JP2013165506A - Image encoder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image encoder for avoiding an increase of an arithmetic processing amount and enlargement of an LSI circuit, which are caused by decoding of parameter information referred to in arithmetic decoding processing.SOLUTION: An image encoder includes: an acquisition section for acquiring image data and an image parameter; an image encoding section which image-encodes the image data by referring to the image parameter, generates a residual coefficient, binarizes and entropy-encodes the generated residual coefficient to generate encoded image data; a parameter processing section which (1) outputs a parameter referred to when the encoded image data is entropy-decoded among parameters included in the image parameters as first parameter information without encoding the parameter, and (2) encodes and outputs the image parameter as second parameter information; and a stream generation section for generating a stream including the encoded image data, the first parameter information and the second parameter information and outputs the stream to an external device.

Description

  The present invention relates to an image encoding apparatus that encodes an image, an image decoding apparatus that decodes encoded image data, and a method thereof. The present invention relates to a parameter information control method referred to when decoding encoded image data arithmetically encoded by H.264 / AVC.

  ITU-T and ISO / IEC JTC1 formed JVT (Joint Video Team) as a new video compression and coding standard, and “H.264 / AVC” was approved as a recommendation in May 2003. . This H. H.264 / AVC is said to achieve about twice the compression rate compared to MPEG-2 used in Japanese satellite digital broadcasting and terrestrial digital broadcasting.

  In order to achieve a high compression ratio, H. H.264 / AVC implements many video compression techniques, and encoding is performed by adaptively combining various compression techniques.

  H. One compression technique implemented in H.264 / AVC is entropy coding (variable length coding). In the entropy encoding method, two encoding methods of “CAVLC” and “CABAC” are prepared, and either one is selected by a predetermined flag.

  Here, “CAVLC” is an abbreviation of Context-Adaptive Variable Length Coding (context adaptive variable length coding method), and when a DCT coefficient is encoded, a run having a continuous length of “0” In this method, the level is encoded from the direction opposite to the scan direction by using a variable length encoding table. “CABAC” is an abbreviation for Context-based Adaptive Binary Arithmetic Coding (context adaptive binary arithmetic coding system), which is a system that performs coding by changing the appearance frequency of a coding object that changes with time. is there.

  CAVLC and CABAC are selected based on “entropy_coding_mode_flag” encoded in “pic_parameter_set_rbsp ()” of the header information. When this “entropy_coding_mode_flag” is “0”, CAVLC is selected, and when it is “1”, CABAC is selected.

  In CABAC, a context index (hereinafter referred to as “ctxIdx”) is attached to each code to be compressed, and the appearance frequency is changed and managed for each ctxIdx, thereby improving the compression rate by arithmetic coding. . Here, the context index is an identification number assigned to each classification by classifying binary data in arithmetic coding by syntax elements, adjacent macroblock information, and the like.

  The encoding process by CABAC is mainly divided into two processes. The first process is a process called binarization that converts multi-value information to be encoded, called a syntax element, into binary data. The second process is a process of calculating ctxIdx for binary data obtained by binarization and performing arithmetic coding.

  The process of decoding image data encoded by CABAC (hereinafter referred to as “encoded image data”) is divided into two processes as in the case of the encoding described above. This is a process of performing arithmetic decoding on the encoded image data and outputting binary data, and a process of performing multi-value conversion that converts the binary data into syntax elements.

  In CABAC, a compression rate is improved by providing parameters for each picture and each slice. Specifically, there are parameters such as “pic_width_in_MBs”, “pic_height_in_MBs”, “Slice_QPy”, “cabac_init_idc”, and “slice_type”. When encoding with CABAC, initialization of random variables and the like is performed with these parameters, and encoding is performed with a set angle of view.

  The above parameters are encoded by “seq_parameter_set_rbsp ()”, “pic_parameter_set_rbsp ()”, “slice_header ()”, etc., and are placed before “VCL Nal unit” in which encoded image data is stored.

  H. When decoding a H.264 / AVC-encoded stream, “seq_parameter_set_rbsp ()”, “pic_parameter_set_rbsp ()”, “slice_header ()”, etc. are decoded and the above “entropy_coding_mode_flagV” is extracted. To determine whether to decode with CABAC or CABAC. When decoding by CABAC, parameters used for CABAC decoding are extracted, random variables and the like are initialized using the extracted parameters, and arithmetic decoding is executed.

  By the way, as described above, the encoding and decoding processes in CABAC are divided into two. It is processing of binarization and multi-value processing of syntax elements, and processing of arithmetic coding and arithmetic decoding. Due to the nature of the arithmetic coding process, only one bit can be encoded in one cycle, and the arithmetic decoding process can only perform one bit decoding in one cycle. For this reason, it is assumed that the binarization process and the arithmetic coding process, and the multi-value process and the arithmetic decoding process are performed asynchronously.

  As a conventional technique, in an image decoding method for decoding input encoded image data, each variable-length codeword in the encoded image data is variable-length decoded into a corresponding intermediate code, and the decoded intermediate data Store code temporarily in buffer. A method has been proposed in which variable length decoding and processing other than variable length decoding are performed asynchronously by reading the intermediate code stored in the buffer, analyzing according to the syntax of image encoding, and decoding the image. (For example, refer to Patent Document 1).

  When the method of Patent Document 1 is applied to decoding by CABAC, an intermediate code is obtained by arithmetically decoding input encoded image data, and this intermediate code is temporarily stored in a buffer. Thereafter, the intermediate code stored in the buffer is read, and the intermediate code is multivalued to decode the image.

  FIG. 6 is a diagram illustrating an example of the structure of conventional encoded image data. As shown in FIG. 6, the conventional encoded image data stream is configured with “NAL unit” as a unit, and parameter information necessary for arithmetic decoding is extracted from a slice header or the like.

  FIG. 7 is a flowchart showing the flow of a conventional decoding process.

  First, SPS and PPS in the encoded image data are decoded (S200), and “entropy_coding_mode_flag” is extracted (S202). When the “entropy_coding_mode_flag” is “1” (S204: NO), parameter information is extracted from the SPS and PPS (S206), and the slice header is decoded (S208). When “entropy_coding_mode_flag” is “0” (S204: YES), the arithmetic decoding process (S212) is not executed (is skipped).

  Thereafter, an arithmetic decoding process is executed with reference to the extracted parameter information (S210) (S212), and when there is no other encoded image data to be decoded, this process ends (S214: YES). .

  In the image decoding apparatus to which the above-mentioned Patent Document 1 is applied, “entropy_coding_mode_flag” is used to determine whether CAVLC or CABAC is used for entropy encoding of input encoded image data. When it is decoded and the value is “1”, it is understood that it is CABAC. Next, in order to perform CABAC arithmetic decoding, “seq_parameter_set_rbsp ()”, “pic_parameter_set_rbsp ()”, “slice_header ()”, etc. are decoded and parameters are extracted. Then, using the extracted parameters, initialization of random variables and the like is performed, and arithmetic decoding processing is performed.

  Here, the intermediate code is composed of decoded image data and “seq_parameter_set_rbsp ()”, “pic_parameter_set_rbsp ()”, “slice_header ()”, etc. In order to reduce the buffer capacity, the intermediate code is decoded. Except for the image data that has been processed, data that has been encoded is stored.

JP 2003-259370 A

  However, in the above-described prior art, when the intermediate code stored in the buffer is read and the image is restored, the above “seq_parameter_set_rbsp ()”, “pic_parameter_set_rbsp ()”, “slice_header ()”, etc. are decoded again. Must be performed (S200, S208), there is a problem that the amount of calculation processing increases and the cost increases with the addition of LSI circuits.

  The present invention has been made in order to solve the above-described problem. An image encoding device and the like that can avoid an increase in the amount of arithmetic processing accompanying the decoding of parameter information referred to in the arithmetic decoding process and an increase in the number of LSI circuits. The purpose is to provide.

  In order to achieve the above object, an image encoding device according to the present invention is an image encoding device that encodes image data, an acquisition unit that acquires image data and image parameters, and the image parameters. An image encoding unit that encodes the image data to generate a residual coefficient, and further generates encoded image data by binarizing and entropy encoding the generated residual coefficient; (1 ) Among the parameters included in the image parameter, a parameter referred to during entropy decoding of the encoded image data is output as first parameter information without encoding, and (2) the image parameter is encoded. A parameter processing unit that outputs the second parameter information, the encoded image data, the first parameter information, and the second parameter information It generates a stream including, and a stream generation unit for outputting to an external device.

  The parameter processing unit generates the second parameter by encoding all the parameters included in the image parameter, and is referred to during the entropy decoding among the parameters included in the image parameter. As for the parameters, both the encoded parameter and the unencoded parameter may be output as the first parameter information and the second parameter information.

  An image encoding device for encoding an image, wherein the image encoding device arithmetically encodes image data representing an image to generate arithmetically encoded image data and performs arithmetic decoding An arithmetic encoding unit that generates encoded image data based on the parameter information referred to and the arithmetically encoded image data, the arithmetic encoding unit, when generating the encoded image data, Parameter information may be integrated into the encoded image data.

  The parameter information is H.264. It may be a parameter referred to when arithmetic coding is performed by CABAC of H.264 / AVC.

  The parameter information includes H.264. It is also possible to include entropy_coding_mode_flag in CABAC of H.264 / AVC.

  The parameter information includes H.264. In nal_unit including slice_data () of H.264 / AVC CABAC, information representing the data amount from the beginning of nal_unit to immediately before slice_data () may be included.

  The parameter information may be arranged before a variable-length codeword to be arithmetically decoded in a stream including the encoded image data.

  The parameter information includes H.264. In H.264 / AVC CABAC, it may be stored in user_data_unsigned_SEI.

  In addition, the arithmetic coding unit may incorporate the parameter information into the coded image data without coding.

  Furthermore, in order to achieve the above object, the present invention provides an image decoding apparatus for decoding an encoded image, wherein the encoded image includes predetermined parameter information and the arithmetically encoded image data. The parameter information is extracted from data, and an arithmetic decoding unit that arithmetically decodes the arithmetically encoded image data with reference to the parameter information is provided, and the arithmetic decoding unit includes the encoded image data The parameter information integrated together in the above may be extracted.

  Note that the present invention is realized as an image encoding method or an image decoding method having characteristic constituent means in the image encoding device or the image decoding device as steps, or for causing a computer to execute these steps. It can also be realized as a program. Needless to say, the program can be widely distributed via a recording medium such as a DVD or a transmission medium such as the Internet.

  According to the present invention, when the encoded image data is converted from the encoded image data to the intermediate code in the image decoding apparatus, the decoding process of parameter information other than the image data is not necessary, and thus the processing amount of the entire apparatus is reduced Thus, the processing time can be shortened and the circuit configuration can be simplified.

It is a block diagram which shows the outline | summary of the function structure of the image decoding apparatus which concerns on embodiment. It is a figure which shows an example of the stream structure of the coding image data which concerns on this invention. It is a structural example of the parameter information referred with the image decoding apparatus of embodiment. It is a figure which shows the list of the value of NAL unit type. It is a flowchart which shows the flow of the decoding process in the image decoding apparatus which concerns on embodiment. It is a figure which shows an example of the stream structure of the conventional encoding image data. It is a flowchart which shows the flow of the conventional decoding process.

  Hereinafter, embodiments according to the present invention will be described with reference to the drawings. In addition, although this invention is demonstrated using the following embodiment and attached drawing, this is for the purpose of illustration and this invention is not intended to be limited to these.

  FIG. 1 is a block diagram showing an outline of a functional configuration of the image decoding apparatus according to the embodiment. An image decoding apparatus 10 shown in FIG. 1 is an apparatus that does not require decoding of parameter information used when arithmetically decoding encoded image data, and includes an arithmetic decoding unit 12, a buffer 14, and syntax elements. A generation unit 15 and an image decoding unit 16 are provided.

  The arithmetic decoding unit 12 variable-decodes each variable-length codeword arithmetically encoded in the input encoded image data into a corresponding intermediate code, and temporarily stores the decoded intermediate code in a buffer To do. More specifically, the arithmetic decoding unit 12 extracts parameter information from encoded image data including unencoded parameter information and arithmetically encoded image data, and refers to the parameter information. The arithmetically encoded image data is arithmetically decoded.

  The buffer 14 is a RAM, for example, and stores the intermediate code generated by the arithmetic decoding unit 12.

  The syntax element generation unit 15 reads the intermediate code 13 stored in the buffer 14 and generates a syntax element and its value.

  The image decoding unit 16 receives the syntax and the value as input, restores the image according to the conventional method, and outputs the image data 17.

  FIG. 2 is a diagram showing an example of a stream structure of encoded image data according to the present invention.

  In general, H.W. The H.264 / AVC data stream is composed of a plurality of “NAL units” (not shown in FIG. 2). The type of “NAL unit” is determined by a parameter encoded at the head of “NAL unit”, that is, “nal_unit_type”. When “nal_unit_type = 9”, “NAL unit” is “access unit delimiter”. A list of “NAL unit types” is shown in FIG.

  In the encoded image data 11 shown in FIG. 2, the head is “access unit delimiter”, and “SEI” is placed immediately after that.

  “SEI (Supplemental Enhancement Information)” is a NAL unit that stores supplementary additional information used when decoding VCL. When the payloadType in the SEI is 5, the SEI is “user_data_unsigned_SEI”. “User_data_unsigned_SEI” can store information uniquely defined by the user. The parameter information referred in the arithmetic decoding is stored in this “user_data_unsigned_SEI”.

  After SEI, SPS and PPS follow, followed by “slice_layer_without_partitioning” which is a VCL NAL unit. The “VCL NAL unit” mainly includes “slice_header ()” and “slice_data ()”, and encoded image data is stored in “slice_data ()”.

  “Access unit delimiter” follows again after “VCL NAL unit”, and a new access unit is started. “SEI” is again placed after “access unit delimiter” (similarly, parameter information is stored in SEI). After SEI, “SPS” and “PPS” may or may not be present. Then, “VCL NAL unit” follows.

  FIG. 3 is a configuration example of parameter information referred to by the image decoding apparatus according to the present embodiment. These parameter information are all present in “SPS”, “PPS”, or “slice_header”.

“MbuffFrameFlag” in FIG.
MbaffFrameFlag = (mb_adaptive_frame_field_flag &&! Field_pic_flag) (1)
The parameter information obtained by Further, “mb_adaptive_frame_field_flag” is a syntax element in “SPS”, and “field_pic_flag” is a syntax element in “slice_header ()”.

  Furthermore, “entropy_coding_mode_flag” is a syntax element in the PPS, and determines a variable length coding mode. When “entropy_coding_mode_flag = 0”, “Exp-Golomb” and “CAVLC” are selected as variable length coding modes, and when “entropy_coding_mode_flag = 1”, “CABAC” is selected. The arithmetic decoding process is performed only when CABAC is selected.

  The other syntax elements are used when performing the “slice_data ()” sequence process and the CABAC arithmetic decoding process.

  Next, the operation of the image decoding apparatus 10 configured as described above will be described.

  FIG. 5 is a diagram illustrating a flowchart of a decoding process in the arithmetic decoding according to the present embodiment.

  Decoding processing in the image decoding apparatus 10 starts with extraction of parameter information (S100). Since the parameter information is stored inside the SEI, the encoded image data 11 is searched to identify “NAL unit” (= SEI) where “nal_unit_type = 6”.

  Parameter information is acquired from the retrieved SEI (S102), and "entropy_coding_mode_flag" is referred to from the parameter information (S104). When “entropy_coding_mode_flag = 0” (S104: YES), since the variable-length coding mode is not CABAC, the process proceeds to the determination of whether or not the encoded image data has been completed without performing the arithmetic decoding process.

  When “entropy_coding_mode_flag = 1” (S104: NO), “VCL_NAL_unit” is searched (S106). When the head of “VCL NAL unit” is searched, the encoded image data 11 is skipped until the slice header ends based on the data amount of the slice header in the parameter information (S108), and “slice_data ()” Advance the data to the beginning.

  Then, the variable length codeword in “slice_data ()” is arithmetically decoded using the parameter information (S212).

  When the arithmetic decoding process is completed, it is confirmed whether or not the encoded image data 11 continues thereafter. If the encoded image data 11 continues, the process returns to the beginning to continue the process. If there is no encoded image data, the process ends (S100 to S110).

  As described above, the image decoding apparatus according to the present embodiment can perform arithmetic decoding of encoded image data without decoding parameter information.

(Modification)
The present invention can constitute an image encoding device corresponding to the image decoding device 10. Specifically, the arithmetic encoding unit of the image encoding device arithmetically encodes image data representing an image to generate arithmetically encoded image data, and parameter information to be referred to when performing arithmetic decoding Encoded image data is generated based on the arithmetically encoded image data. In the arithmetic coding unit, the parameter information is collectively incorporated into the coded image data when the coded image data is generated.

  The above embodiment or modification has been described on the assumption that the parameter information exists in the SEI. However, when the parameter information does not exist in the SEI, it is shown in the flowchart of FIG. It is also possible to configure the image decoding device so that the decoding process in the arithmetic decoding unit can be performed according to the procedure. With this configuration, even when encoded image data other than the encoded image data according to the present invention is input, the encoded image data can be decoded by the same image decoding apparatus.

  The present invention can be used for an image encoding device that encodes an image by CABAC and an image decoding device that decodes the encoded image data. More specifically, H.M. H.264 / AVC image encoding apparatus for encoding high-definition video encoded by a data compression technique, and image decoding for decoding (reproducing) encoded image data generated thereby It can be used for the device.

DESCRIPTION OF SYMBOLS 10 Image decoding apparatus 11 Encoded image data 12 Arithmetic decoding part 13 Intermediate code 14 Buffer 15 Syntax element generation part 16 Image decoding part 17 Image data

Claims (3)

An image encoding device for encoding image data,
An acquisition unit for acquiring image data and image parameters;
An image code for generating encoded image data by encoding the image data with reference to the image parameter to generate a residual coefficient, and further binarizing and entropy encoding the generated residual coefficient And
(1) Out of the parameters included in the image parameter, a parameter referred to in entropy decoding of the encoded image data is output as first parameter information without being encoded, and (2) the image parameter is A parameter processing unit that encodes and outputs the second parameter information;
An image encoding device comprising: a stream generation unit that generates a stream including the encoded image data, the first parameter information, and the second parameter information, and outputs the stream to an external device. The parameter processing section is
Generating the second parameter by encoding all the parameters included in the image parameter;
Of the parameters included in the image parameter, for parameters referred to in the entropy decoding, both the encoded parameter and the uncoded parameter are the first parameter information and the parameter Output as second parameter information,
The image encoding device according to claim 1. An image encoding method for encoding image data, comprising:
Obtain image data and image parameters,
Image encoding the image data with reference to the image parameters to generate a residual coefficient;
Encoded image data is generated by binarizing and entropy encoding the generated residual coefficient,
(1) Out of parameters included in the image parameter, a parameter referred to in entropy decoding of the encoded image data is output as first parameter information without being encoded, and (2) the image parameter is Encode and output as second parameter information,
Generate a stream including the encoded image data, the first parameter information and the second parameter information, and output to an external device,
Image coding method.

Images