vertical and horizontal reduction ratios of the partial image with respect to the one screen, excluding a high-frequency region portion that is insufficient for image reduction;
Resolution conversion means for changing the image size of the partial image by adding a value smaller than the noise level to the high frequency region side for the data after the orthogonal transformation;
An image processing apparatus comprising: compressed data generation means for generating compressed data from the data after the resolution conversion.   The resolution conversion unit adds the small value to a high frequency region portion that is insufficient for image reduction based on a vertical / horizontal reduction ratio of the partial image with respect to the one screen, and at least one image in a horizontal direction or a vertical direction. The image processing apparatus according to claim 1, wherein the size is changed.   The image processing apparatus according to claim 1, wherein the imaging unit outputs an image signal of the partial image by one of thinning readout, pixel combination readout, and partial readout.   The image processing apparatus according to any one of claims 1 to 3, wherein the orthogonal transform unit performs orthogonal transform on the partial image data by discrete cosine transform or discrete wavelet transform. A step of performing image processing on an image signal of a partial image corresponding to an arbitrary plurality of lines or an arbitrary area of one screen to generate partial image data;
Based on the vertical and horizontal reduction ratios of the partial image with respect to the one screen, orthogonally transforming the partial image data excluding a high-frequency region portion that is insufficient for image reduction;
A step of changing the image size of the partial image by adding a value smaller than the noise level to the high frequency region side for the data after the orthogonal transformation
And a step of generating compressed data from the resolution-converted data. A step of performing image processing on an image signal of a partial image corresponding to an arbitrary plurality of lines or an arbitrary area of one screen to generate partial image data;
Based on the vertical and horizontal reduction ratios of the partial image with respect to the one screen, orthogonally transforming the partial image data excluding a high-frequency region portion that is insufficient for image reduction;
A step of changing the image size of the partial image by adding a value smaller than the noise level to the high frequency region side for the data after the orthogonal transformation
A program for causing a computer to execute the step of generating compressed data from the resolution-converted data.

Images