JP2001352543A - Image coding system and image coder and method and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image coder that applies coding to image data without being affected to the utmost by restriction conditions such as a processing time of the coder, a memory and an arithmetic cost under a circumstance where the image data are broken down to frequency components, each frequency component is quantized, and a prescribed entropy coding is applied to quantized data to generate coded data. SOLUTION: The system of this invention is provided with an image input device and an image processing unit, the image input device is provided with a means that inputs image data, a means that converts the image data into frequency components to obtain a conversion coefficient, and a 1st entropy coding means that applies entropy coding to a quantized value of the conversion coefficient, and the image processing unit is provided with a means that decodes the coded data obtained by the 1st entropy coding means and a 2nd entropy coding means that applies again an entropy coding to the conversion coefficient or the quantized value obtained by the decoding by a method different from that of the 1st entropy coding means.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding system for encoding image data, an image encoding apparatus and method, and a storage medium.

[0002]

2. Description of the Related Art In recent years, as the technology of image input devices such as digital cameras and scanners has been improved, the resolution of image data captured by these input devices has been steadily increasing. For low-resolution images, the amount of image data was small and did not hinder processing such as encoding, transmission, and storage.However, as the resolution became higher, the amount of image data became enormous, and data was transmitted. However, there is a problem that it takes a lot of time in the process, and a large storage capacity is required for the encoding process and the accumulation.

As a technique for efficiently transmitting and displaying such a large amount of image data, a stepwise transmission method of the image data has attracted attention. This is because the image quality is restored at the data receiving side by transmitting low-quality images so that the outline of the image can be grasped at the initial stage of image data transmission, and then transmitting data necessary for high-quality images in order. Is to be done.

[0004] As an encoding method suitable for such stepwise transmission, spatial stepwiseness is realized by using a wavelet transform for a sequence conversion, and SNR is realized by using a bit plane encoding for entropy encoding. An encoding method that realizes the stepwise nature of the above has been studied.

FIG. 2 shows an example of an image coding system using the above-described coding method. In the figure, 201
Denotes an image input unit, 202 denotes a discrete wavelet transform unit, 203 denotes a coefficient quantization unit, 204 denotes a bit plane coding unit, 205 denotes a code string forming unit, and 206 denotes a code output unit.

Hereinafter, the operation of the image coding system shown in FIG. 2 will be described.

First, pixel data P (x, y) indicating an image to be encoded is input from the image input unit 201 in raster scan order. x and y represent the horizontal and vertical positions of the pixel. The image input unit 201 is, for example, a storage device such as a hard disk, a magneto-optical disk, or a memory that stores image data, an imaging device such as a scanner, or a network line interface.

The discrete wavelet transform unit 202 performs a two-dimensional discrete wavelet transform while appropriately storing the pixel data P (x, y) input from the image input unit 201 in an internal buffer (not shown), and performs LL, LH1, Decompose into seven sub-bands HL1, HH1, LH2, HL2, HH2 and output the coefficients of each sub-band.
Hereinafter, the coefficient of each subband is represented as C (S, x, y). Where S
Represents a subband, and is one of LL, LH1, HL1, HH1, LH2, HL2, and HH2. Also, x and y represent the horizontal and vertical coefficient positions when the coefficient position at the upper left corner in each subband is set to 0,0.

The two-dimensional discrete wavelet transform is realized by applying a one-dimensional transform to each of the horizontal and vertical directions. Fig. 4 shows that a one-dimensional discrete wavelet transform is applied to the image to be encoded in the vertical direction first, and is decomposed into a low-frequency subband L and a high-frequency subband H. FIG. 11 shows a state in which a transform is applied to decompose into four subbands of LL, HL, LH, and HH. In this image encoding apparatus, N
The one-dimensional discrete wavelet transform on the one-dimensional signals x (n) (n is from 0 to N−1) is performed by the following equation.

H (n) = x (2n + 1)-floor {(x (2n) + x (2n + 2)) / 2} l (n) = x (2n) + floor {(h (n- 1) + h (n) + 2) / 4}

Where h (n) is the coefficient of the high frequency subband,
l (n) represents the coefficients of the low frequency subband, and floor {R} is a real number
Represents the largest integer value not exceeding R. Although not described here, both ends x (n) (n <0 and n ≧ N) of the one-dimensional signal x (n) required for the calculation of the above equation are calculated by a known method. n) (0 ≦ n <N).

By applying the two-dimensional discrete wavelet transform to the sub-band LL obtained by the above two-dimensional discrete wavelet transform two more times, as shown in FIG. 5, LL, LH1, HL1, HH1, Decomposes into 7 subbands of LH2, HL2, HH2. The LL in FIG. 5 is the LL in FIG.
Are not identical because they are re-decomposed.

The coefficient quantizing section 203 has a coefficient C (S, x, y) of each subband generated by the discrete wavelet transform section 202.
Is quantized using a quantization step delta (S) defined for each subband. Assuming that the quantized coefficient value is represented by Q (S, x, y), the quantization process performed by the coefficient quantization unit 203 is represented by the following equation.

Q (S, x, y) = sign {C (S, x, y)} × floor {|
C (S, x, y) | / delta (S)}

Here, sign {I} is a function representing the sign of the integer I, and returns 1 if I is positive and -1 if I is negative. Also,
floor {R} represents the largest integer value not exceeding the real number R.

The bit plane encoding unit 204 encodes the coefficient value Q (S, x, y) quantized by the coefficient quantization unit 203 to generate a code sequence. A method of dividing the coefficients of each subband into blocks and facilitating random access by separately coding is known, but here, coding is performed in units of subbands to simplify the description. I do. Coding of the quantized coefficient Q (S, x, y) (hereinafter simply referred to as coefficient value) of each subband is performed by naturally calculating the absolute value of the coefficient value Q (S, x, y) in the subband. It is represented by a binary number, and is performed by performing binary arithmetic coding with priority on the bit plane direction from the upper digit to the lower digit. Coefficient for each subband
A description will be given by describing the bit of the nth digit from the bottom when Q (S, x, y) is expressed in natural binary notation as Qn (x, y). Note that a variable n representing a digit of a binary number is referred to as a bit plane number, and the LSB of the bit plane number n is the 0th digit.

FIG. 6 shows a flow of a process of encoding the subband S by the bit plane encoding unit 204.

In FIG. 6, step S601 is a sub-band S
Obtaining the maximum value Mabs (S) of the absolute values of the coefficients in S, S
602 is a step for obtaining the significant digit NBP (S) required to represent the absolute value of the coefficient in the subband, S603 is a step for substituting the significant digit for a variable n, and S604 is for n-1 to obtain n. The substituting step, S605, is a step of encoding the n-th bit plane, and S606 is a step of determining whether or not n is 0.

First, in step S601, the absolute value of the coefficient in the subband S to be encoded is checked, and its maximum value Mabs (S)
Ask for.

Next, in step S602, the number of digits NBP (S) required to represent Mabs (S) in a binary number is determined by the following equation.

NBP (S) = ceil {log2 (Mabs (S))}

Here, ceil {R} represents a minimum integer value equal to or greater than the real number R.

In step S603, the number of significant digits NBP (S) is substituted for the bit plane number n.

In step S604, 1 is subtracted from the bit plane number n.

In step S605, the bit plane n is encoded using a binary arithmetic code. In the present embodiment, QM-Coder is used as an arithmetic code. This QM-Cod
For the procedure for encoding a binary symbol generated in a certain state (context) S using er, or for the initialization procedure and the termination procedure for arithmetic coding processing, refer to the ITU-T Recommendation for still images. T.81 | ISO / IEC10918-1
Since the details are described in the recommendations and the like, the description is omitted here. It is assumed that the arithmetic encoder is initialized at the start of encoding of each bit plane, and termination processing of the arithmetic encoder is performed at the end. Immediately after "1" which is coded at the beginning of each coefficient, the sign of the coefficient is represented by 0 or 1, and arithmetically coded. Here, it is 0 if positive, and 1 if negative.
For example, if the coefficient is -5 and the number of significant digits NBP (S) of the subband S to which the coefficient belongs is 6, the absolute value of the coefficient is a binary number.
000101, and is encoded from the upper digit to the lower digit by encoding each bit plane. At the time of encoding the second bit plane (in this case, the fourth digit from the top), the first “1” is encoded, and immediately after that, the sign “1” is arithmetically encoded.

In step S606, the bit plane number n
Is compared with 0, and if n = 0, that is, if the encoding of the LSB plane is performed in step S605, the subband encoding process is terminated; otherwise, the process proceeds to step S604.

By the above-described processing, all the coefficients of the sub-band S are encoded, and the code string CS (S,
Generate n). The generated code sequence is sent to the code sequence forming unit 205, and is temporarily stored in a buffer (not shown) in the code sequence forming unit 205.

When the encoding of coefficients of all subbands is completed by the bit plane encoding unit 204 and the entire code sequence is stored in the internal buffer, the code sequence forming unit 205 stores the code sequence in the internal buffer in a predetermined order. The code sequence is read out, necessary additional information is inserted, a final code sequence to be output from the present encoding device is formed, and output to the code output unit 206.

The final code string generated by the code string forming unit 205 is composed of a header, level 0, level 1 and level 2
It is composed of encoded data hierarchized. Level 0 encoded data is composed of a code string of CS (LL, NBP (LL) -1) to CS (LL, 0) obtained by encoding the coefficients of the LL subband. Level 1 is a code string CS (LH1, NBP (LH1) -1) to CS obtained by coding the coefficients of each subband of LH1, HL1, HH1.
(LH1,0), CS (HL1, NBP (HL1) -1) to CS (HL1,0), and C
S (HH1, NBP (HH1) -1) to CS (HH1,0). Level 2 is a code string CS (LH2, NBP (LH2) -1) to CS (LH2, LH2, HL2, HH2) obtained by encoding the coefficients of each subband.
0), CS (HL2, NBP (HL2) -1) to CS (HL2,0), and CS (HH
2, NBP (HH2) -1) to CS (HH2,0).

FIG. 3 shows the structure of a code string generated by the code string forming unit 205.

The code output section 206 outputs the code string generated by the code string forming section 205 to the outside of the apparatus. This code output unit 20
Reference numeral 6 denotes a storage device such as a hard disk or a memory, an interface of a network line, and the like.

[0032]

However, in the conventional stepwise coding method as described above, since the bit-plane coding is used for the entropy coding, the coefficients to be coded are set in subband units. Alternatively, the encoding process cannot be performed unless the coefficients are once stored in block units obtained by dividing the subband into a predetermined size, so that there is a problem that a large memory for storing the coefficients is required.

Further, when coding each bit plane using arithmetic coding as in the above-described conventional example, the arithmetic processing becomes complicated, requiring a large amount of CPU power, taking a long processing time, and reducing the circuit scale. There is a problem such as becoming larger.

In forming encoded data, in the conventional stepwise encoding method, a code string obtained by encoding each bit plane is temporarily stored in order to generate a final code string. For this reason, a large amount of memory is required for storing bit plane encoded data.

With the recent increase in performance and functionality of personal computers, the above-mentioned problems have been reduced. However, for example, when a large amount of image data must be encoded within a limited time, such as when encoding and storing moving image data captured in real time, it is desirable to solve the above-described problem. It is.

The above-mentioned problems are serious in an apparatus such as a scanner or a digital camera, in which the calculation performance and the memory capacity are restricted.

The present invention has been made in view of the above-mentioned problems, and has as its main object to efficiently perform encoding even when the processing time, memory, and operation cost of the apparatus are restricted. . Specifically, in a situation where image data is decomposed into frequency components, each frequency component is quantized, and the quantized data is subjected to a predetermined entropy encoding to generate encoded data, processing time, memory, operation cost, etc. The purpose of the present invention is to perform encoding that is not affected as much as possible by the constraints.

[0038]

According to an embodiment of the present invention, there is provided an image encoding system including an image input device and an image processing device, wherein the image input device comprises: Image input means for inputting image data, decomposition means for converting the image data into frequency components to obtain conversion coefficients, and entropy conversion coefficients obtained by the decomposition means or quantization values of the conversion coefficients. A first entropy encoding unit for encoding, wherein the image processing apparatus decodes encoded data obtained by the first entropy encoding unit; and a decoding unit for decoding the encoded data obtained by the decoding. The transform coefficient or the quantized value is set to the first
And a second entropy encoding unit that performs entropy encoding again by a method different from the above entropy encoding unit.

[0039]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (First Embodiment) A preferred embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing an image coding system according to a first embodiment of the present invention.
In the figure, 101 is an image input unit, 102 is a discrete wavelet transform unit, 103 is a coefficient quantization unit, 104 is a Golomb encoding unit,
105 and 106 are communication interface units, 107 is a Golomb code decoding unit, 108 is a bit plane coding unit, 109 is a code string forming unit, and 110 is a code output unit.

In this embodiment, the luminance value of one pixel is 8
The description will be made on the assumption that monochrome image data represented by bits is decoded and displayed. However, the present invention is not limited to this, and can also be applied to image data expressing a luminance value with a bit number other than 8 bits such as 4 bits, 10 bits, and 12 bits. Also, the present invention can be applied to color image data in which each pixel is represented by a plurality of color components such as RGB and CMYK or luminance and chromaticity / color difference components such as YCrCb. In this case, each component in the color image data may be regarded as monochrome image data. It is also apparent that the present invention can be applied to a case of encoding index data in which the state of each pixel is represented by multi-valued information, such as a case where the color of each pixel is represented by an index value of a predetermined color table.

Hereinafter, the operation of each section in the present embodiment will be described in detail with reference to FIG.

The image encoding system according to the present embodiment has an image input device 1 having a limited computational performance and a limited memory, and an image processing system having a higher amount of memory than the image input device 1 and having a higher computational performance. It is composed of two devices. The image input device 1 can be, for example, a scanner, a digital camera, or the like. In FIG. 1, an image input unit 101, a discrete wavelet transform unit 102, a coefficient quantization unit 103, a Golomb encoding unit 10
4. It is composed of the communication interface unit 105.

On the other hand, the image processing device 2 can be a personal computer or a general-purpose computer in recent years, a dedicated device, an image processing board, or the like. The communication interface unit 106, the Golomb code decoding unit 107, the bit plane coding unit 108 , A code sequence forming unit 109 and a code output unit 110.

First, the image data P (x, y) to be encoded by the image encoding system of this embodiment is
Are input in the order of raster scan. This image input section
An imaging unit 101 such as a scanner or a digital camera includes an imaging device such as a CCD and various image adjustment circuits such as gamma correction and shading correction.

The discrete wavelet transform unit 102 performs a two-dimensional discrete wavelet transform while appropriately storing the pixel data P (x, y) input from the image input unit 101 in an internal buffer (not shown), and performs LL, LH1, HL1 , HH1, LH2, HL2, and HH2 are decomposed into seven subbands, and the coefficients of each subband are output. Hereinafter, the coefficient of each subband is represented as C (S, x, y). S represents a subband, and is one of LL, LH1, HL1, HH1, LH2, HL2, and HH2. Also, x and y indicate the coefficient positions in the horizontal and vertical directions, with the coefficient position at the upper left corner in each subband being 0,0.

The two-dimensional discrete wavelet transform is realized by applying a one-dimensional transform (filter process) to each of the horizontal and vertical directions. As shown in FIG. 5, LL, LH1, HL1, HH1,
The process of decomposing into seven subbands of LH2, HL2, and HH2 is known, and the description is omitted here.

The coefficient quantization unit 103 calculates the coefficient C (S, x, y) of each subband generated by the discrete wavelet transform unit 102.
Is quantized using a quantization step delta (S) defined for each subband. Assuming that the quantized coefficient value is represented as Q (S, x, y), the quantization process performed by the coefficient quantization unit 103 is represented by the following equation.

Q (S, x, y) = sign {C (S, x, y)} × floor {|
C (S, x, y) | / delta (S)}

Here, sign {I} is a function representing the sign of the integer I, and returns 1 if I is positive and -1 if I is negative. Also,
floor {R} represents the largest integer value not exceeding the real number R.

The Golomb coding unit 104 performs Golomb coding on the quantized coefficient Q (S, x, y) of each subband (hereinafter, simply referred to as a coefficient) in units of one line of each subband. Generate a column. Hereinafter, the Golomb code sequence for the coefficient on the m-th line of the subband S is represented as GCS (S, m).

The Golomb code is a coding method that can generate codes corresponding to several types of probability distributions by setting non-negative integer values as coding targets and appropriately setting coding parameters (k parameters). . In the present embodiment, the k parameter whose code length is the shortest for each line of the coefficient of each subband is selected, and the coefficient Q (S, x, y) is calculated as a non-negative integer value (V and After conversion to
This is Golomb-coded with the selected k parameters. .

[0053]

It is assumed that the selected k parameter is included in the code string and transmitted. The procedure for Golomb-encoding a non-negative integer value V to be encoded with an encoding parameter k is as follows.

First, V is shifted right by k bits to obtain an integer value m. The code for V is composed of a combination of “1” following m “0” and the lower k bits of V. Figure 7 shows k = 0,1,2
Here is an example of the Golomb code in.

FIG. 8 shows a Golomb code string G generated and output by the Golomb coding unit 104 for the coefficient of the m-th line of the subband S.
3 shows the structure of CS (S, m). As is clear from the figure, the code string GCS (S, m) includes an identifier specifying the subband S, a line number m, and the selected k parameter.

The communication interface 105 outputs the code string GCS (S, m) output from the Golomb coding unit 104 to the outside of the device via a communication line. This communication interface
Interface to network such as Ethernet, analog telephone line, ISDN line, or SCSI, I
Interfaces to buses such as DE and ISA.

The communication interface 106 receives a code string GCS (S, m) from the outside of the apparatus (image input apparatus 1) via a communication line and sends it to the Golomb code decoder 107. This communication interface supports Ethernet, analog telephone lines, IS
An interface to a network such as a DN line or an interface to a bus such as SCSI, IDE, or ISA.

Golomb code decoding section 107 decodes Golomb coded data using coding parameter k included in code sequence GCS (S, m), and decodes a coefficient value on the m-th line of subband S. .

The decoding of the Golomb code is performed in a procedure reverse to that of the encoding. First, the number of consecutive “0” s is checked from the decoding start point, and is held at an integer value m. K bits are extracted immediately after “1” that terminates the “0” sequence, m is shifted left by k bits, and an OR operation is performed on the extracted k bits to decode the non-negative integer value V. The coefficient Q (S, x, y) is decoded from the non-negative integer value V by the following operation.

[0061]

The decoded coefficient Q (S, x, y) is sent to bit plane encoding section 108.

The bit plane encoding unit 108 performs bit plane encoding (arithmetic encoding for each bit plane) on the coefficient values Q (S, x, y) decoded by the Golomb code decoding unit 107 in subband units. Generate a column. Bit plane encoding unit
Reference numeral 108 has a buffer (not shown) inside and can store decoded coefficient values. Subband S coefficient Q (S, x,
y) is bit-plane coded, and CS (S, NBP (S) -1) to CS (S,
Since the process of generating (0) is the same as that described above, the description is omitted here.

In the code sequence forming unit 109, when the encoding of the coefficients of all the subbands is completed by the bit plane coding unit 108 and all the code sequences are stored in the internal buffer, they are stored in the internal buffer in a predetermined order. The code sequence is read out, necessary additional information is inserted, a final code sequence to be output from the present encoding device is formed, and output to the code output unit 110.

The final code string generated by the code string forming unit 109 is composed of a header, level 0, level 1 and level 2
It is composed of encoded data hierarchized. Level 0 encoded data is composed of a code string of CS (LL, NBP (LL) -1) to CS (LL, 0) obtained by encoding the coefficients of the LL subband. Level 1 is a code sequence CS (LH1, NBP (LH1) -1) to CS obtained by encoding the coefficients of each subband of LH1, HL1, and HH1.
(LH1,0), CS (HL1, NBP (HL1) -1) to CS (HL1,0), and CS
(HH1, NBP (HH1) -1) to CS (HH1,0). Also,
Level 2 is a code string CS (LH2, NBP (LH2) -1) to CS (LH2, 0), CS (LH2, NBP (LH2) -1) obtained by encoding the coefficients of each subband of LH2, HL2, HH2.
(HL2, NBP (HL2) -1) to CS (HL2,0) and CS (HH2, NBP (HH
2) -1) to CS (HH2,0).

FIG. 3 shows the structure of the code string generated by the code string forming unit 109.

Code output section 110 outputs the code string generated by code string forming section 109 to the outside of the apparatus. This code output unit 11
0 is, for example, a storage device such as a hard disk or a memory, a network line interface, or the like.

As described above, in the image input apparatus 1 having low performance, after performing discrete wavelet transform and quantization, instead of performing predetermined entropy coding (arithmetic coding for each bit plane) which should be performed originally, Simple entropy coding (Golomb coding) is performed, and the obtained coded data is transferred (transmitted) to the high-performance image processing apparatus 2, and the predetermined coding is performed after decoding on the image processing apparatus 2 side. (Bit plane coding) is performed to form a final coded data sequence, so that efficient coding can be performed while avoiding limitations due to the device performance of the image input device 1 as much as possible.

(Second Embodiment) FIG. 9 is a block diagram showing an image coding system according to a second embodiment of the present invention. Portions common to the block diagram of FIG. 1 used in the first embodiment are denoted by the same reference numerals, and description thereof will be omitted.

In the figure, reference numeral 901 denotes a moving image input unit;
A next storage device 903 is an encoded data reading unit.

In the present embodiment, a description will be given on the assumption that monochrome moving image data in which the luminance value of one pixel is 8 bits and which is captured in 15 frames per second is 4 seconds, that is, 60 frames are encoded. However, the present invention is not limited to this, and it is also possible to modify the capturing time and the number of captured frames per second.

As in the first embodiment, this embodiment can encode various image data. That is, image data expressing a luminance value by a bit number other than 8 bits such as 4 bits, 10 bits, and 12 bits, or color image data expressing each pixel by a plurality of color components or luminance and chromaticity / color difference components Alternatively, the present invention can also be applied to a case where index data indicating the state of each pixel by multi-value information is encoded.

Hereinafter, the operation of each unit of the image coding system according to the present embodiment will be described in detail with reference to the block diagram of FIG. The image encoding system according to the present embodiment encodes 15 frames of moving image data per second input from the moving image input unit 901 in real time using simple entropy coding, and performs bit plane coding ( Recompression is performed using arithmetic coding for each bit plane.

Except that the image to be encoded is a moving image, that a secondary storage device 902 is provided in place of the communication interfaces 105 and 106, and that encoded data is stored in the secondary storage device 902, This is almost the same as the image coding system described in the first embodiment.

First, as described above, a moving image (60 frames) corresponding to 4 seconds at 15 frames per second from the moving image input unit 901
Is entered. The moving image input unit 901 sends an input image to the discrete wavelet transform unit 102 frame by frame.

The 1 sent to the discrete wavelet transform unit 102
The frame, that is, one piece of image data, is subjected to a discrete wavelet transform by the discrete wavelet transform unit 102 in the same manner as in the first embodiment, is quantized by the coefficient quantizing unit 103, and G
The olomb encoding unit 104 performs entropy encoding.

Code string CS generated by Golomb coding section 104
(S, m) is temporarily stored in the secondary storage device 902. At this time, the code string CS (S, m) is stored collectively for each frame.

When all the image data of 60 frames input from the moving image input unit 901 has been encoded, the encoded data reading unit 903 encodes the data from the secondary storage device 902 from the first frame to the last frame one frame at a time. The data is read out and passed to Golomb code decoding section 107.

In the same manner as in the first embodiment, the Golomb code decoder 107 restores the coefficient Q (S, x, y) from the code sequence GCS (S, m), and the bit plane encoder 108 (S, x, y) is bit-plane coded in sub-band units to generate and output bit-plane coded data CS (S, NBP (S) -1) to CS (S, 0).
Further, the code sequence forming unit 109 rearranges the code sequence on a frame basis to generate a final code sequence, and generates the final code sequence.
Store in 02.

With the above processing, the moving image data input from the moving image input unit 901 is encoded on a frame basis, and the encoded data is generated in the secondary storage device. Once, it is encoded in real time using simple entropy encoding, stored in the secondary storage device, and then, after decoding, bit-plane encoded again, ensuring excellent processing performance at the time of capture and excellent compression performance. Thus, encoded moving image data suitable for stepwise transmission can be generated.

(Third Embodiment) In the first and second embodiments, the spatial resolution is gradually improved (or the decoded image size is increased) with the transmission of the code sequence as the final code sequence. ).

In the present embodiment, as shown in FIG. 10, a request from a decoding device 5 different from the image input device 3 and the image processing device 4 is transmitted to the code sequence forming unit 1102. As in the second embodiment, a code string whose spatial resolution gradually improves (hereinafter referred to as spatial resolution scalable) and a code string whose image quality gradually improves (hereinafter referred to as SNR scalable) are switched and output. .

FIG. 10 is a block diagram showing an image coding system according to the second embodiment of the present invention. No.
Portions common to the block diagram of FIG. 1 used in the first embodiment are denoted by the same reference numerals, and description thereof will be omitted.

In the figure, reference numeral 1001 denotes a secondary storage device, and 1002 denotes a code string forming unit. As shown in the figure, the image encoding system of the present embodiment has an image input device 3 having almost the same functions as the image encoding system (image input device 1, image processing device 2) of the first embodiment. And an image processing device 4.

The image input device 3 is exactly the same as the image input device 1 of the first embodiment. Further, the image processing device 4 adds the function of the secondary storage device 1001 to the image processing device 2 and includes a code sequence forming unit 1002 having a different operation in place of the code sequence forming unit 109. It has been changed so that a request from the decryption device 5 is input.

The image encoding system according to the present embodiment encodes a large number of image data, stores the bit plane encoded data output from the bit plane encoding unit 108 in the secondary storage device 1001, In response to a request from the decoding device 5, encoded data of predetermined image data is generated and output with a scalability according to the request from the decoding device 5.

The operation of the image input device 3 is the same as that of the image input device 1 according to the first embodiment, and the description is omitted. Hereinafter, the operation of the image processing apparatus 4 will be described.

In the same manner as in the image processing apparatus 2 of the image encoding system according to the first embodiment, the Golomb code decoder 107 restores the coefficient Q (S, x, y) from the code sequence GCS (S, m). Then, Q (S, x, y) is bit-plane coded in sub-band units (arithmetic coding for each bit plane) by the bit-plane coding unit 108,
Bit plane coded data CS (S, NBP (S) -1) to CS (S, 0)
Generate and output

Bit plane coded data CS (S, NBP (S) -1) to CS (S, S
0) is temporarily stored in the secondary storage device 1001. At this time,
Bit plane coded data CS (S, NBP (S) -1) to CS (S, 0)
Are stored together for each image.

The decoding apparatus 5 outputs a plurality of pieces of encoded image data stored in the secondary storage device 1001 from the decoding apparatus 5.
A code sequence forming unit 100 is formed by pairing an image identification signal i for specifying an image required by the scalability selection signal s for specifying whether the spatial resolution is scalable or SNR scalable.
Sent to 2.

The image identification signal may be an image name, a serial number of the image, or the like as long as it can uniquely designate an image, and is not particularly limited. The scalability selection signal s is set to 0 for spatial resolution scalable, and set to 1 for SNR scalable.

The code stream forming unit 1002 reads the bit plane coded data relating to the image data specified by the request signal (i, s) from the decoding device 5 from the secondary storage device 1001, and codes with the specified scalability. Generate and output a column. If the scalability selection signal s is 0, as in the image encoding system according to the first embodiment, a diagram hierarchically divided into a header and three levels 0, 1, and 2 is shown.
The encoded data having the structure shown in FIG. 3 is generated.

On the other hand, if the scalability selection signal s is 1, the coded data of each bit plane is shifted from the upper bit plane to the lower bit plane as shown in FIG.
Coded data is constructed by repeatedly arranging 0, level 1, and level 2 in this order.

The encoded data generated by the code string forming unit 1002 is output from the code output unit 110 to the decoding device 5. The code output unit 110 is a functional unit similar to the image coding system according to the first embodiment, and can be applied to a storage device such as a hard disk or a memory, an interface of a network line, or the like.

With the above processing, according to the present embodiment, in addition to the effects of the first embodiment, it is possible to form a flexible code string according to the request of the decoding device 5.

(Modification) The present invention is not limited to the above embodiment. For example, in the above-described first to third embodiments, an example of encoding using the discrete wavelet transform has been described. However, the discrete wavelet transform is not limited to the one used in the present embodiment. The type of filter and the adaptation method may be changed.
For example, a filter having a longer number of taps than a 9/7 filter or the like may be used, or a two-dimensional discrete wavelet transform may be repeatedly applied to low frequency subbands. Furthermore, a method suitable for encoding image data hierarchically other than the discrete wavelet transform may be used.
An encoding method based on other sequence transform methods such as DCT and Hadamard transform may be applied. Also, the coding method of the coefficient is not limited to the above embodiment, for example, MQ-Coder, etc., may be applied to arithmetic coding methods other than QM-Coder, other entropy code The method of conversion may be applied.

In the above-described embodiment, the Golomb code is used as simple entropy coding. However, a coding method other than the Golomb code such as a Huffman code may be applied.

Further, for simplicity, in each of the above embodiments, bit plane coding in units of subbands has been described. However, in order to enhance random access, each subband is further divided into small blocks. Bit plane coding may be applied to each small block unit. In addition, the bit plane is classified into a plurality of sub-bit planes according to the state of the neighboring coefficient of the coefficient of interest,
Encoding may be performed in a plurality of passes.

Further, the code string forming unit 100 of the third embodiment
In 2, a method of arranging bit plane coded data of each level in bit plane units from an upper bit plane to a lower bit plane as a code string construction method (SNR scalable) in which image quality gradually improves with code transmission However, an efficient transmission order may be obtained from the balance between the code amount of each bit plane data and the image quality improvement effect, and the bit plane encoded data may be arranged in this order.

The image data may be divided into a plurality of tiles in advance, and the encoding process may be performed for each tile.

The present invention can be applied to a single device (for example, a copying machine, a facsimile machine, Digital camera or the like). That is, in FIG. 1 described above, the image input device and the image processing device are described separately, but the image input device 1 and the image processing device 2 may be a single device.

The present invention is not limited only to an apparatus and a method for realizing the above-described embodiment.
Computer (CPU or
(MPU) is supplied with software program code for implementing the above-described embodiment, and the computer of the system or apparatus operates the various devices according to the program code to implement the above-described embodiment. Included in the scope of the invention.

In this case, the program code of the software implements the functions of the above-described embodiment, and the program code itself and means for supplying the program code to the computer, specifically, A storage medium storing the program code is included in the scope of the present invention.

As a storage medium for storing such a program code, for example, a floppy disk, hard disk, optical disk, magneto-optical disk, CD-ROM, magnetic tape, nonvolatile memory card, ROM and the like can be used.

In addition to the case where the computer controls various devices in accordance with only the supplied program codes, not only the functions of the above-described embodiment are realized, but also the program codes operate on the computer. Such a program code is also included in the scope of the present invention when the above-described embodiment is realized in cooperation with an operating system (OS) or another application software.

Further, after the supplied program code is stored in a memory provided in a function expansion board of a computer or a function expansion unit connected to the computer, the function expansion board or the function is stored based on the instruction of the program code. The present invention also includes a case where a CPU or the like provided in the extension unit performs part or all of the actual processing, and the above-described embodiment is realized by the processing.

[0107]

As described above, according to the present invention, coding can be performed efficiently even in a state where the processing time, memory, operation cost, and the like of the apparatus are restricted. In particular, in a situation where image data is decomposed into frequency components, each frequency component is quantized, and the quantized data is subjected to predetermined entropy encoding to generate encoded data, processing time, memory, operation costs, and the like of the apparatus are reduced. It is possible to perform encoding that is as insensitive to constraints as possible.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image encoding system according to a first embodiment.

FIG. 2 is a block diagram showing a configuration of an example of a conventional image encoding system.

FIG. 3 is a diagram showing an example of final encoded data.

FIG. 4 is a diagram for explaining a state of a two-dimensional wavelet transform.

FIG. 5 is a diagram illustrating subband division.

FIG. 6 is a flowchart illustrating an encoding process in a bit plane encoding unit 108;

FIG. 7 is a diagram illustrating an example of a Golomb code in a case where coding parameters k = 0, 1, 2, and 3.

FIG. 8 is a diagram illustrating a configuration of Golomb encoded data output from the Golomb encoding unit 104.

FIG. 9 is a block diagram illustrating a configuration of an image encoding system according to a second embodiment.

FIG. 10 is a block diagram showing a configuration of an image encoding system according to a third embodiment.

FIG. 11 is a diagram showing a configuration of encoded data output by an image encoding system according to a third embodiment.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 image input device 2 image processing device 101 image input unit 102 discrete wavelet transform unit 103 coefficient quantization unit 104 Golomb coding unit 105, 106 communication interface 107 Golomb code decoding unit 108 bit plane coding unit 109 code sequence forming unit 110 code Output section

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Hiroki Kishi 3-30-2 Shimomaruko, Ota-ku, Tokyo F-term in Canon Inc. (reference) 5C059 KK08 KK11 KK41 MA22 MA23 MA24 MA32 MA34 MA35 MC11 MC38 ME01 ME02 ME11 PP01 PP04 PP15 PP16 PP17 SS11 SS14 SS20 UA02 UA06 UA11 5C078 AA04 BA21 CA27 CA31 DA00 DA01 DA02

Claims (8)

[Claims] 1. An image encoding system comprising an image input device and an image processing device, said image input device comprising: an image input unit for inputting image data; and converting the image data into a frequency component; And a first entropy encoding unit that entropy-encodes the transform coefficient obtained by the decomposer or the quantized value of the transform coefficient. Decoding means for decoding the coded data obtained by the first entropy coding means, and the transform coefficient or the quantization value obtained by the decoding by a method different from that of the first entropy coding means. A second entropy encoding unit for performing entropy encoding again. 2. The image encoding system according to claim 1, wherein the conversion into the frequency component is a wavelet transform. 3. The method according to claim 1, wherein the first entropy encoding means includes a Go
The image encoding system according to claim 1, wherein lomb encoding is performed. 4. The image encoding system according to claim 1, wherein said second entropy encoding means is bit plane encoding. 5. The image coding system according to claim 4, wherein said second entropy coding means performs arithmetic coding for each bit plane. 6. An image input unit for inputting image data, a decomposition unit for converting the image data into a frequency component to obtain a conversion coefficient, a conversion coefficient obtained by the decomposition unit, or a quantum of the conversion coefficient First entropy encoding means for entropy encoding a coded value; decoding means for decoding encoded data obtained by the first entropy encoding means; and the transform coefficient obtained by the decoding or the An image encoding apparatus, comprising: second entropy encoding means for entropy encoding a quantized value again by a method different from that of the first entropy encoding means. 7. An image input step of inputting image data; a decomposition step of converting the image data into frequency components to obtain a conversion coefficient; and a conversion coefficient obtained by the decomposition means or a quantum of the conversion coefficient. A first entropy encoding step of entropy encoding the coded value; a decoding step of decoding the encoded data obtained in the first entropy encoding step; and the transform coefficient obtained by the decoding or the A second entropy encoding step of entropy encoding the quantized value again by a method different from the first entropy encoding step. 8. An image input step of inputting image data, a decomposition step of converting the image data into frequency components to obtain a conversion coefficient, a conversion coefficient obtained by the decomposition means, or a quantum of the conversion coefficient A first entropy encoding step of entropy encoding the coded value; a decoding step of decoding the encoded data obtained in the first entropy encoding step; and the transform coefficient obtained by the decoding or the A storage medium storing a second entropy encoding step of entropy encoding a quantization value again by a method different from the first entropy encoding step in a computer-readable state.