JP2007020010A - Image encoder and method of encoding image, and image decoder and method of decoding image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image encoder and an image encoding method that improve reversible compression performance in images, and are capable of compressing images with high image quality and a high compression ratio, and to provide a corresponding decoder and an image decoding method. <P>SOLUTION: A control unit 13 controls the operation of a prediction section 11 and a JPEG2000 encoder 12 according to an external switching signal. For example, when for example characters/graphic images are encoded, the predictor 11 predicts and processes an inputted image, and the JPEG2000 coder 12 allows encoding processing after an EBCOT to be made in a bypass mode. As a result, the amount of coding can be restrained while the image quality is maintained by reversible encoding even in the characters/graphic images in which the amount of coding increases in JPEG2000. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a technique for encoding and decoding an image.

  As a technique for encoding and decoding still images, a JPEG (Joint Photographic Expert Group) method has been widely used. Recently, JPEG2000 (hereinafter abbreviated as J2K) has been newly standardized. FIG. 4 is a block diagram showing an outline of the encoding process in J2K. In the figure, 41 is a pre-processing unit, 42 is a wavelet transform unit, 43 is a quantization unit, 44 is an EBCOT unit, and 45 is a post-processing unit. The preprocessing unit 41 appropriately performs processing such as DC level shift, color conversion, and tile division, and the wavelet conversion unit 42 performs discrete wavelet conversion. The coefficient generated by the wavelet transform is quantized by the quantization unit 43. The processing of the quantization unit 43 is optional. Thereafter, the EBCOT unit 44 performs entropy coding by bit modeling and arithmetic coding. This EBCOT (Embedded Block Coding with Optimized Truncation) is also a feature of J2K. Finally, processing such as rate control and bit stream generation is performed by the post-processing unit 45 to obtain an encoded image.

  Unlike JPEG, J2K seamlessly realizes lossless encoding and lossy encoding, and realizes high image quality and a high compression rate. However, as described above, since J2K performs wavelet transform, which is one of orthogonal transform coding, in principle, it is not suitable for compression of ideal edges included in many characters and graphics images, and has lossless compression performance. There is a problem of being bad.

  For example, in Patent Document 1, an image is divided into blocks, whether or not a character exists in the block is determined, lossless encoding is performed when the character exists, and orthogonal transform encoding is performed when the character does not exist. It is carried out. Thereby, the reproducibility of characters and the like can be improved, and the compression rate can be increased. Similarly, a technique of separating images such as characters and graphics from photographs and encoding them using an encoding method suitable for each is being standardized as J2K part 6 (hereinafter referred to as MRC: Mixed Raster Content). This method may increase the lossless compression rate of the text / graphics image. However, these systems have a problem that the scale of the apparatus increases because an encoding system suitable for each plane is applied.

JP-A-6-152982

  The present invention has been made in view of the above-described circumstances. An image encoding apparatus that can improve the lossless compression performance of an image, and that can be compressed with high image quality and a high compression rate, and a corresponding image decoding apparatus. It is another object of the present invention to provide an image decoding method.

  The present invention is characterized in that a prediction process is performed on each pixel of an input image, and the prediction result is encoded by the EBCOT method. This EBCOT method is a method used in J2K. By instructing the bypass mode to the JPEG2000 encoding means, the processing before J2K EBCOT is bypassed and the processing after EBCOT is performed. Can do.

  Further, JPEG2000 encoding means may be provided so that normal J2K encoding and encoding by executing processing after EBCOT in J2K bypass mode after prediction processing can be switched from the outside. it can. For example, a character and a graphics image (region) can be encoded by executing prediction processing and processing after EBCOT, and other images (region) can be encoded by J2K.

  In the present invention, the encoded data encoded by the EBCOT method after performing the prediction process as described above is first decoded by the EBCOT method, and the inverse prediction process is performed on the decoded data, and the original image is converted. It is characterized by restoring. This EBCOT decoding is also used for J2K decoding. By instructing the JPEG2000 decoding means to bypass mode, the processing after J2K EBCOT is bypassed, and the processing up to EBCOT is performed. Can be done.

  As in the case of encoding, JPEG2000 decoding means is provided, and normal J2K decoding and decoding for performing prediction processing after executing EBCOT and subsequent processing in J2K bypass mode are input code data. It can be configured to switch depending on

  According to the present invention, since the prediction process is performed and the encoding is performed by the EBCOT method, the compression rate can be greatly improved even for the lossless compression of characters and graphics images that are unsuitable in J2K. is there. Further, only by providing the predicting means or the inverse predicting means, encoding or decoding can be performed using the JPEG2000 encoding means or the JPEG2000 decoding means as they are. Furthermore, if JPEG2000 encoding means or JPEG2000 decoding means is provided, normal J2K encoding and decoding and prediction processing and encoding and decoding by EBCOT can be switched appropriately and used for encoding. It can be used properly according to the image.

  FIG. 1 is a block diagram illustrating a configuration example of encoding according to an embodiment of the present invention. In the figure, 11 is a prediction unit, 12 is a JPEG2000 encoding unit, and 13 is a control unit. The prediction unit 11 performs a prediction process for each pixel of the input image. The prediction method is arbitrary. As the prediction unit 11, for example, a JPEG-LS predictor, a source coder of an existing predictive coding method, such as PNG or DPCM, can be applied.

  The JPEG2000 encoding unit 12 has a configuration schematically shown in FIG. 4, for example, and encodes an image according to J2K. J2K defines a bypass mode in which the processing up to the quantization unit 43 in FIG. 4 is bypassed and encoding is performed by the processing after the EBCOT unit 44, and an operation in this bypass mode is also possible.

  The control unit 13 performs control to switch encoding processing in accordance with an external switching signal. When encoding the input image by J2K, control is performed so that the prediction unit 11 is bypassed and the JPEG 2000 encoding unit 12 performs the encoding process. Further, when the prediction unit 11 is switched by a switching signal, an input image is input to the prediction unit 11 to perform prediction processing, and the JPEG2000 encoding unit 12 performs an operation in the bypass mode. An instruction is made to perform the encoding process after the EBCOT unit 44 on the data after the prediction process.

  When JPEG 2000 encoding is performed by the JPEG2000 encoding unit 12, discrete wavelet transform, which is one of orthogonal transform encoding methods, is performed as shown in FIG. If lossless compression is performed on an image for which an edge is to be stored, the compression rate is reduced and the code amount is increased. In the present invention, such an image is switched to use the prediction unit 11 and the EBCOT method by a switching signal, thereby performing a prediction process (a process by a source coder for predictive coding) without performing orthogonal transform coding. Thus, the code amount can be suppressed while maintaining the reversibility, and the compression rate can be improved.

  Switching between the J2K method and the encoding method according to the present invention using prediction and EBCOT may be performed for each region in the image by separately providing image region recognition means, for example, for each image. Which encoding method is used may be left separately, may be included in the header portion of the code data, or may be identified by the extension of the file name. Alternatively, a part of the bits of the image data to be encoded may be used as switching information, and the switching information may be added when a prediction result is input from the prediction unit 11 to the JPEG2000 encoding unit 12.

  In the configuration example shown in FIG. 1, the JPEG2000 encoding unit 12 is used as it is, and the operation is switched by the control unit 13, but the result predicted by the prediction unit 11 is always the bypass mode of the JPEG2000 encoding unit 12. 4, only the configuration after the EBCOT unit 44 shown in FIG. 4 in the JPEG2000 encoding unit 12 is installed, and the prediction result of the prediction unit 11 is input to the EBCOT unit 44. Good.

  FIG. 2 is a block diagram showing a configuration example at the time of decoding in the embodiment of the present invention. In the figure, 21 is a JPEG2000 decoding unit, 22 is an inverse prediction unit, and 23 is a control unit. The JPEG2000 decoding unit 21 decodes the code data according to J2K. Further, in accordance with J2K regulations, it is also possible to perform an EBCOT process as a bypass mode and bypass the subsequent processes.

  The reverse prediction unit 22 performs reverse prediction processing on the output when the JPEG2000 decoding unit 21 is operated in the bypass mode, and restores the original image. In the inverse prediction process, a process corresponding to the prediction process at the time of encoding is performed.

  If the input code data is encoded by JPEG2000, the control unit 13 performs control so that the JPEG2000 decoding unit 21 decodes and obtains the original image, and the input code data performs the prediction process. In the case of code data encoded by the EBCOT method after being performed, the JPEG2000 decoding unit 21 is instructed to perform bypass mode decoding, and the result is subjected to reverse prediction processing by the reverse prediction unit 22 to perform the original prediction. Control to get the image.

  Which process is selected depends on, for example, obtaining information about the encoding method at the time of encoding from the file name of the encoded data, its extension, the header of the file, etc. If so, that information can be obtained. Further, if it exists in the code data, the JPEG2000 decoding unit 21 may be operated in the bypass mode to extract the data, and the encoding method may be switched by referring to the specific bit. Of course, the decoding method may be switched from the outside.

  With such a configuration, the code data encoded by the configuration shown in FIG. 1 is switched in its encoding method (whether it is encoded by the J2K method or the EBCOT method after the prediction process), and the original image is changed. Can be restored. At this time, if the encoding is performed by the EBCOT method after the prediction process, it is possible to reproduce characters and graphics images faithfully because of lossless compression, and to reproduce an image from code data with a small code amount. it can.

  If only the encoded data obtained by encoding the prediction result in the bypass mode of the JPEG2000 encoding unit 12 is decoded, only the configuration up to the EBCOT unit in the JPEG2000 decoding unit 21 is installed, and the output from the EBCOT unit is provided. May be configured to be reversely predicted by the reverse prediction unit 22.

  FIG. 3 is a block diagram showing an application example of the present invention. In the figure, 31 is a decomposer, 32 is an encoding unit, 33 is an image storage unit, 34 is a decoding unit, 35 is an image output unit, and 36 is a compression / decompression control unit. In this example, an example in which an image is internally encoded and temporarily stored in the image storage unit 33 in the printer, and then decoded and an image output unit 35 forms an image is shown. The encoding unit 32 has the configuration of the present invention shown in FIG. 1, and the decoding unit 34 has the configuration of the present invention shown in FIG.

  The decomposer 31 receives data (PDL data) described by, for example, PDL (Print Description Language) and performs a process of drawing an image according to a drawing command. During the rendering process, it is determined whether the rendered image is an image with many characters and graphics or an image with many photographs, and a switching signal is output to the compression / decompression control unit 36. In addition, the rendered image is output to the encoding unit 32.

  The image storage unit 33 stores the code data encoded by the encoding unit 32.

  The compression / decompression control unit 36 has functions of a control unit of the encoding unit 32 and a control unit of the decoding unit 34, and the operations of the encoding unit 32 and the decoding unit 34 according to a switching signal from the decomposer 31. To control.

  The outline of the operation will be briefly described below. When the PDL data is input to the decomposer 31, the decomposer 31 performs a drawing process according to the PDL data. At this time, it is determined whether the drawn image is a character or graphic image or other image including raster data such as a photograph. Of course, this determination method is arbitrary, and it suffices to determine whether it is appropriate to encode in J2K or whether it is appropriate to perform encoding after the J2K EBCOT by performing a prediction process. This determination result is passed to the compression / expansion control unit 36 as a switching signal.

  The compression / decompression control unit 36 receives the switching signal from the decomposer 31 and sets the encoding unit 32. When the switching signal indicates that the encoding target is, for example, a character or graphic image, and the encoding processing is instructed by the prediction processing and the processing after J2K EBCOT, the compression / decompression control unit 36 performs the prediction of the encoding unit 32. And the JPEG2000 encoding unit is set to operate in the bypass mode.

  Then, the decomposer 31 inputs the rendered image to the encoding unit 32. In the encoding unit 32, the input image is subjected to prediction processing by the prediction unit, and the prediction result is encoded by processing after the EBCOT unit of the JPEG2000 encoding unit. The code data is stored in the image storage unit 33. At this time, the fact that the code data is encoded by the prediction process and the processes after EBCOT is stored in the image storage unit 33 in association with the code data or included in the code data.

  When the switching signal indicates a J2K encoding, for example, other than a character or graphic image, the compression / decompression control unit 36 bypasses the prediction unit of the encoding unit 32, and the JPEG2000 encoding unit normally Set to operate in the mode. Then, the decomposer 31 inputs the rendered image to the encoding unit 32, and the image is encoded by the JPEG2000 encoding unit of the encoding unit 32 using the J2K method. The code data is stored in the image storage unit 33. At this time, the fact that the code data is encoded by the J2K system is stored in the image storage unit 33 in association with the code data or included in the code data. Or By not adding such information, it may be distinguished from code data encoded by prediction processing and EBCOT.

  When the image output unit 35 is ready, the code data stored in the image storage unit 33 is decoded by the decoding unit 34 and supplied to the image output unit 35. At this time, the compression / decompression control unit 36 sets the decoding unit 34 according to the code data. That is, when the encoded data is encoded by the encoding unit 32 using the prediction process and EBCOT, the compression / decompression control unit 36 operates the JPEG2000 decoding unit of the decoding unit 34 in the bypass mode, and outputs the result. Is set to be subjected to reverse prediction processing by the reverse prediction unit. Then, the code data is read from the image storage unit 33, the JPEG2000 decoding unit of the decoding unit 34 performs the decoding process up to EBCOT, and the inverse prediction unit performs the reverse prediction process. The image obtained by the reverse prediction process is used for image formation in the image output unit 35 after performing other processes (not shown).

  When the encoded data is encoded by the J2K system, the compression / decompression control unit 36 sets the JPEG2000 decoding unit of the decoding unit 34 to operate in the normal mode and bypasses the inverse prediction unit. . Then, the encoded data is read from the image storage unit 33, and the decoding process is performed by the JPEG2000 decoding unit of the decoding unit 34 to obtain the original image. The inverse prediction unit is bypassed, and the image decoded by the JPEG2000 decoding unit is output. The obtained image is subjected to other processing (not shown) and then used for image formation in the image output unit 35.

  In this way, since the encoding process is switched according to the image drawn by the decomposer 31, the optimum code or process is applied to each image, and the code amount can be suppressed. Accordingly, since the data amount of the code data stored in the image storage unit 33 is reduced, the capacity of the image storage unit 33 can be suppressed, or more images can be stored in the image storage unit 33. Also, the code amount can be suppressed by using the J2K encoding unit and decoding unit as they are and adding the prediction unit and the inverse prediction unit.

  In the above description, until encoding is performed by the encoding unit 32 and stored in the image storage unit 33, and until encoded data is extracted from the image storage unit 33, decoded by the decoding unit 34, and sent to the image output unit 35. Are described separately. In this way, in addition to operating both separately, the image storage unit 33 can also be used as a buffer for timing adjustment. In this case, the compression / decompression control unit 36 can also be configured to perform a series of processing by setting the decoding unit 34 together with the encoding unit 32 by the switching signal from the decomposer 31.

  Of course, it is not limited to the application to the printer as described above, but is used for encoding and transmitting at the time of communication and decoding on the receiving side to obtain an image, or encoding and saving, and decoding when used The present invention can be applied to various uses.

It is a block diagram which shows the structural example of the encoding in one Embodiment of this invention. It is a block diagram which shows the structural example at the time of the decoding in one embodiment of this invention. It is a block diagram which shows the application example of this invention. It is a block diagram which shows the outline | summary of the encoding process in J2K.

Explanation of symbols

DESCRIPTION OF SYMBOLS 11 ... Prediction part, 12 ... JPEG2000 encoding part, 13 ... Control part, 21 ... JPEG2000 decoding part, 22 ... Inverse prediction part, 23 ... Control part, 31 ... Decomposer, 32 ... Encoding part, 33 ... Image storage part 34 ... Decoding unit, 35 ... Image output unit, 36 ... Compression / decompression control unit, 41 ... Pre-processing unit, 42 ... Wavelet transformation unit, 43 ... Quantization unit, 44 ... EBCOT unit, 45 ... Post-processing unit.

Claims (12)

  An image encoding apparatus comprising: a prediction unit that performs a prediction process on each pixel of an input image; and an entropy encoding unit that encodes an output of the prediction unit using an EBCOT method.   2. The image encoding apparatus according to claim 1, wherein the entropy encoding unit is operated by instructing a bypass mode to a JPEG2000 encoding unit.   A prediction unit that performs a prediction process on each pixel of the input image, a JPEG2000 encoding unit that performs a JPEG2000 encoding process on the input image, and the JPEG2000 encoding unit according to an external signal. An image encoding apparatus comprising control means for switching encoding or encoding in the prediction means and JPEG2000 encoding means in bypass mode.   An entropy decoding unit that decodes code data encoded by the EBCOT method after performing the prediction processing by the EBCOT method, and an inverse that performs reverse prediction processing on the output from the entropy decoding unit to restore the original image An image decoding apparatus comprising a prediction unit.   5. The image decoding apparatus according to claim 4, wherein the entropy decoding unit is operated by instructing a bypass mode to a JPEG2000 decoding unit.   JPEG2000 decoding means for decoding an image encoded by the JPEG2000 method, reverse prediction means for performing reverse prediction processing on the prediction-processed data and restoring the original image, and input code data encoded by JPEG2000 In the case of code data encoded by the EBCOT method after being decoded by the JPEG2000 decoding means if it has been converted to the JPEG2000 decoding means, the JPEG2000 decoding means performs decoding in the bypass mode and then performs the reverse processing. An image decoding apparatus comprising: control means for controlling to perform reverse prediction processing by the prediction means.   An image encoding method, wherein a prediction process is performed on each pixel of an input image by a prediction unit, and the prediction result is encoded by an entropy encoding unit by an EBCOT method.   8. The image encoding method according to claim 7, wherein the encoding by the EBCOT method is performed by instructing a JPEG2000 encoding unit to operate in a bypass mode.   A first encoding process for encoding the input image by JPEG2000 encoding means using JPEG2000, a prediction process for each pixel of the input image by the prediction means, and a JPEG2000 encoding means for the prediction result And a second encoding process in which the encoding process is performed in the JPEG 2000 bypass mode is switched by a control unit in accordance with an external signal.   Code data encoded by the EBCOT method after performing the prediction process is decoded by the EBCOT method by the entropy decoding unit, and the reverse prediction process is performed on the decoded data by the reverse prediction unit to restore the original image An image decoding method characterized by the above.   11. The image decoding method according to claim 10, wherein the decoding by the EBCOT method is performed by instructing the JPEG2000 decoding unit to operate in a bypass mode. If the input code data is encoded by the JPEG2000 system, the code data is decoded by the JPEG2000 decoding means, and the input code data is encoded by the EBCOT system after performing the prediction process In such a case, the image decoding method is characterized in that the control means controls the reverse prediction process by the reverse prediction means after the JPEG2000 decoding means performs decoding in the bypass mode.

Images