US-12627817-B2 - Device and method of video encoding with first and second encoding code

US12627817B2US 12627817 B2US12627817 B2US 12627817B2US-12627817-B2

Abstract

A video image encoding device, in a first mode, variable-length-encodes a residual coefficient to generate a coefficient code string, outputs the coefficient code string and the header information in a state in which the header information is associated with the coefficient code string, in a second mode, directly uses a differential image as a coefficient code string without variable-length-encoding the differential image, and outputs the coefficient code string and the header information in a state in which the header information is associated with the coefficient code string.

Inventors

Hideyuki Ohgose
Kiyofumi Abe
Hiroshi Arakawa
Tatsuro Juri
Kazuhito Kimura

Assignees

PANASONIC INTELLECTUAL PROPERTY CORPORATION OF AMERICA

Dates

Publication Date: 20260512
Application Date: 20240730
Priority Date: 20110309

Claims (5)

1 . An image encoding device that encodes an image, the image encoding device comprising: a processor; and a non-transitory memory having stored thereon executable instructions, which cause the processor to: generate a predictive image on a predictive unit for a current coding unit using predictive information, the prediction unit being included in the coding unit; and wherein in a first mode, the processor variable-length-codes a residual coefficient to generate a coefficient code string, the residual coefficient being generated by performing an orthogonal transformation process and a quantizing process on a differential image between the image to be encoded and the generated predictive image, wherein in a second mode, the processor directly uses the differential image to generate the coefficient code string without performing the orthogonal transformation process and the quantizing process, and wherein a syntax of the current coding unit includes flag information indicating whether the first mode or the second mode is applied for the current coding unit and the syntax is included in a syntax of the coding unit separate from a syntax of the prediction unit included in the syntax of the coding unit, and wherein in a second mode, the orthogonal transformation process and the quantizing process are bypassed.
2 . An image encoding method that encodes an image, the image encoding method comprising: generating a predictive image on a predictive unit for a current coding unit using predictive information, the prediction unit being included in the coding unit; and wherein in a first mode, variable-length-coding a residual coefficient to generate a coefficient code string, the residual coefficient being generated by performing an orthogonal transformation process and a quantizing process on a differential image between the image to be encoded and the generated predictive image, wherein in a second mode, directly using the differential image to generate the coefficient code string without performing the orthogonal transformation process and the quantizing process, and wherein a syntax of the current coding unit includes flag information indicating whether the first mode or the second mode is applied for the current coding unit and the syntax is included in a syntax of the coding unit separate from a syntax of the prediction unit included in the syntax of the coding unit, and wherein in a second mode, the orthogonal transformation process and the quantizing process are bypassed.
3 . An image decoding device that decodes an encoded image, the image decoding device comprising: a processor; and a non-transitory memory having stored thereon executable instructions, which cause the processor to: generate a predictive image on a predictive unit for a current coding unit using predictive information, the prediction unit being included in the coding unit; and wherein in a first mode, the processor variable-length-decodes a coefficient code string to output a residual coefficient and performs an inverse quantizing process and an inverse orthogonal transformation process on the residual coefficient to output a residual decoded image, the residual decoded image is a differential image between the image to be decoded and the generated predictive image, wherein in a second mode, the processor directly uses the coefficient code string to output the differential image without performing the inverse quantizing process and the orthogonal transformation process, and wherein a syntax of the current coding unit includes flag information indicating whether the first mode or the second mode is applied for the current coding unit and the syntax is included in a syntax of the coding unit separate from a syntax of the prediction unit included in the syntax of the coding unit, and wherein in a second mode, the orthogonal transformation process and the quantizing process are bypassed.
4 . An image decoding method that decodes an encoded image, the image decoding method comprising: generating a predictive image on a predictive unit for a current coding unit using predictive information, the prediction unit being included in the coding unit; and wherein in a first mode, variable-length-decoding a coefficient code string to output a residual coefficient and performing an inverse quantizing process and an inverse orthogonal transformation process on the residual coefficient to output a residual decoded image, the residual decoded image is a differential image between the image to be decoded and the generated predictive image, wherein in a second mode, directly using the coefficient code string to output the differential image without performing the inverse quantizing process and the orthogonal transformation process, and wherein a syntax of the current coding unit includes flag information indicating whether the first mode or the second mode is applied for the current coding unit and the syntax is included in a syntax of the coding unit separate from a syntax of the prediction unit included in the syntax of the coding unit, and wherein in a second mode, the orthogonal transformation process and the quantizing process are bypassed.
5 . A non-transitory computer readable medium storing a bitstream, the bitstream including an encoded signal and syntax information according to which a decoder performs a method comprising: generating a predictive image on a predictive unit for a current coding unit using predictive information, the prediction unit being included in the coding unit; and wherein in a first mode, variable-length-decoding a coefficient code string to output a residual coefficient and performing an inverse quantizing process and an inverse orthogonal transformation process on the residual coefficient to output a residual decoded image, the residual decoded image is a differential image between the image to be decoded and the generated predictive image, wherein in a second mode, directly using the coefficient code string to output the differential image without performing the inverse quantizing process and the orthogonal transformation process, and wherein a syntax of the current coding unit includes flag information indicating whether the first mode or the second mode is applied for the current coding unit and the syntax is included in a syntax of the coding unit separate from a syntax of the prediction unit included in the syntax of the coding unit, and wherein in a second mode, the orthogonal transformation process and the quantizing process are bypassed.

Description

CROSS REFERENCE TO RELATED APPLICATIONS This application is a continuation of U.S. application Ser. No. 18/385,588 filed on Oct. 31, 2023, which is a continuation of U.S. application Ser. No. 17/686,698, now U.S. Pat. No. 11,849,124, filed on Mar. 4, 2022, which is a continuation of U.S. application Ser. No. 16/751,967, now U.S. Pat. No. 11,303,910, filed on Jan. 24, 2020, which is a continuation of U.S. application Ser. No. 14/021,207, now U.S. Pat. No. 10,757,422, filed on Sep. 9, 2013, which is a U.S. continuation application of PCT International Patent Application Number PCT/JP2012/001665 filed on Mar. 9, 2012, which claims priority of Japanese Patent Application No. 2011-051221 filed on Mar. 9, 2011. The entire disclosures of the above-identified applications, including the specifications, drawings, and claims are incorporated herein by reference in their entirety. BACKGROUND 1. Technical Field The present disclosure relates to a video image encoding device that encodes a video image such that the video image is divided into blocks. 2. Related Art In recent years, with development of multimedia applications, it has been popularized that information of all kinds of media such as images, sounds, and texts are uniformly handled. Since a digitized image has huge volumes of data, an image information compression technique is indispensable to accumulate and transmit the image. On the other hand, to mutually operate compressed image data, standardization of compression techniques is also important. For example, as standards of measure of image compression techniques, H.261, H.263, and H.264 of ITU-T (telecommunication standardization sector in International Telecommunication Union), MPEG-1, MPEG-3, MPEG-4, and MPEG-4AVC of ISO/IEC (International Organization for Standardization), and the like are given. At present, a standardization action for a next-generation screen encoding scheme obtained by the collaboration between ITU-T and ISO/IEC and called HEVC has been advanced. In such encoding of a video image, each picture to be encoded is divided into encoding unit blocks, and redundancies in time and spatial directions are reduced in units of blocks to compress an amount of information. In inter predictive encoding to reduce a time redundancy, detection of motion and formation of a predictive image are performed in units of blocks with reference to a forward or a backward picture to obtain a differential image between the obtained predictive image and an input image of a block to be encoded. In intra predictive encoding to reduce a spatial redundancy, a predictive image is generated from pixel information of peripheral encoded blocks to obtain a differential image between the obtained predictive image and an input image of a block to be encoded. In addition, orthogonal transformation such as discrete cosine transformation and quantization are performed to the obtained differential image, and a code string is generated by using variable-length-coding to compress an amount of information. In decoding, the code string generated by the encoding process is analyzed to obtain prediction information and residual coefficient information, inter predictive decoding and intra predictive decoding are performed by using the prediction information to generate a predictive image, inverse quantization and inverse orthogonal transformation are performed to the residual coefficient information to generate a differential image, and the obtained predictive image and the differential image are added to each other to decompress a final output image. In H.264 (ITU-T H.264: Advanced video coding for generic audiovisual services (March 2010)), in order to restrict an upper limit of an amount of processing in each block, a maximum value of an amount of codes generated in each block is defined (more specifically, 3200 bits). When the above normal encoding process is performed, a code string including an amount of code larger than the maximum value of the amount of generated codes may be generated depending on the quality of an input image or conditions of a quantization process. For this reason, a special encoding mode called an IPCM is used to make it possible to suppress the amount of code of the code string to be smaller than the maximum value. The IPCM is different from a normal encoding mode, and is a mode in which pixel values of an input image are directly described as a bit string in a code string without performing generation, orthogonal transformation/quantization of a differential image by intra/inter prediction. In use of the mode, for example, when the format of an input image is YUV4:2:0 in which each pixel has 8 bits, a block of a luminance component has 16×16 pixels, and each block of two color-difference components has 8×8 pixels. For this reason, the total number of bites is 384 bytes, and the number of bits of the input image including information required for a header can be suppressed in an amount equal to or smalle