JP2008263589A - Encoding device, code conversion device, code expansion device, encoding method, code conversion method, and code expansion method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an encoding device and an encoding method that can prevent a memory consumption and a processing speed from decreasing while accessing data in the order of resolution by arranging codes in an order different from that of a conventional one. <P>SOLUTION: The encoding device generates a code stream by encoding input data and includes a dividing unit 31 that divides the input data into at least one rectangular area, a conversion unit 33 that converts data of the rectangular area divided by the dividing unit 31 into conversion data as specified, an ordering unit 35 that adds the priority order to the conversion data, an encoding unit 37 that rearranges the conversion data based on the priority order added by the ordering unit 35 and encodes the conversion data into the code stream, and marker coupling units 39a to 39d that couple a marker that can identify the priority order to the header part of each conversion data in the code stream generated by the encoding unit 37. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an encoding apparatus and an encoding method for encoding an image and generating a compressed data stream. In particular, the present invention relates to an encoding apparatus and an encoding method that can prevent a decrease in the amount of memory used and the processing speed while enabling access in the order of resolution by making the order of arrangement of code data different from the conventional method. is there.

In recent years, a method of dividing an image into rectangular areas and independently compressing the rectangular areas has been proposed. Each rectangular area in this method is subjected to color conversion, resolution conversion, area division, and bit plane conversion. Then, arithmetic processing is performed on the data subjected to these processes to generate code data, and the code data is aligned to generate a compression code.
The order of arranging the code data is called a progression order, and five types are prepared. One of these progression orders is selected and the code data is aligned. When the user wants to display an image from a low resolution, if the image is compressed with a progression order called RPCL, the display device can easily display the images in the order of resolution.
Furthermore, if the image is divided into a plurality of rectangular areas, the code data of all the rectangular areas are arranged in the order of resolution, and even if the image is code data divided into a plurality of rectangular areas, display from a low resolution is possible. It is possible.
As a prior art, Patent Document 1 records information on data transmitted to the client terminal device in the past as a history. When a transmission request for data in a logical unit in a tile necessary for obtaining a desired image is received from the client terminal device, the type of progression order used in the client terminal device by referring to the above history Is determined. A technique for determining the transmission order of logical unit data in each tile to be transmitted to the client terminal device according to the determination result, and transmitting the logical unit data in the tile to the client terminal device based on the determined transmission order is disclosed. Yes.
Japanese Patent Application Laid-Open No. 2005-12585

However, when generating code data in the order of resolution, it is necessary to simultaneously process the code data of all rectangular areas in order to arrange the code data of all rectangular areas in the order of resolution. This process requires a large amount of processing time and the amount of memory used by the code data. As a result, there have been problems such as a decrease in processing speed and unstable operation of the device due to the reasons described above.
Also, when code data is formed by hardware, there is a problem that the cost is very high for the same reason.
The present invention has been made in view of the above-described conventional problems. By making the code data alignment order different from that of the conventional method, the amount of memory used and the processing speed are reduced while enabling access in the order of resolution. It is an object of the present invention to provide an encoding device, an encoding method, a computer program, and a recording medium that can prevent the above.

In order to achieve the above-mentioned object, the invention according to claim 1 is an encoding device for encoding input data and generating a code stream, and dividing means for dividing the input data into at least one rectangular area A conversion unit that performs predetermined conversion on the rectangular area divided by the division unit and converts the rectangular area into conversion data, an ordering unit that assigns a priority order to the conversion data, and the priority assigned by the ordering unit Arranging the converted data based on the order, encoding processing means for generating a code stream by performing an encoding process, and at the head of each conversion data in the code stream generated by the encoding processing means, And a marker combining unit that combines markers whose priority order can be identified.
The invention according to claim 2 is characterized in that the conversion data is any one of frequency band data, color component data, bit plane aggregate data, and small area data. And

According to a third aspect of the present invention, there is provided a code conversion device for converting a code stream generated by dividing image data into rectangular areas and encoding each rectangular area, wherein at least one of the code streams is converted. Boundary detection means for reading area data and detecting a boundary portion of the area data, and marker combination for combining a marker capable of identifying the area data at the head of at least one area data detected by the boundary detection means And a code conversion device.
According to a fourth aspect of the present invention, there is provided a code conversion device for converting a code stream generated by dividing image data into rectangular areas and encoding each rectangular area. Identification means for identifying area data; setting means for setting the alignment order of the area data identified by the identification means; and reordering means for rearranging code data based on the alignment order set by the setting means And a code conversion device that converts code data relating to the rectangular area into a code stream that is combined and arranged.

According to a fifth aspect of the present invention, in the code conversion device according to the third or fourth aspect, the area data is any one of frequency band data, color component data, bit plane aggregate data, and small area data. It is characterized by.
According to a sixth aspect of the present invention, there is provided a code expansion device for dividing image data into rectangular areas and expanding a code stream to which an identifier that can identify a boundary of the data is added, wherein the rectangular area is extracted from the code stream. Identifying means for identifying the area data; setting means for setting the alignment order of the area data identified by the identifying means; and reordering means for rearranging the code data based on the alignment order set by the setting means And decompression means for decompressing the code stream;
A code expansion apparatus having the following features.

The invention according to claim 7 is characterized in that the region data is any one of frequency band data, color component data, bit plane set data, and small region data. And
The invention according to claim 8 is an encoding method for encoding input data and generating a code stream, wherein the input data is divided into at least one rectangular area, and the division is performed in the division step. A conversion step of performing predetermined conversion on the rectangular area thus converted to conversion data, an ordering step of assigning a priority order to the conversion data, and the conversion data based on the priority order assigned in the ordering step And a marker capable of identifying the priority order at the head of each conversion data in the code stream generated in the encoding step, and generating a code stream by performing an encoding process. And a marker combining step for combining.

The invention according to claim 9 is the encoding method according to claim 8, wherein the conversion data is any one of frequency band data, color component data, bit plane set data, and small area data. And
The invention according to claim 10 is a code conversion method for converting a code stream generated by dividing image data into rectangular areas and encoding each rectangular area, and at least one of the code streams. A boundary detection step for reading region data and detecting a boundary portion of the region data, and a marker combination for combining a marker capable of identifying the region data at the head of at least one region data detected in the boundary detection step And a code conversion method including the steps.

The invention according to claim 11 is a code conversion method for converting a code stream generated by dividing image data into rectangular areas and encoding each rectangular area, wherein An identification step for identifying data, a setting step for setting the alignment order of the data identified in the identification step, and a rearrangement step for rearranging the code data based on the alignment order set in the setting step. And a code conversion method for converting the code data related to the rectangular area into a code stream arranged in combination.
The invention according to claim 12 is the code conversion method according to claim 10 or 11, wherein the area data is any one of frequency band data, color component data, bit plane aggregate data, and small area data. It is characterized by.

According to a thirteenth aspect of the present invention, there is provided a code decompression method for dividing image data into rectangular areas and decompressing a code stream to which an identifier capable of identifying the boundary of the area data is attached. An identification step for identifying data in the area; a setting step for setting the alignment order of the data identified in the identification step; a rearrangement step for rearranging the code data based on the alignment order set in the setting step; A code decompression method comprising: a decompressing step of decompressing the code stream.
The invention according to claim 14 is characterized in that the region data is any one of frequency band data, color component data, bit plane aggregate data, and small region data. And

According to the present invention, a code stream that can be easily accessed with any one element of resolution, color components, and layers can be generated at high speed and in a memory-saving manner.
Further, according to the present invention, code data that is easy to access in a partial area in a tile can be generated at high speed and with a small amount of memory.
Further, according to the present invention, code data divided into tile parts can be decompressed at high speed and with a small amount of memory.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
The present invention is realized by a program in a general-purpose computer such as a PC (Personal Computer), an information terminal having an image display function, or a built-in microcomputer of an information processing device such as a mobile phone. It is.
Such an embodiment will be described with reference to FIG. FIG. 1 is a block diagram of one embodiment of a computer according to the present invention.
As shown in FIG. 1, the computer includes a CPU (Central Processing Unit) 1, a first storage device 2, a display device 3, a second storage device 4 such as a hard disk, an input device 5 such as a keyboard and a pointing device, A communication interface (I / F) 6, an input / output interface (I / F) 7, and a medium drive device 8 for reading and writing various recording media (magnetic disk, optical disk, magneto-optical disk, semiconductor storage element, etc.) . They are configured to be connected to each other via a system bus 9.

Then, a recording medium on which a program 10 for implementing the present invention (hereinafter referred to as an image processing program) is recorded is read by the medium drive device 8, and the image processing program 10 is stored in the second storage device 7. Alternatively, it is fetched from the communication interface 11 via the network and stored in the second storage device 4.
The image processing program 10 is read from the second storage device 4 to the first storage device 2. By executing the image processing program 10 by the CPU 1, the image processing according to the present invention is realized on the computer.

Here, code data of a still image or a moving image to be processed by the encoding device of the present invention is stored in, for example, a server on a network. Code data is taken from this server via the communication interface 6. Further, the image data is taken in from an imaging device such as a digital camera connected to the input / output interface 7. There is also a form in which code data is taken from the second storage device 4.
According to one typical aspect, the image processing program 10 takes a form that operates in cooperation with another program 12 (hereinafter referred to as a viewer / browser) for browsing an image called a viewer or a browser. .

In the present invention, an image is divided into rectangular areas, and encoding processing is performed independently for each rectangular area. Conventionally, a function called a tile part is used when arranging code data divided into rectangular areas in the order of resolution. If this tile part function is used, the code data of all the rectangular areas can be arranged in the order of resolution. However, the code data requires a large amount of memory for the compression / decompression processing for both software and hardware, and the processing time is very long.
Therefore, in the system according to the embodiment of the present invention, the code data of all the rectangular areas are not rearranged, the process is performed in units of rectangular areas, and the tile part identifier is inserted at the resolution break.
As shown in FIG. 2, each rectangular area is subjected to color space conversion, wavelet conversion, quantization, entropy encoding, and tag processing during compression to generate code data.

Ｕ＝Ｒ−Ｇ
Ｖ＝Ｂ−Ｇ
・・・・式１
但し：

は、ｘを超えない最大の整数を表す。
なお、上記の色変換に先立ち、画像に対してＤＣレベルシフトが行われる。ＤＣレベルシフトは、例えば入力される画像がＲＧＢ信号の８ｂｉｔである場合は、式２にて表される。

・・・・式２
色変換処理後の信号は、ＤＷＴ処理部３０３において各々のコンポーネント毎にＤＷＴ（離散ウェーブレット変換）が行われ、ウェーブレット係数が生成される。ＤＷＴは２次元にて行われるが、通常は、リフティング演算と呼ばれる演算方法により１次元フィルタ演算のコンボリューションにて実施される。１次元の変換式を式３に示す。

・・・・式３
但し、Ｌ（ｋ）は低周波成分、Ｈ（ｋ）は高周波成分、ｘ（ｋ）は画素値、ｋは座標を表す。
なお、ＤＷＴは、ダウンサンプリングを伴うため、上記Ｌ（ｋ）、Ｈ（ｋ）は入力画像と比較して１／２の解像度となる。 The input image data is divided into rectangular areas by the dividing unit 301. The division of the rectangular area is arbitrary, and when the image is not divided into rectangular areas, the number of rectangular areas is handled as one. Next, the color conversion processing unit 302 performs color conversion processing on the image data divided by the dividing unit 301. For example, reversible color conversion processing is performed according to the following Equation 1.

U = RG
V = B-G
.... Formula 1
However:

Represents the largest integer not exceeding x.
Prior to the color conversion, a DC level shift is performed on the image. The DC level shift is expressed by Equation 2 when the input image is 8 bits of RGB signals, for example.

.... Formula 2
The signal after the color conversion process is subjected to DWT (discrete wavelet transform) for each component in the DWT processing unit 303 to generate wavelet coefficients. Although DWT is performed in two dimensions, it is usually performed in a convolution of a one-dimensional filter operation by an operation method called lifting operation. A one-dimensional conversion formula is shown in Formula 3.

.... Formula 3
However, L (k) represents a low frequency component, H (k) represents a high frequency component, x (k) represents a pixel value, and k represents a coordinate.
Since DWT involves downsampling, the above L (k) and H (k) have half the resolution compared to the input image.

FIG. 3 is a diagram illustrating wavelet coefficients that have been divided into octaves.
In the DWT, directional components called four subbands of LL, HL, LH, and HH are generated for each decomposition (decomposition) level. Then, by performing DWT recursively on the subband LL, the composition level is increased to a lower resolution. The coefficient of one decomposition level with the highest resolution is represented as 1HL, 1LH, 1HH, and hereinafter represented as 2HL, 2LH... NHH. The figure is an example of division into three decomposition levels. On the other hand, the resolution level is referred to as 0, 1, 2, 3 because of the low resolution coefficient in the opposite direction to the decomposition level.
A subband in each decomposition level can be divided into areas called precincts to form a set of codes. Encoding is performed in predetermined block units called code blocks.

FIG. 4 is a diagram illustrating the relationship among tiles, precincts, and code blocks in the intra-tile wavelet coefficients.
The wavelet coefficients output from the DWT processing unit 303 are subjected to scalar quantization by the quantization processing unit 304 in FIG. 2, but when performing reversible transformation, the scalar quantization is not performed or is quantized by “1”. To do. Also, in post-quantization at the subsequent stage, the same effect as that of quantization can be obtained. In scalar quantization, parameters can be changed for each tile.
In FIG. 2, the entropy encoding unit 305 performs entropy encoding on the quantized data output from the quantization processing unit 304. In the entropy encoding method in JPEG2000, a subband is divided into rectangular areas called code blocks (however, when the size of the subband area is equal to or smaller than the code block size, it is not divided) and encoded in units of code blocks.

Further, after the data in the code block is decomposed into bit planes as shown in FIG. 5, the three bit paths (Significance propagation path, Magnitude refinement path) are represented according to the state representing the influence of the transform coefficient on the image quality. , Clean up path), and each is encoded by an arithmetic encoding method called MQ coder. The bit plane has an MSB side, and the encoding pass has an increasing importance (contribution to image quality) in the order of significance propagation, magnitude refinement, and cleanup. The end of each path is also called a truncation point, which is a unit that can be used for truncating a post-quantization code at a later stage.
The post-quantization unit 306 in FIG. 2 truncates the code of the entropy-encoded code as necessary. Note that when it is necessary to output a reversible code, post-quantization is not executed. JPEG2000 has a configuration that requires no feedback to control the code amount (1 pass encoding) because the code amount can be truncated after encoding. The code data after post-quantization is the code shown in FIG. The stream generation processing unit 307 rearranges codes and adds headers based on a predetermined progressive order (decoding order of code data) to complete a code stream for the tile.

FIG. 6 is a diagram showing the entire code stream by layer progression in JPEG2000. The whole code is composed of a main header and a plurality of tiles obtained by dividing an image. A tile code is composed of a plurality of layers in which a tile header and an intra-tile code are divided into code units called layers (details will be described later). Layer 0, layer 1,... Are lined up. The configuration of the layer code is composed of a layer tile header and a plurality of packets, and the packet is composed of a packet header and code data.
The packet is a minimum unit of code data, and is composed of code data of one layer in one precinct at one resolution level (decomposition level) in one tile component.

Next, the progressive order in JPEG2000 will be described.
In JPEG2000, by changing the priority order of the four image elements of image quality (layer (L)), resolution (R), component (C), and position (precinct (P)), the following five progressions are performed. Is defined.
LRCP progression Since decoding is performed in the order of precinct, component, resolution level, and layer, the image quality of the entire image is improved each time the layer index advances, and the progression of image quality can be realized. Also called layer progression.
-RLCP progression Since the decoding is performed in the order of precinct, component, layer, and resolution level, the progression of resolution can be realized.
RPCL progression Since decoding is performed in the order of layer, component, precinct, and resolution level, it is a progression of resolution as in RLCP, but the priority of a specific position can be increased.
-PCRL progression Since decoding is performed in the order of layer, resolution level, component, and precinct, decoding of a specific part is prioritized, and progression of spatial position can be realized.
CPRL Progression Since decoding is performed in the order of layer, resolution level, precinct, and component, for example, in the progressive decoding of a color image, it is possible to realize component progression that first reproduces a gray image.

FIG. 7A is a diagram schematically showing the progressive order of LRCP progression (hereinafter referred to as layer progression), and FIG. 7B is a diagram illustrating the progressive order of RLCP progression or RPCL progression (hereinafter referred to as resolution progression).
7 (a) and 7 (b), the horizontal axis represents the decomposition level (the higher the number, the lower the resolution), and the vertical axis represents the layer number (the higher the number, the higher layer, and the upper layer code in the lower layer) By adding and decoding, reproduction with higher image quality becomes possible. In the figure, a filled rectangular figure represents a code in the decomposition level and layer, and its size schematically represents the ratio of the code amount. The dotted arrows in the figure indicate the code order.

FIG. 7A shows a code order for decoding in layer progression, and decoding of all resolutions of the same layer number is performed to decode the upper layer of the next stage. If it sees at a wavelet coefficient level, it will decode from the high-order bit of a coefficient, and the progression which improves an image quality gradually is realizable. FIG. 7B shows a code order for decoding in resolution progression, in which all layers at the same decomposition (resolution) level are decoded and decoding at the next decomposition (resolution) level is performed. Thus, it is possible to realize a progression in which the resolution is gradually improved.
Through the above processes, code data (packets) of a plurality of images classified by resolution, color component, layer, and precinct are generated. Then, in the tag processing unit (not shown), the code data is rearranged in a specific progression order as shown in FIG.

FIG. 8 shows an RPCL progression order when 12 packets are generated for each rectangular area, classified into 4 resolutions (R0 to R3), 3 color components (C0 to C2), 1 precinct, and 1 layer. It is explanatory drawing which shows the code data at the time of rearrangement. In FIG. 8, MH is a main header, and SOT is a marker, which is added to the head of code data in a rectangular area.
When the code data is rearranged in the order shown in FIG. 8, since the code data is rearranged in the order of resolution in each rectangular area, the image after the decompression process can be displayed so that the resolution is sequentially improved. However, in JPEG2000, when an image is divided into rectangular areas, the tile part function is used.
When this tile part function is used, it is possible to expand and display the code data so that the resolution is sequentially improved from the low resolution over the entire rectangular area. However, this tile part function requires access to the code data of all the rectangular areas, and the amount of used memory is greatly increased, and an increase in processing time due to accessing all becomes a problem. When the tile part function is used, code data as shown in FIG. 9 is generated. FIG. 9 shows an example of tile parts for each resolution.

Incidentally, the viewer / browser 11 of FIG. 1 opens an image display window on the screen of the display device 3 and displays an image there.
Usually, in addition to designation of an image to be displayed (designation of a file name and URL), the operation of a pointing device such as a mouse included in the input device 5 changes the vertical and horizontal sizes of the image display window, changes the image display magnification, It is possible to scroll an image larger than the display window and specify a specific area (position). Conditions such as the vertical and horizontal sizes of the image display window, the image display magnification, and the position of the image area are notified to the image processing program 10. In addition, various conditions relating to moving images or still images can be specified by using buttons or slide bars displayed in the control area, and the viewer / browser 11 can display resolution, image quality, color components, position, important area, frame rate. The image processing program 10 is notified of conditions relating to the code amount, processing time, and the like.
The image processing program 10 encodes a still image or a moving image so that the still image or the moving image can be displayed in accordance with the conditions such as the decoding processing capacity and the size of the image display window and in accordance with the user's intention. The code of the data is selectively decoded to reproduce the image data, and the viewer / browser 11 displays the image data on the image display window. Thus, the selective decoding process of the code of the encoded data is performed by the image processing program 10, but the display of the image on the screen of the display device 5 is controlled by the viewer / browser 11.

Next, an embodiment of an encoding apparatus and an encoding method achieved by applying the image processing program 10 to the computer shown in FIG. 1 according to the present invention will be described.
FIG. 10 is a functional configuration block diagram of the encoding apparatus achieved by applying the image processing program 10 to the computer shown in FIG.
The encoding apparatus shown in FIG. 10 encodes an input image and generates a compressed data stream. 10 includes a dividing unit 31, a frequency converting unit 33, an ordering unit 35, a compressing unit 37, a marker combining unit 39a, a marker combining unit 39b, a marker combining unit 39c, and a marker combining unit 39d. The dividing unit 31 divides an input image into rectangular areas. In the present embodiment, the image is divided into a plurality of rectangular areas, but it is needless to say that the number of rectangular areas may not be plural if the image is regarded as one rectangular area. The frequency converting unit 33 performs frequency conversion on the rectangular area divided by the dividing unit 31 to generate a plurality of frequency band data. The ordering unit 35 assigns a priority order to each frequency band data generated by the frequency conversion unit 33. The compression unit (encoding processing means) 37 arranges the frequency band data based on the priority order given by the ordering unit 35 and generates a code stream by performing compression processing on the frequency band data. The marker combining unit 39a combines a marker whose priority order can be identified with the head portion of the band data in the code stream of each band generated by the compression unit 37.

As described above, in this embodiment, the code data of all tiles is not rearranged as in the prior art, but the process is executed in units of each rectangular area, and the identifier is inserted at the band break of the code stream.
As another embodiment, the compression unit 37 includes a marker combining unit 39b that combines a marker whose priority order can be identified at the head of the color component data in the code stream of each band.
Furthermore, as another embodiment, the compression unit 37 includes a marker combining unit 39c that combines a marker capable of identifying the priority order at the head of the image quality data in the code stream of each band.
Furthermore, as another embodiment, the compression unit 37 is configured to include a marker combining unit 39d that combines markers capable of identifying the priority order at the head of the small area data in the code stream of each band.

FIG. 11 is a flowchart of the image processing program 10 in the above encoding apparatus.
Hereinafter, it demonstrates along a processing flow.
Step 101: The following encoding conditions are set by referring to the setting information and default settings by the user. Here, only the setting according to the present invention will be described.
The resolution divided as tile parts is set as follows in the resolution division level number RM, the color component number CM, the precinct division number PM, the layer number LM, and the tile part division boundary resolution Ri.
Ri = {R0, R1, R2, R3}
In this description, the progression order RPCL will be described, but other progression orders can be implemented. Further, the tile number T1, resolution level R, precinct T2, color component C, and layer number L of the image being processed are set to initial values.

Step 102: An image to be encoded is read.
This step may be performed at the time of encoding in step 110, or may be performed for each specific size before encoding.
Step 103: It is determined whether or not the tile T1 of the image being processed has reached the number of encoded tiles TM. If reached, the process moves to END, and if not, the process moves to Step 104.
Step 104: It is determined whether or not the resolution level R has reached the encoded resolution level number RM. If it has reached, the process moves to Step 111, and if not, the process moves to Step 105.
Step 105: It is determined whether or not the resolution level to be processed is included in the set of the set Ri of the tile parts. If the resolution level to be processed is included in the set of the boundary Ri of the tile part, the process proceeds to step 106, and otherwise, the process proceeds to step 107. Here, the tile part boundary Ri may be changed for each tile, and in a tile without tile part division, the tile part boundary set Ri is an empty set.
Step 106: A marker indicating a tile part boundary is assigned. This marker includes a parameter indicating the length of the tile part, but the value is replaced when the tile part length is determined. Therefore, the value is initially indefinite.

Step 107: It is determined whether or not the precinct T2 has reached the number of encoded precincts PM. When the number of encoded precincts PM has been reached, the routine proceeds to step 112, and otherwise, the routine proceeds to step 108.
Step 108: It is determined whether or not the color component C has reached the number of encoded color components CM. If the number of encoded color components CM has been reached, the process proceeds to step 113, and if not, the process proceeds to step 109.
Step 109: It is determined whether or not the layer number L has reached the encoding layer number LM. If the number of coding layers LM has been reached, the process moves to step 114, and if not, the process moves to step 110.
Step 110: Encode the target resolution R, precinct P, color component C, and layer number L data, and go to Step 15.
Step 111: The resolution level R is incremented to the initial value R0, and the tile number T1 is incremented by one.

Step 112: The precinct number T2 is incremented to the initial value P0 and the resolution level R is incremented by one.
Step 113: The color component number C is incremented to the initial value C0, and the precinct number T2 is incremented by one.
Step 114: Increment the layer number L to the initial value L0 and the color component number C by one.
Step 115: Increment the layer number L by one.

FIG. 12 shows tile part codes generated as a result of the processing based on the above flowchart.
12 is an explanatory diagram of tile part codes generated by the image processing shown in FIG.
The tile part codes generated in the prior art have tile parts arranged for each resolution as shown in FIG. However, if the present invention is implemented, an identifier is inserted for each rectangular area as shown in FIG. 12, and code data is collected. That is, the code data for each rectangular area is divided into tile parts for each resolution.
Since the above code data only needs to be processed only for the target rectangular area, the amount of memory used is much smaller than that of the conventional method. That is, when the present invention is implemented, the processing time is greatly shortened as compared with the prior art.

FIG. 13 is a functional block diagram of the code conversion apparatus achieved by applying the image processing program 10 to the computer shown in FIG.
As shown in FIG. 13, this code conversion apparatus divides image data into rectangular areas, performs frequency conversion, reads at least one area data of a code stream generated by compression processing, A band detection unit 41a that detects a boundary part and a marker combination unit 43a that combines a marker capable of identifying the band at the head of at least one band data detected from the band detection unit 41a. Yes.
Further, in the modification of the code conversion apparatus, as shown in FIG. 13, the image data is divided into rectangular areas, color conversion processing is performed, and compression processing is performed. A band detection unit 41b that reads color component data and detects a boundary portion of the color component, and a marker that can identify the color component is coupled to the head of at least one color component data detected from the band detection unit 41b. The marker coupling portion 43b is included.

Further, in another modification of the code conversion apparatus, as shown in FIG. 13, the image data is divided into rectangular areas, the rectangular areas are divided into bit planes, and a set of a plurality of bit planes is formed. A band detector 41c that reads at least one bit plane set data of a code stream from which a stream is generated and detects a boundary portion of the set, and a head portion of at least one bit plane set detected by the band detector 41c, And a marker combining unit 43c that combines markers capable of identifying a set of bit planes.
Further, in another modification of the code conversion device, as shown in FIG. 13, the image data is divided into rectangular areas, the divided rectangular areas are further divided into small areas, and the data is generated for each small area. A band detection unit 41d that reads at least one small region data of the code stream and detects a boundary portion of the small region, and a small region is identified at the head of at least one small region data detected by the band detection unit 41d. And a marker coupling portion 43d that couples possible markers.

FIG. 14 is a schematic process flowchart of the image processing program 21 in the code conversion apparatus having the above functional configuration.
Hereinafter, it demonstrates along a processing flow.
Step 201: The following encoding conditions are set by referring to the setting information and default settings by the user. Here, only the setting according to the present invention will be described.
The tile number T1, resolution level R, precinct T2, color component C, and layer number L of the image being processed are set to initial values.
The resolution to be divided as tile parts into the tile part division boundary resolution Ri is set as follows.
Ri = {R0, R1, R2, R3}
Step 202: A determination is made whether the code to be converted has reached the end. If it is the end, the process proceeds to END. Otherwise, go to step 203.
Step 203: Read the header of the code to be converted.
Step 204: Based on the header information read in Step 203, it is determined whether or not it is a tile part division boundary. If it is a boundary, the process moves to Step 205, and if not, the process moves to Step 206.

Here, the boundary is determined based on the read header information when the resolution R ′ processed immediately before and the resolution R being processed are different from each other.
If it is a resolution boundary, it is compared whether the resolution R is included in the set of tile part division boundaries Ri, and if it is included, it becomes a tile part division boundary.
Step 205: A tile part boundary marker SOT is inserted.
This marker includes a parameter indicating the length of the tile part, but the value is replaced when the tile part length is determined. Therefore, the value is initially indefinite.
Step 206: The code data corresponding to the packet length read in the header is skipped.
As a result of the above processing, the code shown in FIG. 8 is converted into the code shown in FIG.
The tile part code generated in the prior art has tile parts arranged for each resolution. However, the code data according to the present invention is grouped for each tile, and the tile data is divided into tile parts for each resolution. is doing.
Since the code data need only be processed for the target tile when generating the code data, the amount of memory used is less than that of the conventional method, resulting in a reduction in processing time.

Next, a code conversion apparatus and a code conversion method embodying the present invention will be described.
FIG. 15 is a functional block diagram of a code conversion apparatus achieved by applying the image processing program 10 to the computer shown in FIG.
As shown in FIG. 15, in this code conversion apparatus, image data is divided into rectangular regions, frequency conversion is performed, compression processing is performed, and a frequency band boundary can be identified as at least one frequency band boundary. An identification unit 51a for identifying a frequency band of a rectangular region from a code stream generated with an identifier, a setting unit 53a for setting the alignment order of rectangular frequency data identified by the identification unit 51a, and a setting unit 53a The reordering unit 55 reorders the code data generated by the compression process in the set order of arrangement, and is configured to generate a code stream in which code data related to at least one rectangular area is combined and arranged. .
Further, as shown in FIG. 15, the modification of the code conversion apparatus divides image data into rectangular areas, performs color conversion processing, performs compression processing, and sets a color at the boundary of at least one color component. A setting for setting an arrangement order of an identification unit 51b for identifying a color component of a rectangular area and a rectangular color data identified by the identification unit 51b from a code stream generated with an identifier that can identify a component boundary. Unit 53b and rearrangement unit 55 that rearranges the codes in the arrangement order set by setting unit 53b, and generates a code stream in which code data related to at least one rectangular area is combined and arranged. Yes.

Further, as shown in FIG. 15, another modification example of this code conversion apparatus divides image data into rectangular areas, performs bit plane conversion, forms a set of bit planes, and performs compression processing. An identification unit 51c for identifying a bit plane set in a rectangular area from a code stream generated by assigning an identifier capable of identifying the boundary of the bit plane set to the boundary of at least one bit plane set, and the identification unit 51c A setting unit 53c that sets the arrangement order of the rectangular bit plane set data; and a rearrangement unit 55c that rearranges the codes in the arrangement order set by the setting unit 53c, and combines the code data related to at least one rectangular area. The code stream is arranged to be generated.
Further, as shown in FIG. 15, another modification of the code conversion apparatus divides the image data into rectangular areas, further divides the divided rectangular areas into small areas, and performs compression processing. An identification unit 51d for identifying a small region of a rectangular region from a code stream generated by assigning an identifier that can identify the boundary of the small region to the boundary of two small regions, and the rectangular small region data identified by the identifying unit 51d A code stream that includes a setting unit 53d that sets an arrangement order and a rearrangement unit 55 that rearranges codes in the arrangement order set by the setting unit 53d, and is arranged by combining code data related to at least one rectangular area Is generated.

FIG. 16 is a schematic process flowchart of the image processing program 10 in the code conversion apparatus having the above functional configuration.
Hereinafter, it demonstrates along a processing flow.
Step 301: The following encoding conditions are set by referring to the setting information and default settings by the user. Here, only the setting according to the present invention will be described.
Setting of a set Ti of tile numbers to combine tile parts Ti = {T0, T1, T2, T3}.

Further, the tile number T, tile part number TP, code pointer FP1, and code pointer FP2 of the image data being processed are set to initial values.
Step 302: It is determined whether or not the code conversion process has been completed. If the sign change process is completed, the process proceeds to END. Otherwise, the process proceeds to step 303.
Step 303: The tile part header is read according to the code reading position FP, and the tile number T is acquired.
Step 304: It is determined whether or not the tile number T is included in Ti. If the tile number T is included in Ti, the process proceeds to step 305. If the tile number T is not included in Ti, the process proceeds to step 310.
Step 305: Record the code reading position FP1.
FP2 = FP1

The length of the tile part is acquired from the tile part header, the tile part data is read, and combined with the generated code data.
Step 306: It is determined whether all tile parts of the tile T being processed have been read. If all tile parts have been read, the process proceeds to step 309. If all tile parts have not been read, the process proceeds to step 307.
Step 307: The tile parts are sequentially searched from the FP, and the tile parts having the same tile number are searched.
Step 308: Combine the found tile part with the generated code data.
Step 309: Reset the pointer FP1 to FP2.
Here, FP2 = FP1
Step 310: Copy the tile part being read into the generated code data.

As a result of the above steps, the code data shown in FIG. 9 is converted into the code data shown in FIG.
The tile part code generated in the prior art has tile parts arranged for each resolution. However, the code data according to the present invention is grouped for each tile, and the tile data is divided into tile parts for each resolution. Is done.
Since the code data need only be processed for the target tile when generating the code data, the amount of memory used is less than that of the conventional method, resulting in a reduction in processing time.

Next, a code expansion apparatus and code expansion method embodying the present invention will be described.
FIG. 17 is a functional block diagram of a code decompression apparatus achieved by applying the image processing program 10 to the computer shown in FIG.
As shown in FIG. 17, in this code decompression apparatus, an image data is divided into rectangular areas, subjected to frequency conversion, compression processing is performed, and an identifier that can identify a frequency band boundary as at least one frequency band boundary. Is expanded. That is, the identification unit 61a for identifying the frequency band of the rectangular area from the code stream, the setting unit 63a for setting the alignment order of the rectangular frequency data identified by the identification unit 61a, and the alignment order set by the setting unit 63a are encoded. It has a rearrangement unit 65 that rearranges data and a decompression unit 67 that decompresses the code stream, and converts the code data related to at least one rectangular area into code data arranged and further expands the code data. .
Further, in this modification of the code expansion device, as shown in FIG. 17, the image data is divided into rectangular areas, subjected to color conversion processing, compression processing, and color components at the boundary of at least one color component. The code stream to which an identifier that can identify the boundary is attached is decompressed. That is, the identifying unit 61b for identifying the color component of the rectangular area from the code stream, the setting unit 63b for setting the alignment order of the rectangular color data identified by the identifying unit 63b, and the alignment order set by the setting unit 63b are encoded. A rearrangement unit 65 that rearranges the code stream, and a decompression unit 67 that decompresses the code stream. The code data relating to at least one rectangular area is converted into code data that is combined and arranged, and further decompressed.

Further, in another modification of this code decompression apparatus, as shown in FIG. 17, the image data is divided into rectangular areas, bit-plane converted, a set of bit planes is formed, and compression processing is performed. A code stream in which an identifier that can identify a boundary of a bit plane set is attached to a boundary of at least one bit plane set is converted. That is, an identification unit 61c for identifying a bit plane set of a rectangular area from the code stream, a setting unit 63c for setting the alignment order of the rectangular bit plane set data identified by the identification unit 61c, and an alignment set by the setting unit 63c A configuration that includes a rearrangement unit 65 that rearranges codes in order and a decompression unit 67 that decompresses the code stream, converts the code data related to at least one rectangular area into code data that is arranged, and further decompresses the code data It is.
Further, in another modification of this code expansion apparatus, as shown in FIG. 17, the image data is divided into rectangular areas, the divided rectangular areas are further divided into small areas, and compression processing is performed. A code stream in which an identifier that can identify the boundary of a small area is added to the boundary of two small areas is expanded. That is, an identification unit 61d for identifying a small region of a rectangular region from the code stream, a setting unit 63d for setting the alignment order of the rectangular small region data identified by the identification unit 61d, and an alignment order set by the setting unit 63d. It has a rearrangement unit 65 for rearranging codes and a decompression unit 67 for decompressing the code stream, and converts the code data related to at least one of the rectangular areas into code data arranged and further expands the code data. is there.

FIG. 18 is a schematic processing flowchart of the image processing program 21 for a still image in the code expansion device.
Hereinafter, it demonstrates along a processing flow.
Step 401: The following encoding conditions are set by referring to the setting information and default settings by the user. Here, only the setting according to the present invention will be described.
Setting of a set Ti of tile numbers to combine tile parts.
Ti = {T0, T1, T2, T3}.

Further, the tile number T, tile part number TP, code pointer FP1, and code pointer FP2 of the image data being processed are set to initial values.
Step 402: It is determined whether or not the code data expansion processing is completed. If it is completed, the process proceeds to END. Otherwise, the process proceeds to step 403.
Step 403: The tile part header is read from the code reading position FP, and the tile number T is acquired.
Step 404: It is determined whether or not the tile number T is included in Ti. If the tile number T is included in Ti, the process proceeds to step 405. If the tile number T is not included in Ti, the process proceeds to step 410.
Step 405: Record the code reading position FP1.
Here, FP2 = FP1
The length of the tile part is obtained from the tile part header, the tile part data is read and combined with the generated code data.

Step 406: It is determined whether all tile parts of the tile T being processed have been read. If all the tile parts have been read, the process proceeds to step 409. If all the tile parts have not been read, the process proceeds to step 407.
Step 407: The tile parts are sequentially searched from the FP, and the tile parts having the same tile number are searched.
Step 408: Combine the found tile part with the code data to be decompressed.
Step 409: Reset the pointer FP1 to FP2.
Here, FP2 = FP1
Step 410: The read tile data is decompressed.
As a result of the above steps, the code data shown in FIG. 9 is converted into the code data shown in FIG. 12, and decompression processing is performed.
Since the code data need only be processed for the target tile when generating the code data, the amount of memory used is less than that of the conventional method, resulting in a reduction in processing time.

It is a block diagram of a configuration of an embodiment of a computer according to the present invention. It is explanatory drawing which shows the detail of the process at the time of compression. It is a figure showing the wavelet coefficient by which the octave division was carried out. It is a figure which shows the relationship between the tile in a wavelet coefficient in a tile, a precinct, and a code block. It is a figure which shows the relationship between a bit plane and a sub bit plane. It is a figure showing the whole code stream by the layer progression in JPEG2000. FIG. 7A schematically shows the progressive order of LRCP progression (hereinafter referred to as layer progression), and FIG. 7B schematically illustrates the progressive order of RLCP progression or RPCL progression (hereinafter referred to as resolution progression). It is explanatory drawing which shows the example of the code | cord | chord at the time of arranging in RPCL progression order when it classify | categorizes into 4 resolution per 1 tile, 3 color components, 1 precinct, 1 layer, and 12 packets are produced | generated. is there. It is explanatory drawing which shows the example of the tile part for every resolution. It is a functional block diagram of the encoding apparatus achieved by applying the image processing program 10 to the computer shown in FIG. 12 is a schematic processing flowchart of an image processing program 10 in the code decompression apparatus of FIG. 10. It is explanatory drawing of the tile part code | symbol produced | generated by the image processing shown by FIG. FIG. 2 is a functional configuration block diagram of a code conversion apparatus achieved by applying an image processing program 10 to the computer shown in FIG. 1. 14 is a schematic process flowchart of an image processing program 10 in the code conversion apparatus of FIG. 13. FIG. 2 is a functional configuration block diagram of a code conversion apparatus achieved by applying an image processing program 10 to the computer shown in FIG. 1. 16 is a schematic process flowchart of an image processing program 10 in the code conversion apparatus of FIG. FIG. 2 is a functional configuration block diagram of a code expansion device achieved by applying an image processing program 10 to the computer shown in FIG. 1. 18 is a schematic processing flowchart of an image processing program 10 in the code decompression apparatus of FIG.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... CPU, 2 ... 1st memory | storage device, 3 ... Display apparatus, 4 ... 2nd memory | storage device, 5 ... Input device, 7 ... Input / output interface, 8 ... Medium drive device, 9 ... System bus, 10 ... Image Processing program 11 ... Communication interface 12 ... Viewer / browser 31 ... Splitting unit 33 ... Conversion unit 35 ... Ordering unit 37 ... Compression unit 39 ... Marker combining unit 41 ... Band detecting unit 43 ... Marker combining , 51 ... Identification unit, 53 ... Setting unit, 55 ... Rearrangement unit, 61 ... Identification unit, 63 ... Setting unit, 65 ... Rearrangement unit, 67 ... Extension unit

Claims (14)

An encoding device that encodes input data and generates a code stream,
Dividing means for dividing the input data into at least one rectangular area;
Conversion means for performing a predetermined conversion on the rectangular area divided by the dividing means and converting it into converted data;
Ordering means for giving a priority order to the converted data;
An encoding processing means for arranging the converted data based on the priority order assigned by the ordering means and generating a code stream by performing an encoding process;
Marker combining means for combining a marker that can identify the priority order at the head of each conversion data in the code stream generated by the encoding processing means;
An encoding device comprising:   The encoding apparatus according to claim 1, wherein the conversion data is any one of frequency band data, color component data, bit plane set data, and small area data. A code conversion device that divides image data into rectangular areas and converts a code stream generated by encoding each rectangular area,
Boundary detection means for reading at least one region data of the code stream and detecting a boundary portion of the region data;
Marker combining means for combining a marker capable of identifying the area data at the head of at least one area data detected by the boundary detection means;
A code conversion apparatus comprising: A code conversion device that divides image data into rectangular areas and converts a code stream generated by encoding each rectangular area,
Identification means for identifying region data of a rectangular region from the code stream;
Setting means for setting an arrangement order of the area data identified by the identification means;
Reordering means for reordering the code data based on the alignment order set by the setting means;
Have
A code conversion apparatus that converts code data related to the rectangular area into a code stream that is combined and arranged.   5. The code conversion apparatus according to claim 3, wherein the area data is any one of frequency band data, color component data, bit plane aggregate data, and small area data. A code expansion device that divides image data into rectangular regions and expands a code stream to which an identifier that can identify a boundary of the data is attached,
Identification means for identifying region data of a rectangular region from the code stream;
Setting means for setting the arrangement order of the area data identified by the identification means;
Reordering means for reordering the code data based on the alignment order set by the setting means;
Decompression means for decompressing the code stream;
A code expansion device comprising:   The code expansion device according to claim 6, wherein the area data is any one of frequency band data, color component data, bit plane aggregate data, and small area data. An encoding method for encoding input data and generating a code stream,
A dividing step of dividing the input data into at least one rectangular region;
A predetermined conversion is performed on the rectangular area divided in the dividing step, and a conversion step for converting into the conversion data,
An ordering step of giving a priority order to the converted data;
An encoding step of aligning the converted data based on the priority order assigned in the ordering step and generating a code stream by performing an encoding process;
A marker combining step of combining a marker whose priority order can be identified at the head of each converted data in the code stream generated in the encoding step;
An encoding method characterized by comprising:   9. The encoding method according to claim 8, wherein the conversion data is any one of frequency band data, color component data, bit plane aggregate data, and small area data. A code conversion method for converting a code stream generated by dividing image data into rectangular areas and encoding each rectangular area,
A boundary detection step of reading at least one region data of the code stream and detecting a boundary portion of the region data;
A marker combining step for combining a marker capable of identifying the region data at the head of at least one region data detected in the boundary detection step;
A code conversion method characterized by comprising: A code conversion method for converting a code stream generated by dividing image data into rectangular areas and encoding each rectangular area,
An identification step of identifying rectangular area data from the code stream;
A setting step for setting the alignment order of the data identified in the identification step;
A rearrangement step of rearranging the code data based on the alignment order set in the setting step;
Have
A code conversion method comprising: converting code data relating to the rectangular area into a code stream arranged by being combined.   12. The code conversion method according to claim 10, wherein the area data is any one of frequency band data, color component data, bit plane aggregate data, and small area data. A code decompression method for decompressing a code stream to which image data is divided into rectangular regions and an identifier capable of identifying the boundary of the region data is added,
An identification step of identifying rectangular area data from the code stream;
A setting step for setting the alignment order of the data identified in the identification step;
A rearrangement step of rearranging the code data based on the alignment order set in the setting step;
A decompression step of decompressing the code stream;
A code expansion method characterized by comprising:   14. The code expansion method according to claim 13, wherein the area data is any one of frequency band data, color component data, bit plane aggregate data, and small area data.

Images