JP4039609B2 - Image coding apparatus and moving picture coding apparatus using the same - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image encoding device performed for the purpose of reducing the amount of information, and more particularly to an image encoding device capable of decoding while sequentially increasing image quality and a moving image encoding device using the same.
[0002]
[Prior art]
Generally, since the quality of a decoded image when an image is encoded depends on the encoding rate at the time of encoding, the quality cannot be manipulated on the decoding side. The same applies to not only still images but also moving images. However, if hierarchical encoding is performed at the time of encoding, decoding can be performed while changing the image quality on the decoding side. In the conventional hierarchical coding technology represented by the MPEG-2 SNR scalability coding technology, the quality of a moving image that can be manipulated at the time of decoding can only be changed in about two stages, and the performance of the decoding device and the bandwidth of the transmission path can be changed. It has been difficult to provide a decoded moving image having an optimum quality corresponding to a small difference in width. That is, in order to reproduce a high-quality image, the image quality can be improved only after all the encoded data of the enhancement layer can be decoded. The decoding result does not contribute to the quality of the decoded image at all, and only an image with the same quality as when only the base layer is decoded can be obtained.
[0003]
The bit plane coding technique is cited as a technique for solving such a problem that decoding up to the middle stage of hierarchical coding does not make sense and reflecting all the results of decoding up to the middle stage in improving image quality. be able to. Even when all the encoded data cannot be decoded by the bit plane encoding technique, an image having a quality corresponding to the amount of information decoded halfway can be obtained. Specifically, in MPEG-4 FGS coding technology, which is a moving picture coding method using bit-plane coding technology, the basic layer is created by normal MPEG-4 coding, and the extension layer is bit-plane coding. It is configured by using
[0004]
[Problems to be solved by the invention]
However, a conventional bit-plane encoding technique such as that used in MPEG-4 FGS encoding technique uses a one-pass binary zero-run-length encoding method. Since “0” symbols appear continuously, the coding efficiency is good, but as the lower plane is reached, there is a problem that the symbol appearance probability becomes random and the coding efficiency extremely decreases. Also, since encoding is not performed in consideration of the order of improving the decoded image quality at the time of encoding, characteristics of decoded image quality in the case of decoding a part of encoded data instead of all encoded data, that is, an intermediate stage There was also a problem that was not good.
[0005]
The present invention has been made to solve the problems of the prior art described above, and its object is to provide an image coding apparatus capable of enhancing a decoded image quality characteristic of the decoding intermediate stage. Still another object is to provide a moving picture coding apparatus using the picture coding apparatus.
[0006]
[Means for Solving the Problems]
In order to achieve the above-described object, the present invention provides means for expressing an image signal coefficient in a bit plane, and a bit that becomes a symbol “1” only after scanning each bit expressed in the bit plane in order from the upper bit. A significant bit group consisting of bits belonging to a higher rank than this bit; means for classifying into a significant bit group consisting of bits belonging to a lower rank than the significant bit group; means for encoding the significant bit group; Means for encoding a bit group , wherein the means for encoding the significant bit group preferentially encodes the significant bit group, and the means for encoding the significant bit group includes After encoding of significant bit groups is completed for all bit planes, encoding of significant bit groups is started. To have a first feature in that as will encode.
According to this feature, it is possible to improve the image quality characteristic in the middle of decoding. In addition, the encoding efficiency of significant bit groups can be improved, and the amount of encoded information can be reduced.
[0009]
The second feature is that when the image signal to be encoded is a differential signal used in the extended hierarchy of hierarchical encoding, entropy encoding such as Huffman encoding is performed on the already significant bits. There is. According to this feature, the encoding efficiency of the already significant bit group can be improved, and the amount of encoded information generated can be reduced.
[0010]
Furthermore, there is a third feature in that the image coding apparatus according to the present invention is applied not only to still images but also to coding of an extended hierarchy of a hierarchical moving picture coding apparatus. According to this feature, it is possible to reduce the amount of encoded information of the generated moving image encoded data and to perform decoding while changing the quality of the moving image at the time of decoding.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an image encoding apparatus according to the present invention. In the input image signal a, the sign bit b is first separated by the sign bit separation unit 1 and the absolute value c of the image signal is output. The separated sign bit b is held in the sign bit holding unit 5. The image signal expressed by the absolute value c is expressed as a bit plane by the pre-encoding processing unit 2, and when the input signal is DCT, a zigzag scan or the like is further performed and the binary bit sequence is rearranged in order. d is output.
[0012]
The binary bit series d is classified into two types of bit groups, a significant bit group e and a significant bit group f, by the significant / significant bit separation unit 3. The significant bit group e is encoded by the significant bit encoding unit 4, while the significant bit group f is rearranged by the significant bit rearranging unit 6 as described later with reference to FIG. It is encoded by the encoding unit 7. Here, when the significant bit e is encoded, the output request signal g is output from the significant bit encoding unit 4, and the code bit is output from the code bit holding unit 5 simultaneously with the encoded signal. The encoded significant bit i, the encoded already significant bit j, and the encoded bit h are combined by the combining processing unit 8 and output as encoded data k.
[0013]
Hereinafter, each process described above will be described in detail. For convenience of explanation, the case where the input image signal uses DCT coefficients subjected to orthogonal transformation as the input image signal will be described. However, the apparatus of the present invention can be applied to an image signal which has not been transformed, regardless of the type. Is possible.
[0014]
First, as shown in FIG. 2, since the DCT coefficients (for example, 11, -4, 23, -9,...) That are the input image signal a have positive and negative values, the sign bit separation processing unit 1 Separates a sign bit (Sign) b representing positive or negative from the DCT coefficient. As an example, the sign bit b is represented by 0 for positive and 1 for negative, and is stored in the sign bit holding unit 5. At this time, the sign bit separation unit 1 takes the absolute value of the DCT coefficient and outputs a signal (| DCT |) c representing the magnitude of the DCT coefficient.
[0015]
Next, the pre-encoding processing unit 2 performs bit plane representation of the absolute value of the DCT coefficient. In the bit plane representation, it is necessary to determine the number of bit planes. This is obtained by finding the one having the largest DCT coefficient size, and adjusting it to the number of bits necessary to represent the maximum DCT coefficient. In the illustrated example, since the maximum DCT coefficient is 23, the number of bit planes is 5. Here, the bit plane means a bit string in the horizontal direction composed of 0 and 1 corresponding to each of [0] to [4] in FIG.
[0016]
Further, the encoding preprocessing unit 2 rearranges the sequences by zigzag scanning, and sequentially outputs binary bit sequences in units of planes. FIG. 3 shows an encoding order of two sequences that are sequentially output in units of planes. The subsequent encoding is performed sequentially from the top of the upper plane.
[0017]
Next, as shown in FIG. 4, the significant / significant bit separation unit 3 converts each bit into two types of bit groups, that is, a significant bit group and a significant bit group based on the distribution of the significant coefficient. Classify into: Here, the significant bit is a bit that becomes the symbol “1” only when the DCT coefficient is expressed in bits and scanned in order from the upper bit. That is, all the bits belonging to the significant bits of the DCT coefficient are “0” symbols. Significant bits and bits that are higher than significant bits are collectively referred to as a significant bit group. On the contrary, the bits belonging to the lower order than the significant bits are collectively referred to as the already significant bit group.
[0018]
In FIG. 4, the bits marked with ○ are significant bits, the bits belonging to the range surrounded by the dotted frame are the significant bit group, and the bits belonging to the other ranges are the already significant bit group.
[0019]
Next, the encoding process of the significant bit encoding unit 4 is shown in FIG. The significant bit encoding unit 4 performs binary zero-run length encoding on the significant bit group. Since all the bits located above the significant bits are “0” symbols, the run of “0” symbols continues for a long time. Also, the run of the “0” symbol continues for a long time from the result of excluding the significant bits whose symbol appearance probability is random. The long run of the “0” symbol is shown in the upper column of FIG. As a result, efficient encoding can be performed by binary zero-run length.
[0020]
Next, as shown in the figure, binary zero-run length coding first creates (RUN, EOP) symbols, and then creates (RUN, EOP) symbols with variable-length coding. To do. In this variable length coding, RUN represents the number of consecutive “0” symbols, EOP represents the end of the plane if 1 and 0 indicates that the plane has not ended.
[0021]
The (RUN, EOP) symbols of this significant bit group are as shown in the lower column of the figure. In addition, in this variable length coding, (RUN, EOP) symbols are created with significant bits of the significant bit group as delimiters, so at the same time one (RUN, EOP) symbol is created, It is necessary to output h.
[0022]
Next, operations of the significant bit rearrangement unit 6 and the significant bit encoding unit 7 that perform rearrangement and encoding processing of the remaining significant bit groups will be described with reference to FIG. FIG. 6 shows an example of encoding processing of a significant bit group. The significant bit group can also be performed in units of planes as shown in FIG. 3, but here, in consideration of image characteristics, as shown by the dotted arrow in FIG. Change to the left. That is, as shown in FIG. 6, starting from 0 of [2], 0, 0, 0, 1, 0, 0, 0, 0 is defined as one plane. Next, starting from 1 of [3], 0,1,0,1,1,1 is the second plane, then starting from 1 of [4], 1,1,0 is the third The plane and the remaining 1 are the fourth plane. By this order, encoding is performed preferentially from information (that is, higher-order bits) that contributes to image quality at the time of decoding, and image quality characteristics in the middle of decoding can be improved. FIG. 7 shows the result of rearranging the significant bit groups. Entropy coding such as Huffman coding is performed on the already significant bit group after the rearrangement.
[0023]
The significant bit group and the already significant bit group encoded by the above procedure are combined by the combining processing unit 8 to generate encoded data k having a structure as shown in FIG. That is, the first half part is composed of the encoded data of the significant bit group, and the encoded data of the already significant bit group is arranged in the second half part. However, code bit information is also arranged at an appropriate location in the encoded data of the significant bit group.
[0024]
Next, another embodiment of the present invention will be described. In this embodiment, the coding apparatus of the present invention is applied to moving picture coding, and the apparatus configuration is shown in FIG. FIG. 9 shows a hierarchical video encoding apparatus that performs motion compensation prediction such as MPEG. As is clear from the configuration of the figure, the hierarchical video encoding apparatus is composed of a basic hierarchy and an extension hierarchy. The basic layer is created by an encoding device similar to normal MPEG that does not perform layering. On the other hand, the encoding apparatus 11 of the present invention is applied to the enhancement layer. Since the MPEG layered moving image encoding apparatus is known, detailed operation description is omitted.
[0025]
The enhancement layer is configured by encoding a difference image between the original image and the decoded image of the base layer. When the encoding apparatus of the present invention is applied to the enhancement layer encoding, decoding can be performed at a variable bit rate during decoding. In addition, by encoding the difference data in the enhancement layer, the DCT coefficients follow a Laplace distribution, and the appearance probability of symbols of the significant bit group is biased. Therefore, entropy such as Huffman code is applied to the significant bit group. Improvement of the encoding efficiency by using encoding can be obtained.
[0026]
As is apparent from the above description, according to the invention of claim 1, it is possible to improve the image quality characteristics in the middle of the decoding.
[0028]
According to the second aspect of the invention, it is possible to improve the encoding efficiency of the significant bit group and reduce the amount of encoded information. According to the invention of claim 3 , the encoding efficiency of the already significant bit group can be improved, and the amount of encoded information generated can be reduced.
[0029]
According to the invention of claim 4 , it is possible to reduce the amount of encoded information of the generated encoded moving image data, and at the same time, it is possible to provide a moving image having a quality according to the decoded data amount.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of an embodiment of the present invention.
FIG. 2 is an explanatory diagram of a bit plane of an image signal.
FIG. 3 is an explanatory diagram of an encoding order.
FIG. 4 is an explanatory diagram showing classification of a significant bit group and an already significant bit group.
FIG. 5 is an explanatory diagram of encoding significant bit groups by binary zero-run length.
FIG. 6 is an explanatory diagram of an encoding order of a significant bit group.
FIG. 7 is an explanatory diagram of an encoding order of significant bit groups in consideration of the degree of contribution to image quality.
FIG. 8 is an explanatory diagram of the structure of synthesized encoded data.
FIG. 9 is a block diagram showing a configuration of a second exemplary embodiment of the present invention.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Code bit separation part, 2 ... Pre-encoding process part, 3 ... Significant / pre-significant bit separation part, 4 ... Significant bit encoding part, 5 ... Sign bit holding part, 6 ... Significant bit rearrangement unit, 7 ... Significant bit coding unit, 8 ... Combination processing unit, 11 ... Coding apparatus of the present invention.

Claims (4)

Means for representing the image signal coefficients in a bit plane;
A significant bit group consisting of a bit that becomes a symbol “1” for the first time when each bit expressed in the bit plane is scanned in order from the upper bit and a bit that belongs to a higher rank than this bit, and a bit that belongs to the lower order of the significant bit group Means for classifying into significant bit groups consisting of:
Means for encoding the significant bits;
Means for encoding the significant bit group ,
The means for encoding the significant bit group is performed by first giving priority to the encoding of the significant bit group,
The means for encoding the significant bit group starts encoding the significant bit group after the encoding of the significant bit group is completed for all the bit planes, and the encoding order of the significant bit group An image encoding device characterized in that encoding is performed preferentially from the upper bits . The image encoding device according to claim 1,
An image encoding apparatus characterized in that the means for encoding the significant bit group applies variable length encoding. The image encoding device according to claim 1,
The means for encoding the significant bit group applies entropy coding to the significant bit group when the image signal to be encoded is a differential signal used in an extended layer of hierarchical coding. An image encoding device. The enhancement layer of the hierarchical video encoding apparatus including a base layer and enhancement layer, the video encoding apparatus characterized by using the image coding apparatus according to any one of claims 1 to 3.

Images