JPH0750828A - Picture coder - Google Patents

Abstract

PURPOSE:To encode other picture simultaneously during coding of a picture by adopting a bit plane coding method so as to apply reverse coding at a high coding efficiency even when original pictures such as a black/white picture, a color picture and a motion picture are in existence in mixture. CONSTITUTION:When a binary bit lane picture 10 is generated from a multi- value digital original picture 1, attribute information 9 is added respectively to each of binary bit plane pictures 10. Even when binary bit plane pictures 10 whose original picture differs are stored in bit plane memories 13, they are distinguished by referencing the attribute information 9. Thus, while refering the binary bit plane pictures 10 whose multi-value digital original picture 1 is in common, a coding section 15 makes prescribed coding.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding apparatus which compresses image information and codes it with high efficiency.

[0002]

2. Description of the Related Art When image information is handled by an information processing device, the amount of data is compressed in order to shorten the transfer time and save the storage location. FIG. 2 shows an explanatory diagram of a general image coding procedure. As shown in the figure, for example, when a multi-valued digital original image 1 is given, the compression processing is performed paying attention to the redundancy of each image signal forming this image, and the transfer 3 is performed via a communication line. , Store in memory 4 is performed. On the receiving side, the original multi-valued digital original image 1 is reproduced by a process reverse to this compression process 2. Further, even when the data stored in the memory is read out, the multivalued digital original image 1 is reproduced by the process reverse to the compression process 2.

A high-efficiency coding method known as a bit-plane coding method has been developed as a compression processing method as described above. In this method, a binary bit plane image is obtained from a multi-valued digital original image. FIG. 3 shows an explanatory diagram of a binary bit plane image generation operation. As shown in this figure, for example, each signal 6 that constitutes the multilevel digital original image 1
Is composed of 4 bits. In this case, only the most significant bits of each signal are collected to obtain the binary bit plane image 5-1. Further, only the second bit is collected to obtain the binary bit plane image 5-2. Similarly, binary bit plane images 5-3 and 4 in which only the third bit is collected
Binary bit-plane image that collects only the th bit 5-
Get 4.

The four binary bit plane images 5-1 to 5-4 thus obtained have strong correlation between pixels, and are compressed with high coding efficiency by various well-known compression methods. It becomes possible to do. Also, in the case of a color image, a technique for obtaining high compression efficiency by arranging MSB bits to LSB bits in a fixed order for the corresponding bit planes of red R, green G, and blue B is introduced. (Japanese Patent Laid-Open No. 4-154276
Issue).

[0005]

The various well-known high-efficiency coding methods conventionally have the following problems to be solved. Generally, in order to perform high-efficiency encoding, it is necessary that each image signal of the input original image is read by the encoding unit in a predetermined fixed order. However, this order differs depending on whether the original image is a simple binary image, a monochrome multi-valued image, a color image, or a moving image. That is, it is necessary to rearrange the most suitable image signals in advance depending on the nature of the original image.

Therefore, it is not easy to uniformly encode original images of various properties with one device. Moreover,
As described above, it is necessary to input image signals in an order suitable for encoding, and it has been difficult to accept a multiplexed signal in which two or more types of images are mixed and perform compression processing.

The present invention has been made by paying attention to the above points. Even when original images such as black-and-white images, color images, and moving images are mixed, the bit plane coding method as described above is used. It is an object of the present invention to provide an image coding apparatus capable of performing lossless coding with high coding efficiency and further coding another image at the same time while coding one image.

[0008]

An image coding apparatus according to the present invention comprises a bit plane image generation unit for generating a binary bit plane image obtained by collecting corresponding bits from a multilevel digital original image, and An attribute adding unit that adds attribute information indicating a relationship with a multi-valued digital original image to each binary bit plane image, a bit plane memory that stores a plurality of the binary bit plane images, and the binary bit plane image. With respect to any one of the binary bit plane images stored in the bit plane memory and the encoding condition setting unit that sets the encoding condition of, the corresponding binary bit plane image is selected by the attribute information. With reference to the above, an encoding unit for executing the encoding set by the encoding condition setting unit is provided.

The encoding unit also stores a bit plane memory in which a plurality of binary bit plane images each having a different multi-valued digital original image are mixed and stored, and one of the two bit plane memories stored in the bit plane memory. Regarding the value bit plane image, the encoding set by the encoding condition setting unit is executed while selecting and referring to another binary bit plane image having common multivalued digital original images by the attribute information.

Further, the other encoding unit, for any one of the binary bit plane images stored in the bit plane memory, outputs another arbitrary binary bit plane image having common multilevel digital original images, The encoding set by the encoding condition setting unit is executed while selecting and referring to the attribute information regardless of the storage order in the bit plane memory. In addition to this, a memory storage control unit may be provided for setting a priority order when deleting from the bit plane memory for each bit plane image in which the multivalued digital original images stored in the bit plane memory are common to each other. .

[0011]

According to the present invention, when the binary bit plane image is generated from the multi-valued digital original image, the attribute information is added to each binary bit plane image. Even if binary bit plane images having different original images are stored in the bit plane memory, they can be distinguished by referring to the attribute information. Therefore, while referring to the binary bit plane image in which the multi-valued digital original images are common to each other,
The encoding unit performs predetermined encoding. At the time of this encoding, any other binary bit plane image stored in the bit plane memory can be referred to, so that it is possible to refer to with high encoding efficiency regardless of the storage order in the memory.

Further, each binary bit plane image is distinguished by attribute information, and a monochrome image, a color image, and a moving image can be mixed in the bit plane memory. It is possible to arbitrarily select and execute the encoding processes concurrently. Further, when a new binary bit plane image is stored in the bit plane memory, the binary bit plane image to be erased is selected instead according to a certain priority order, which is useful for high encoding. Is left on the memory for a long time, and the encoding process proceeds.

[0013]

The present invention will be described in detail below with reference to the embodiments shown in the drawings. FIG. 1 is a block diagram showing an embodiment of an image coding apparatus of the present invention. In the figure, this device includes a bit plane image generation unit 11, an attribute addition unit 12, a bit plane memory 13, an encoding condition setting unit 14, and an encoding unit 15. Bit plane image generation unit 11
3 is a part for receiving the multi-valued digital original image 1 and generating the binary bit plane image 10.
This is the same as the conventionally performed processing described using.

The attribute adding unit 12 is newly provided in the apparatus of the present invention, and the relationship between the plurality of binary bit plane images 10 generated by the bit plane image generating unit 11 and the multilevel digital original image 1 is related. Is a part to which attribute information indicating is added. Thus, the attribute information 9 and the binary bit plane image 10 are stored in the bit plane memory 13 as a set. FIG. 4 shows an explanatory diagram of the contents of the attribute information as described above. In the present invention, binary bit plane images generated from different multi-valued digital original images can be mixed in the bit plane memory 13 shown in FIG.

The memory numbers shown in FIG. 4 indicate the oldest stored binary bit plane image as "n" and the most recently stored binary bit plane image as "0". That is, in this example, n binary bit plane images can be stored in the bit plane memory. Here, in the binary bit plane having the memory number “0”, as shown in FIG. 4, the image name is d1, the frame rate is “30”, the frame number is “2”, and the component is “3”. Attribute information such that the component number is "2", the number of bit planes is "8", and the bit plane number is "2" is added.

The image name d1 is a symbol for identifying a multilevel digital original image. Further, the frame rate indicates that, for example, a moving image is composed of 30 frames per second, and each of these frames becomes a multivalued digital original image. The frame number is 30
This is a part that represents which of the frames is the multi-valued digital original image. Further, in the case of a color image, there are three types of components, such as red R, green G, and blue B. The component number is a part showing the number of the component. The number of bit planes is
It represents how many bits each image signal is composed of. In this example, since the image signal is 8 bits, the number of bit planes is eight. Further, the bit plane number is a part indicating, for example, which bit plane the binary bit plane image stored in the memory number “0” corresponds to.

In the example shown in this figure, for example, in the area of the memory number "2", the frame rate is "0" so that it is a still image, and since the constituent elements are "1", there is a black and white image. It is stored. Also, in this example, although the image is a multi-valued image with 10 bit planes, the memory number is i-
Since the number of bit planes of the binary bit plane of the image name b5 stored in the area 1 is 1, the original image is a black and white binary image. As described above, in the present invention, since attribute information is added to each binary bit plane image, any image, whether a moving image or a still black and white image, is converted into a binary bit plane image and stored in the bit plane memory 13 in an arbitrary order. Can be stored. In this case, if there is a sufficient margin in the transmission speed and the coding speed, the coding can be performed without lowering the efficiency.

The binary bit plane image 10 stored in the bit plane memory 13 shown in FIG. 1 is then encoded by the encoding unit 15 according to the conditions set by the encoding condition setting unit 14, and the high efficiency encoded data 16 is obtained. To get FIG. 5 shows an explanatory diagram of the encoding procedure according to the present invention. First, at the start of encoding, of the binary bit plane image 10 stored in the bit plane memory, the memory number “0”
The binary bit plane image of is set as the encoding target image (step S1). Next, in step S2, the attribute information of the binary bit plane image is referred to,
Search for binary bit-plane images with the same image name. Then, the binary bit plane image selected as the reference target is taken out and encoded.

Upon completion of such encoding, the next new binary bit plane image is stored in the bit plane memory (step S3). At the same time, any unnecessary binary bit plane image is deleted from the memory (step S4). The binary bit plane image to be erased is stored in the area having the memory number "n". Normally, the binary bit plane image is generated in the order of the most significant bit MSB to the least significant bit LSB. By doing so, hierarchical display can be performed, and therefore it is widely adopted in general. However, in the case of the present invention, no problem occurs even if the binary bit plane image of another original image is interrupted during such processing.

That is, using the attribute information, it is possible to freely refer to binary bit plane images that are not continuous and perform encoding. Furthermore, if the transmission rate of the communication line for transferring the output of the encoding unit 15 shown in FIG.
Various types of images such as value images can be encoded and transferred by simultaneous parallel processing.

The coding process by the coding unit 15 shown in FIG. 1 is performed by a well-known method. For example, a pixel of interest is set, the surrounding pixels are referred to obtain the correlation between the pixels, and entropy coding is performed using the probability of occurrence of occurrence by the conditional symbol of the Markov model as a coding parameter. There are various selection methods for the reference pixel at this time. FIG. 6 illustrates a reference pixel model (part 1).
Indicates. For example, in this figure, consider a neighborhood 10 pixel reference model. In this figure, the pixel indicated by the number "0" is the pixel of interest. Then, the pixels to be referred to are determined in the numerical order shown in the figure.

For example, as shown in FIG. 2B, when the value of each reference pixel is "0" or "1", the reference image signal is "0111100110" in this example. The coding condition setting unit 14 shown in FIG. 1 estimates the symbol appearance prediction probability from such a reference image signal. The symbol appearance prediction probability is determined for each combination of bit strings of "0" and "1" of the reference image signal. For example, when 10 pixels are referred to, there are 1024 cases, and the symbol appearance prediction probability is calculated for each case. A coding parameter is generated based on this probability. Then, the encoding unit 15 sequentially reads each image signal of the binary bit plane to be encoded from the bit plane memory 13, performs binary entropy encoding using the encoding parameter, and outputs an encoded sequence. To do.

In the example shown in FIG. 6, the reference pixel is set only in the binary bit plane which is being encoded. However, multiple 2
When the value bit plane image is generated, the reference pixel is set to the already-encoded binary bit plane that has a strong correlation with the binary bit plane that is being encoded. As a result, the coding efficiency can be improved.

FIG. 7 is an explanatory view of the reference pixel model (No. 2).
Indicates. Also in the example of this figure, the pixel displayed with the number “0” is the pixel of interest. Then, the pixels having the numbers “1” to “10” are set as the reference pixels. In the example of (a), the pixels with numbers “1” to “4” belong to the same binary bit plane image, and the pixels with numbers 5 to 10 belong to another binary bit plane. In the example of this figure, three flags are prepared to represent the reference bit plane. The first flag means a reference frame. If it is "0", it is in the same frame, and if it is -1, it is 1.
It means to refer to the previous frame. Also,
The second flag means a reference constituent element, and if the constituent element types are red R, green G, and blue B, and the constituent element of the coded bit plane is G, the second flag is set to "0". G
Will refer to the bit plane. If it is -1, it refers to the bit plane of R, which is the immediately previous component.

The third flag means a reference bit plane. If the flag is -1 when the coded bit plane is the third bit plane, the second bit plane immediately before is referred to. For example, the example of FIG. 7A is a model that refers to a bit plane that is one level higher than the coded bit plane and has the same component in the same frame. (B) is
It is a model that refers to the same bit plane of the immediately preceding component of the same frame. (C) is a model that refers to the same bit plane of the same constituent element of the previous frame.

FIG. 8 shows another reference pixel model explanatory view (No. 3). In this example, reference bits when a color moving image is encoded are shown. That is, the 0th pixel is the pixel of interest, and the 1st to 16th pixels are referred to. Here, Nos. 1 and 2 are pixels of the same bit plane, and the other pixels are dispersed in each of the five binary bit plane images. The flag in this case is also shown in FIG.
It is a flag displayed by the same rule as described in. In this way, high efficiency coding is performed with reference to a plurality of various binary bit plane images.

The coding efficiency can be improved by referring to the already-coded binary bit plane stored in the binary bit plane memory. The decoding side is also provided with the same binary bit plane memory as that on the encoding side, and there is no need to send extra information to the decoding side to store the binary bit plane that has been decoded. The same information can be referenced.

When the binary bit plane to be referred to is not stored in the bit plane memory side, the following is performed. The reference bit plane specified by the reference pixel model may have already been erased due to the limited storage capacity of the bit plane memory and may not be stored. Also, since the number of RGB bit planes in a color image is a primary color expression, R is 2 bits, G is 3 bits, and B is 3 bits.
In the case of a bit configuration, the R 3 bit plane may not be encoded and may not be stored in the bit plane memory. In such a case, these binary bit plane images cannot be selected as reference bit planes. Therefore, such a reference bit plane is excluded from the reference pixels in advance for encoding. If the same process is performed on the decoding side, no inconvenience occurs between the encoding side and the decoding side.

However, if the number of binary bit planes that can be referred to is reduced, the number of reference pixels may be reduced, and the coding efficiency may be reduced. To deal with this, the erase order of the stored bit planes is prioritized. That is, for bit planes that are likely to be referenced, the storage rate in the bit plane memory is increased. For example, the higher the MSB plane is stored, the higher the storage rate is, and the lower LSB plane or the binary image is preferentially erased. By doing so, the higher the MSB plane is, the more general information about the image is stored, so there is a high possibility that it will be referenced, and it is possible to prevent a decrease in coding efficiency. For this purpose, the apparatus shown in FIG. 1 is provided with a storage control unit 17 for setting the priority order of the binary bit plane image to be deleted from the bit plane memory 13. For example, when the bit plane memory 13 is a first-in first-out memory, the storage control unit 17 changes the order of the signals output from the attribute adding unit 12 and arranges them in the order of priority to control the storage in the bit plane memory 13. To do.

If the binary bit plane images in the bit plane memory 13 can be erased in an arbitrary order, the storage control unit 17 sequentially outputs the binary bit plane images in accordance with a preset priority order. Erase. It is also possible to arbitrarily select and refer to other binary bit planes having similar characteristics to the reference bit plane stored in the bit plane memory in order to improve the coding efficiency. Therefore, it is possible to select the most efficient encoding method without being influenced by the input order of the binary bit plane image to the encoding unit as in the conventional case.

The present invention is not limited to the above embodiments. The apparatus of the present invention can accept an original image of any property in principle, form a binary bit plane image, encode these in arbitrary ranks, and output as highly efficient encoded data. Of course, as in the conventional art, it is also possible to generate a binary bit plane image by paying attention to one original multi-valued digital image, store it in the bit plane memory in that order, and perform the encoding process. However, in this case, as described above, the reference pixel can be extracted from an arbitrary binary bit plane image having the highest coding efficiency, and high coding efficiency can be obtained. Further, in order to read a target binary bit plane image from the bit plane memory and perform a coding process, for example, when processing a plurality of original images, a plurality of coding units may be provided.

[0032]

The image coding apparatus of the present invention described above generates a binary bit plane image obtained by collecting corresponding bits from a multilevel digital original image,
Since the attribute information indicating the relationship with the multivalued digital original image is added to each binary bitplane image, the binary bitplane image stored in the bitplane memory is encoded while being selected according to the attribute information. Can be performed and the encoding can be performed efficiently.

Further, since it is possible to identify by attribute information,
Free original images such as moving images, still images, and binary images can be mixed in the bit plane memory, and a single encoding device can simultaneously encode various original images. become. Therefore, for example, when moving image communication is performed, other still images and moving images can be transmitted without interrupting the transmission of the moving image, which is advantageous for multimedia communication and the like. Further, by setting a priority order for a plurality of binary bit plane images stored in the bit plane memory and deleting them from the bit plane memory in that order, binary bits that are likely to be referenced It is possible to leave the plane in the bit plane memory and perform encoding with high encoding efficiency.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of an image encoding device of the present invention.

FIG. 2 is an explanatory diagram of a general image encoding procedure.

FIG. 3 is an explanatory diagram of a binary bit plane image generation operation.

FIG. 4 is an explanatory diagram of attribute information contents.

FIG. 5 is an explanatory diagram of an encoding procedure according to the present invention.

FIG. 6 is a reference pixel model explanatory view (No. 1).

FIG. 7 is a reference pixel model explanatory diagram (2).

FIG. 8 is a reference pixel model explanatory view (No. 3).

[Explanation of symbols]

 1 multi-level digital original image 9 attribute information 10 binary bit plane image 11 bit plane image generation unit 12 attribute addition unit 13 bit plane memory 14 encoding condition setting unit 15 encoding unit 16 high efficiency encoded data 17 storage control unit

Claims (5)

[Claims] 1. A bit-plane image generation unit for generating a binary bit-plane image obtained by collecting corresponding bits from a multi-valued digital original image, and a multi-valued digital original image for each of the binary bit-plane images. And a bit plane memory for storing a plurality of the binary bit plane images, and an encoding condition setting unit for setting an encoding condition for the binary bit plane images. With respect to one of the binary bit plane images stored in the bit plane memory, the setting of the encoding condition setting unit is performed while selecting and referring to another corresponding binary bit plane image according to the attribute information. An image encoding apparatus comprising: an encoding unit that performs encoding. 2. A bit plane memory in which a plurality of binary bit plane images each having a different multi-valued digital original image are mixed and stored, and one of the binary bit plane images stored in the bit plane memory, An encoding unit for executing the encoding set by the encoding condition setting unit while selecting another binary bit plane image having common multi-valued digital original images by the attribute information and referring to the other binary bit plane images. The image coding apparatus according to claim 1, wherein 3. An order of storing, in any one of the binary bit-plane images stored in the bit-plane memory, another arbitrary binary bit-plane image having common multi-valued digital original images in the bit-plane memory. 2. The image coding apparatus according to claim 1, further comprising: a coding unit that executes coding set by the coding condition setting unit while selecting and referring to the attribute information. . 4. An encoding unit is provided which simultaneously and in parallel performs encoding set by an encoding condition setting unit for a plurality of binary bit plane images having different multi-valued digital original images. The image coding apparatus according to claim 2. 5. A memory storage control unit is provided for setting a priority when deleting from the bit plane memory, for each bit plane image in which multivalued digital original images stored in the bit plane memory are common to each other. The image coding device according to claim 1 or 3, characterized in that: