WO2011009289A1 - 一种在数字电视接收中实现图像旋转的方法和系统 - Google Patents

一种在数字电视接收中实现图像旋转的方法和系统 Download PDF

Info

Publication number
WO2011009289A1
WO2011009289A1 PCT/CN2010/001079 CN2010001079W WO2011009289A1 WO 2011009289 A1 WO2011009289 A1 WO 2011009289A1 CN 2010001079 W CN2010001079 W CN 2010001079W WO 2011009289 A1 WO2011009289 A1 WO 2011009289A1
Authority
WO
WIPO (PCT)
Prior art keywords
frame
block
module
motion vector
image block
Prior art date
Application number
PCT/CN2010/001079
Other languages
English (en)
French (fr)
Inventor
林国辉
刘宗伟
Original Assignee
北京中天联科微电子技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京中天联科微电子技术有限公司 filed Critical 北京中天联科微电子技术有限公司
Publication of WO2011009289A1 publication Critical patent/WO2011009289A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/60—Rotation of whole images or parts thereof
    • G06T3/606—Rotation of whole images or parts thereof by memory addressing or mapping
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/387—Composing, repositioning or otherwise geometrically modifying originals
    • H04N1/3877—Image rotation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/44—Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/85—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression

Definitions

  • the present invention relates to a method and system for implementing a digital television reception, such as mobile television, terrestrial broadcasting, and the like. Background technique
  • Digital TV is considered to be one of the most promising industries in the 21st century. It is also the focus of digital consumer electronic information products in the international arena. At present, digital TV reception has penetrated into various terminals, including mobile phones, PMPs, PSPs, PCs, and set-top boxes. and many more. China's domestic terrestrial broadcasting standards and mobile TV standards have also been established.
  • FIG. 1 is a schematic diagram of a typical block-based digital television decoding.
  • the user When decoding an image display, the user generally needs to rotate the image for display, especially for a mobile phone or a PMP screen, due to different viewing habits.
  • the technical problem to be solved by the present invention is to provide a method and system for realizing image rotation in digital television reception, which realizes rotational display of a video decoded image without increasing the number of accesses and storage space.
  • an image rotation is implemented in digital television reception.
  • the system includes: an entropy decoding module, receiving and decoding a video elementary stream, obtaining a discrete cosine transform DCT transform domain matrix of a block or block residual and a motion vector; and an inverse quantization module, performing weighted inverse quantization on the DCT transform domain matrix, and performing DCT
  • the coefficients of the transform domain matrix are inversely transformed to the values before quantization, and the coefficients of the weighted inverse quantized DCT transform domain matrix correspond to the result of the DCT transform of the coded image block or block residual;
  • the inverse DCT module the weighted inverse
  • the quantized DCT transform domain matrix is converted into a spatial domain matrix, that is, converted into 4x4 or 8x8 original image block or original image block residual; matrix transpose module, matrix transpose operation on original image block or original image block residual
  • the transposed image block is stored into the frame memory as part of the reconstructed frame; the frame memory stores the display frame,
  • the system for implementing image rotation further includes: a motion vector correction module, which corrects an original motion vector in the entropy decoding module according to the set rotation angle to obtain a rotation motion vector, and sends the rotation motion vector to the frame/ Inter prediction module; an intra/inter prediction module, when the prediction type is intra prediction, reading a pixel of a region pointed by the rotation motion vector from the frame memory as a prediction pixel, and using the predicted pixel and the original after transposition The image block residuals are added to obtain a reconstructed block, and the reconstructed block is written to the frame memory.
  • a motion vector correction module which corrects an original motion vector in the entropy decoding module according to the set rotation angle to obtain a rotation motion vector, and sends the rotation motion vector to the frame/ Inter prediction module
  • an intra/inter prediction module when the prediction type is intra prediction, reading a pixel of a region pointed by the rotation motion vector from the frame memory as a prediction pixel, and using the predicted pixel and the original after transposition The image block residual
  • the intra/inter prediction module further reads, as a prediction block, a pixel of a region pointed by the rotation motion vector from the predicted frame position of the frame memory when the prediction type is inter prediction, and the original after the prediction block and the transposition The image block residuals are added to obtain a reconstructed block, and the reconstructed block is written to the frame memory.
  • a method for implementing image rotation in digital television reception comprising the steps of: entropy decoding module receiving a video elementary stream, decoding the video elementary stream, and obtaining a discrete cosine of a block or block residual Transforming the DCT transform domain matrix and the motion vector; the inverse quantization module will weight the DCT transform domain matrix Quantization, inverse transforming each coefficient of the DCT transform domain matrix to the value before quantization, and the coefficients of the weighted inverse quantized DCT transform domain matrix correspond to the result of the DCT transform of the coded image block or block residual; the inverse DCT module Transforming the weighted inverse quantized DCT transform domain matrix into a spatial domain matrix, that is, converting to a 4x4 or 8x8 original image block or original image block residual; the matrix transpose module performs matrix transformation on the original image block or the original image block residual After the transposition, the original image block is stored as part of the reconstructed frame into the frame memory; the
  • the method further includes the following steps: according to the set rotation angle, the motion vector correction module corrects the original motion vector in the entropy decoding module to obtain a rotation motion vector, and rotates The motion vector is sent to the intra/inter prediction module; when the prediction type is intra prediction, the intra/inter prediction module reads the pixel of one region pointed by the rotation motion vector from the frame memory as the prediction pixel; The reconstructed block is added to the original image block residual after the transposition, and the reconstructed block is written into the reconstructed frame of the frame memory.
  • the method further includes the following steps: according to the set rotation angle, the motion vector correction module corrects the original motion vector in the entropy decoding module to obtain a rotation motion vector, and rotates The motion vector is sent to the intra/inter prediction module; when the prediction type is inter prediction, the intra/inter prediction module reads the pixel of one region pointed by the rotation motion vector from the predicted frame position of the frame memory as the prediction block; The prediction block is added to the transposed original image block residual to obtain a reconstructed block, and the reconstructed block is written into the reconstructed frame of the frame memory.
  • FIG. 1 is a schematic diagram of a system for decoding video in digital television reception in the prior art.
  • Figure 2 is a schematic diagram of a prior art system for implementing map transfer.
  • FIG. 3 is a schematic diagram of a system for implementing image block rotation according to the present invention.
  • FIG. 4 is a schematic diagram of a system for implementing image block residual rotation according to the present invention.
  • Fig. 5 shows an embodiment in which the 8 x 8 matrix is transposed at 90 degrees in the present invention.
  • Figure 6 shows a schematic representation of the relative position before rotation.
  • Figure 6 (b) shows the relative position after rotation.
  • FIG. 7 is a flow chart showing a method for implementing image block rotation according to the present invention.
  • FIG. 8 is a flow chart showing a method for implementing image block residual rotation by intra prediction according to the present invention.
  • FIG. 9 is a schematic diagram showing the positional relationship between a 4 ⁇ 4 image block and adjacent pixels when performing intra prediction, where a, b, c, d, e, f, g, h, I, j, k, l , m, n, o, p are the respective pixels of the 4 ⁇ 4 image block, ABCD is the lowest 4 pixels of the adjacent image block above the image block, and IJKL is the rightmost side of the left adjacent image block of the image block. 4 pixels, M is the pixel at the bottom right corner of the adjacent image block at the upper left of the image block.
  • FIG. 10 is a flow chart showing a method for implementing image block residual rotation by inter prediction according to the present invention. detailed description
  • FIG. 1 is a schematic diagram of a system for decoding video in digital television reception in the prior art, including an entropy decoding module 10, an inverse DCT module 20, an inverse quantization module 30, an intra/inter prediction module 40, and a frame memory 50.
  • the entropy decoding module 10 receives the video elementary stream ES, and decodes the video elementary stream to obtain a discrete cosine transform DCT transform domain matrix of image blocks or image block residuals and corresponding motion vectors.
  • the image macroblock and the image macroblock residual are basic units, and generally one image macroblock will include several luminance images. Blocks (such as 4 in MPEG2) and several chroma image blocks (such as 2 in MPEG2). In the same image macroblock, the operations of the image blocks are the same. In the following, the image block is used as a unit of description.
  • the image block may be a matrix of 4x4 (such as H.264) or 8x8 (such as MPEG2).
  • the entropy decoding module 10 receives the video elementary stream of other modules of the digital television receiving system, for example, the demultiplexing module demultiplexes the digital television transmission stream TS to obtain a video elementary stream and an audio elementary stream, wherein the video elementary stream is used as an input of the entropy decoding 10 .
  • the digital television receiving system obtains the video elementary stream directly from the hard disk or the network as an input to the entropy decoding module 10.
  • the digital television receiving system can be a digital television set top box, including terrestrial, satellite, cable or network. It can also be a mobile terminal with digital TV decoding function, such as mobile phone, PMP, etc.
  • the video elementary stream ES is a compressed code stream based on a DCT block transform, such as H.264, MPEGl/2/4, AVS, and the like.
  • the decoding algorithm according to the entropy decoding module 10 may be Huffman coding, such as MPEG1/2/3, or CABIC, CAVLC, such as H.264.
  • the resulting DCT transform domain matrix can be 4x4 (such as H.264) or 8x8 (such as MPEG2).
  • the inverse quantization module 30 performs weighted inverse quantization on the DCT transform domain matrix, and inverse transforms each coefficient of the DCT transform domain matrix to a value before quantization.
  • the coefficients of the inverse quantized DCT transform domain matrix correspond to the result of the DCT transform of the coded block or block residual.
  • the inverse DCT module 20 converts the weighted inverse quantized DCT transform domain matrix into a spatial domain matrix, i.e., into a 4x4 or 8x8 original image block or original image block residual.
  • the original image block When converted to the original image block, the original image block is sent to the frame memory 50 to construct a reconstructed frame 502.
  • a prediction block When converted to the original image block residual, a prediction block is obtained according to the prediction type, and the prediction block is added to the original image block residual to obtain a reconstructed block.
  • the prediction type includes intra-frame and inter-frame prediction, and the operation of acquiring the prediction block is performed by the intra/inter prediction module 40, which will be described below.
  • the intra/inter prediction module 40 predicts the original image block residual to predict The block adds the prediction block to the original image block residual to obtain a reconstructed frame 502.
  • the intra/inter prediction module 40 obtains the corresponding predicted pixel from the reconstructed frame 502 that has been decoded, and the predicted pixel will be from the corresponding position in the reconstructed frame 502 according to the intra prediction type (eg, Obtained from the boundary pixels of adjacent image macroblocks).
  • the predicted pixel is added to the corresponding pixel in the image block residual obtained by the inverse DCT module 20 to obtain a reconstructed block, and the reconstructed block is written into the corresponding position of the reconstructed frame 502 in the frame memory 50.
  • the intra/inter prediction module 40 will obtain the image block from the reconstructed frame 502 that has been decoded.
  • the bottom 4 pixels of the upper image block as shown in Figure 9 ABCD, and the four pixels and the current image block residuals ( a, b, c, d; e, f, g, h; i, j, k, 1; m, n, o, p)
  • the vertical addition adds the reconstructed block to (a+A, b+B, c+C, d+D; e+A, f+B, g+C, h+D; i+A, j+B, k+C, 1+D; m+A, n+B, o+C, p+D), the reconstructed block is written into the frame memory 50 to reconstruct the frame 502 The corresponding location.
  • the intra/inter prediction module 40 obtains a corresponding prediction block from the position of the prediction frame 503 in the frame memory 50 according to the motion vector obtained from the entropy decoding module 10, and the prediction block will be based on the frame.
  • the inter prediction type is derived from the corresponding position in the prediction frame 503 and is calculated by filtering.
  • the intra/inter prediction module 40 will read an image block, that is, a prediction block, from the position of (k+A, B+1) of the prediction frame 503, which is obtained by the prediction block and the inverse DCT module 20.
  • the image block residuals are added to obtain a reconstructed block which is written to the corresponding position of the reconstructed frame 502 in the frame memory 50.
  • the frame memory 50 stores a display frame 501, a reconstructed frame 502, and a predicted frame 503.
  • the reconstructed frame 502 is transformed into a predicted frame 503 during decoding for use in future decoding predictions.
  • the number of predicted frames 503 reaches the set value during decoding, that is, the solution
  • the code process no longer requires a certain prediction frame 503, which will be the display frame 501 for display.
  • FIG. 2 is a schematic diagram of a prior art system for implementing image rotation.
  • a rotation operation module 60 is further included to perform a rotation operation on the display frame 501.
  • the rotation operation module 60 reads out the display frame 501 from the frame memory 50, and performs a rotation operation on the entire display frame 501 according to the set rotation angle, and the operation result is obtained.
  • the write back to the frame memory 50 becomes a new display frame 501, and the display frame is read and displayed.
  • FIG. 3 is a schematic diagram of a system for implementing image block rotation according to the present invention, and the matrix transposition module 80 is added as compared with FIG.
  • the frame memory stores the display frame, the reconstructed frame and the predicted frame, and the reconstructed frame is transformed into a predicted frame during the decoding process.
  • the predicted frame becomes a display frame.
  • the matrix transposition module 80 performs a matrix transposition operation on the original image block output by the inverse DCT module 20 or the original image block residual according to the set rotation angle, and stores the transposed image block as part of the reconstruction frame.
  • the matrix transposition operation may be a rotation of 0, 90, 180, 270 degrees or a mirror image or the like.
  • Fig. 5 shows an embodiment in which the 8 x 8 matrix is transposed at 90 degrees in the present invention.
  • Performing a rotation operation on an image block avoids the operation of reading the display frame from the frame memory before constructing the display frame, eliminating the need for frame memory read times and additional storage space.
  • the prior art system may include only the entropy decoding module 10, the inverse DCT module 20, the inverse quantization module 30, and the frame memory 50.
  • this operation can also be performed by a system that also includes the intra/inter prediction module 40.
  • FIG. 4 is a schematic diagram of a system for implementing image block residual rotation according to the present invention, in FIG.
  • the motion vector correction module 70 is also included. The operations performed by the module will be specifically described below in conjunction with other modules of the prior art.
  • the motion vector correction module 70 corrects the original motion vector in the entropy decoding module 10 according to the set rotation angle to obtain a rotation motion vector, and transmits the rotation motion vector to the intra/inter prediction module 40.
  • the intra/inter prediction module 40 reads, from the frame memory reconstruction frame 502, a pixel of a region pointed by the rotation motion vector as a prediction pixel, the predicted pixel and the transposed original image block. The residuals are added to obtain a reconstructed block which is written to the corresponding location of the reconstructed frame 502 in frame memory 50.
  • the prior art and the content of the present invention are consistent in this partial reading manner, except that the prior art does not have a rotary motion vector, which is based on the original motion vector obtained from the code stream.
  • the intra/inter prediction module 40 will obtain the image block from the reconstructed frame 502 that has been decoded.
  • the bottom 4 pixels of the upper image block as shown in Figure 9 ABCD, and the four pixels and the current image block residuals (a, b, c, d; e, f, g, h; i, j, k, 1; m, n, o, p)
  • the vertical addition adds the reconstructed block to (a+A, b+B, c+C, d+D; e+A, f+B, g+C, h+D; i+A, j+B, k+C, 1+D; m+A, n+B, o+C, p+D), the reconstructed block is written into the frame memory 50 to reconstruct the frame 502 The corresponding location.
  • the intra/inter prediction module 40 reads, from the prediction frame 503 in the frame memory 50, the pixel of one region pointed by the rotation motion vector as a prediction block, and the original after the prediction block and the transposition The image block residuals are added to obtain a reconstructed block, and the reconstructed block is written to the frame memory.
  • the prediction block will be derived from the corresponding position in the prediction frame 503 according to the inter prediction type and calculated by filtering.
  • Figure 6 (b) A schematic of an embodiment of the relative position after 90 degrees of rotation is shown.
  • the motion vector When rotated 90 degrees, the motion vector will be corrected to (-B, A), and the corrected reference block position will be (VY, X), ie (VB-1, k+A), intra/inter prediction module 40 will read an image block from the (Vbl, k+A) position of the prediction frame 503, that is, a prediction block, and the prediction matrix is added to the reconstructed image block residual to obtain a reconstructed block, and the reconstructed block is written. The corresponding location of frame 502 is reconstructed in frame memory 50.
  • the correction of other angles can be obtained in the same way.
  • the specific correction method is as follows:
  • the motion vector is corrected to (-A, -B), and the corrected reference block position is (H-X, V-Y), ie (H-k-A, V-B-I).
  • the motion vector is corrected to (B, -A), and the corrected reference block position is (Y, H-X), that is, (B+l, H-A-k).
  • the output of the motion vector correction module 70 is also corrected. Rotational motion vector.
  • the rotated image block residual is obtained, and then the reconstructed block obtained by adding the prediction pixel or the prediction block to the image block residual is compared with the original image. .
  • the matrix transposition module 80 is added in the decoding process of the prior art without increasing the capacity and the number of readings of the frame memory 50, thereby realizing the rotation operation of the image block and displaying, and further, the motion vector correction can be added.
  • the module 70 implements a rotation operation on the image block residual and displays it.
  • FIG. 7 is a flow chart of a method for implementing image block rotation according to the present invention, including the following steps:
  • the entropy decoding module 10 receives the video elementary stream, and decodes the video elementary stream to obtain a discrete cosine transform DCT transform domain matrix of block or macro residuals and a motion vector.
  • step 702 the inverse quantization module 30 performs weighted inverse quantization on the DCT transform domain matrix, and inverse transforms each coefficient of the DCT transform domain matrix to a value before quantization, and the coefficients of the weighted inverse quantized DCT transform domain matrix correspond to The result of the DCT transform after the coding end image block or block residual.
  • the inverse DCT module 20 converts the weighted inverse quantized DCT transform domain matrix into a spatial domain matrix, i.e., into a 4x4 or 8x8 original image block or original image block residual.
  • the matrix transposition module 80 performs a matrix transposition operation on the original image block or the original image block residual output by the inverse DCT module 20, and stores the transposed image block as a part of the reconstructed frame into the frame memory 50.
  • the reconstructed frame 502 is transformed into a predicted frame 503 during decoding for use in future decoding predictions. When the number of predicted frames reaches the set value during decoding, the predicted frame will become the display frame.
  • the matrix transposition operation may be a rotation of 0, 90, 180, 270 degrees or a mirror image or the like.
  • FIG. 8 is a flowchart of a method for implementing image block residual rotation by intra prediction according to the present invention.
  • the method further includes the following steps in step 801, according to the set rotation angle.
  • motion vector correction module 70 The original motion vector in the entropy decoding module 10 is modified to obtain a rotated motion vector, and the rotated motion vector is transmitted to the intra/inter prediction module 40.
  • the intra/inter prediction module 40 reads the pixel of one region pointed by the rotational motion vector from the reconstructed frame 502 of the frame memory 50 as the predicted pixel.
  • step 803 the predicted pixel and the transposed image block residual are added to obtain a reconstructed block, and the reconstructed block is written into the corresponding position of the reconstructed frame 502 in the frame memory 50.
  • the prior art and the content of the present invention are consistent in this partial reading manner, except that the prior art does not have a rotary motion vector, which is based on the original motion vector obtained from the code stream.
  • FIG. 10 is a flowchart of a method for implementing image block residual rotation by inter prediction according to the present invention. Steps 701 to 704 of FIG. 7 further include the following steps.
  • step 1001 according to the set rotation angle, the motion vector correction module 70 corrects the original motion vector in the entropy decoding module 10 to obtain a rotation motion vector, and transmits the rotation motion vector to the intra/inter prediction module 40. .
  • step 1002 when the prediction type is inter prediction, the intra/inter prediction module 40 reads the pixel of one region pointed by the rotational motion vector from the position of the prediction frame 503 of the frame memory 50 as a prediction block.
  • step 1003 the prediction block is added to the transposed original image block residual to obtain a reconstructed block, and the reconstructed block is written into the reconstructed frame 502 of the frame memory 50.
  • the motion vector correction module 70 outputs the corrected rotational motion vector. .
  • the rotated image block residual is obtained, and then the reconstructed block obtained by adding the prediction pixel or the prediction block to the image block residual is compared with the original image. .

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Description

一种在数字电视接收中实现图像旋转的方法和系统 技术领域
本发明涉及在数字电视接收中， 如移动电视、 地面广播等， 实 现图^ t转的方法和系统。 背景技术
数字电视被认为是 21 世纪前景最广阔的产业之一， 也是国际 范围内数字消费类电子信息产品^的重点， 目前数字电视接收已 经渗透到各种终端， 包括手机、 PMP、 PSP、 PC、 机顶盒等等。 中国本土的地面广播标准和移动电视标准也已经确立。
基于块的 DCT 变换（离散余弦变换）算法被广泛用于各种视 频标准中， 包括 MPEG1/2/4, H.26x 和 AVS中。 图 1为一典型的 基于块变换的数字电视解码示意图。
在解码图像显示时， 由于观看习惯的不同， 用户一般都需要对 图像进行旋转显示， 尤其是对手机或者 PMP屏幕。
传统的旋转显示的实现方法需要从显示帧緩存中读取数据， 然 后旋转， 最后存入到显示帧緩存中 （很可能是另一段区域）， 如图 2 所示。 由于解码后的图像数据量较大， 这样的实现将会增加对存 储器的存取次数， 甚至需要增加对存储大小的要求， 从而导致终端 价格增加、 功耗增大。 发明内容
本发明要解决的技术问题提供一种在数字电视接收中实现图像 旋转的方法和系统， 在不增加存取次数和存储空间的情况下实现对 视频解码后的图像进行旋转显示。
根据本发明一方面， 提出一种在数字电视接收中实现图像旋转 的系统， 包括： 熵解码模块， 接收并解码视频基本流， 得到块或者 块残差的离散余弦变换 DCT 变换域矩阵以及运动矢量； 反量化模 块， 对 DCT 变换域矩阵进行加权反量化， 将 DCT 变换域矩阵的 各个系数反变换到量化前的数值， 经加权反量化后的 DCT 变换域 矩阵的系数对应于编码端图像块或块残差经过 DCT 变换后的结 果； 反 DCT模块， 将加权反量化后的 DCT 变换域矩阵转换为空 间域矩阵， 即转换为 4x4或 8x8的原始图像块或原始图像块残差； 矩阵转置模块， 对原始图像块或原始图像块残差进行矩阵转置运 算， 将转置后图像块作为重建帧的一部分存入到帧存储器； 帧存储 器， 存储显示帧、 重建帧和预测帧， 重建帧在解码过程中变换为预 测帧， 当解码过程中预测帧的数目达到设定值时， 预测帧将会成为 显示帧。
进一步， 所述实现图像旋转的系统还包括： 运动矢量修正模 块， 根据设定的旋转角度， 对熵解码模块中的原始运动矢量进行修 正得到旋转运动矢量， 并将旋转运动矢量发送给帧内 /帧间预测模 块； 帧内 /帧间预测模块， 当预测类型为帧内预测时， 从帧存储器 中读取旋转运动矢量指向的一个区域的像素作为预测像素， 将预测 像素与转置后的原始图像块残差相加得到重建块， 并将重建块写入 帧存储器。
进一步， 帧内 /帧间预测模块还在预测类型为帧间预测时， 从 帧存储器的预测帧位置读取旋转运动矢量指向的一个区域的像素作 为预测块， 将预测块与转置后的原始图像块残差相加得到重建块， 并将重建块写入帧存储器。
根据本发明另一方面， 还提出一种在数字电视接收中实现图像 旋转的方法， 包括以下步骤： 熵解码模块接收视频基本流， 将视频 基本流进行解码， 得到块或者块残差的离散余弦变换 DCT 变换域 矩阵以及运动矢量； 反量化模块将对 DCT 变换域矩阵进行加权反 量化， 将 DCT 变换域矩阵的各个系数反变换到量化前的数值， 经 加权反量化后的 DCT 变换域矩阵的系数对应于编码端图像块或块 残差经过 DCT 变换后的结果； 反 DCT 模块将加权反量化后的 DCT变换域矩阵转换为空间域矩阵， 即转换为 4x4或 8x8的原始 图像块或原始图像块残差； 矩阵转置模块对原始图像块或原始图像 块残差进行矩阵转置运算， 将转置后原始图像块作为重建帧的一部 分存入到帧存储器； 帧存储器存储显示帧、 重建帧和预测帧， 重建 帧在解码过程中变换为预测帧， 当解码过程中预测帧的数目达到设 定值时， 预测帧将会成为显示帧。
进一步， 对原始图像块残差进行矩阵转置运算后， 还包括以下 步骤： 根据设定的旋转角度， 运动矢量修正模块对熵解码模块中的 原始运动矢量进行修正得到旋转运动矢量， 并将旋转运动矢量发送 给帧内 /帧间预测模块； 当预测类型为帧内预测时， 帧内 /帧间预测 模块从帧存储器中读取旋转运动矢量指向的一个区域的像素作为预 测像素； 将预测像素与转置后的原始图像块残差相加得到重建块， 将重建块写入帧存储器的重建帧中。
进一步， 对原始图像块残差进行矩阵转置运算后， 还包括以下 步骤： 根据设定的旋转角度， 运动矢量修正模块对熵解码模块中的 原始运动矢量进行修正得到旋转运动矢量， 并将旋转运动矢量发送 给帧内 /帧间预测模块； 当预测类型为帧间预测时， 帧内 /帧间预测 模块从帧存储器的预测帧位置读取旋转运动矢量指向的一个区域的 像素作为预测块； 将预测块与转置后的原始图像块残差相加得到重 建块， 将重建块写入帧存储器的重建帧中。
与现有技术相比， 本发明消除了对存储器读取次数的额外要 求， 消除了对存储器空间的额外要求， 并且降低了功耗。 附图说明 图 1 所示为现有技术的数字电视接收中视频解码的系统示意 图。
图 2所示为现有技术的实现图傳 转的系统示意图。
图 3所示为本发明实现图像块旋转的系统示意图。
图 4所示本发明实现图像块残差旋转的系统示意图。
图 5所示为本发明中 8 x 8矩阵经过 90度转置的实施例。
图 6 )所示为旋转前相对位置示意图。
图 6 ( b )所示为旋转后相对位置示意图。
图 7所示为本发明实现对图像块旋转的方法流程图。
图 8所示为本发明通过帧内预测实现对图像块残差旋转的方法 流程图。
图 9所示为一 4x4的图像块在做帧内预测时与相邻像素的位置 关系的示意图， 其中 a，b，c，d，e，f，g，h，I，j，k，l，m，n，o，p为所述 4x4图像块 的各个像素， ABCD为此图像块上方相邻图像块的最下方的 4个像 素， IJKL为此图像块左方相邻图像块的最右方的 4个像素， M为 此图像块左上方相邻图像块的最右下角的像素。
图 10 所示为本发明通过帧间预测实现对图像块残差旋转的方 法流程图。 具体实施方式
图 1 所示为现有技术的数字电视接收中视频解码的系统示意 图， 包括熵解码模块 10、 反 DCT模块 20、 反量化模块 30、 帧内 / 帧间预测模块 40和帧存储器 50。
熵解码模块 10接收视频基本流 ES， 将所述视频基本流进行解 码， 得到图像块或者图像块残差的离散余弦变换 DCT 变换域矩阵 和相应运动矢量。 所涉及的基于块的压缩标准中， 图像宏块和图像 宏块残差为基^算单位， 一般一个图像宏块将包括几个亮度图像 块 （如 MPEG2 中包括 4个）和几个色度图像块 (如 MPEG2中 包括 2 个）。 在同一图像宏块中， 所述图像块的运算都一致， 以下 将以图像块作为叙述单位， 此图像块可以是 4x4 (如 H.264 )也可 以是 8x8 (如 MPEG2 ) 的矩阵。
熵解码模块 10 接收数字电视接收系统其他模块的视频基本 流， 如解复用模块将数字电视传输流 TS解复用后得到视频基本流 和音频基本流， 其中视频基本流作为熵解码 10 的输入。 再如数字 电视接收系统直接从硬盘或者网络得到视频基本流， 作为熵解码模 块 10 的输入。 所述数字电视接收系统可以是数字电视机顶盒， 包 括地面、 卫星、 有线或者是网络。 也可以是带有数字电视解码功能 的移动终端， 如手机、 PMP等。 所述视频基本流 ES是基于 DCT 块变换的压缩码流， 如 H.264、 MPEGl/2/4、 AVS等。 熵解码模块 10 依据的解码算法可以是霍夫曼编码、 如 MPEG1/2/3 , 或者是 CABIC、 CAVLC、 如 H.264 等。 得到的 DCT 变换域矩阵可以是 4x4 (如 H.264 )也可以是 8x8 (如 MPEG2 )。
反量化模块 30对所述 DCT 变换域矩阵进行加权反量化， 将 DCT 变换域矩阵的各个系数反变换到量化前的数值。 经反量化后 的 DCT变换域矩阵的系数对应于编码端图像块或块残差经过 DCT 变换后的结果。
反 DCT模块 20将加权反量化后的 DCT变换域矩阵转换为空 间域矩阵， 即转换为 4x4或 8x8的原始图像块或原始图像块残差。
当转换为原始图像块时， 将所迷原始图像块发送给帧存储器 50， 以构建重建帧 502。 当转换为原始图像块残差时， 则根据预测 类型得到预测块， 将所述预测块与原始图像块残差相加得到重建 块。 其中， 预测类型包括帧内、 帧间预测， 获取预测块的操作由帧 内 /帧间预测模块 40执行， 将在下面进行说明。
帧内 /帧间预测模块 40 对原始图像块残差进行预测得到预测 块， 将预测块与原始图像块残差相加得到重建帧 502。
当预测类型为帧内预测时， 帧内 /帧间预测模块 40从已经解码 完成的重建帧 502得到相应的预测像素， 该预测像素将根据帧内预 测类型从重建帧 502 中的相应位置 （如相邻的图像宏块的边界像 素）得到。 该预测像素与反 DCT模块 20得到的图像块残差中相应 像素相加得到重建块， 将该重建块写入帧存储器 50 中重建帧 502 的相应位置。 以 H.264标准为例， 如果帧内预测类型为 4x4亮度帧 内预测模式 1， 即垂直预测， 则帧内 /帧间预测模块 40将从已经解 码完成的重建帧 502中得到所述图像块上方图像块的最下方的 4个 像素， 如图 9 所示的 ABCD, 并将此四个像素与当前图像块残差 ( a， b， c， d; e， f， g， h; i， j， k， 1; m， n， o， p )垂直相加 得到重建块为 （a+A， b+B, c+C, d+D; e+A, f+B, g+C, h+D; i+A, j+B, k+C， 1+D; m+A, n+B, o+C, p+D ), 将该重建块写 入帧存储器 50中重建帧 502的相应位置。
当预测类型为帧间预测时， 帧内 /帧间预测模块 40根据从熵解 码模块 10中得到的运动矢量， 从帧存储器 50中预测帧 503位置得 到相应的预测块， 该预测块将根据帧间预测类型从预测帧 503中的 相应位置得到像素并经过滤波计算得到。 假设帧间预测类型为 P 帧， 即前向预测， 而且运动矢量为 （A， B ), 而所述图像块残差在 重建帧 502的位置（块的第一个像素位置） 为 （k, 1 )， 则帧内 /帧 间预测模块 40将会从预测帧 503的 （k+A， B+1 )的位置读取一个 图像块， 即预测块， 该预测块与反 DCT模块 20得到的图像块残差 相加得到重建块， 将该重建块写入帧存储器 50 中重建帧 502的相 应位置。
帧存储器 50存储显示帧 501、 重建帧 502和预测帧 503。 其 中， 重建帧 502在解码过程中变换为预测帧 503， 以供未来的解码 预测所用。 当解码过程中预测帧 503的数目达到设定值时， 即该解 码过程不再需要某一预测帧 503， 该预测帧 503 将会成为显示帧 501 , 供显示所用。
图 2所示为现有技术的实现图像旋转的系统示意图。 在图 1的 基础上， 还包括旋转运算模块 60， 对显示帧 501执行旋转操作。
在图 1过程中得到位于帧存储器 50中的显示帧 501后， 旋转 运算模块 60从帧存储器 50中读出显示帧 501， 根据设定的旋转角 度对整个显示帧 501 做旋转运算， 将运算结果写回到帧存储器 50 中成为新的显示帧 501， 并对该显示帧进行读取显示。
下面结合附图和实施例进一步详细说明本发明的工作原理。 图 3所示为本发明实现对图像块旋转的系统示意图， 与图 1相 比增加了矩阵转置模块 80。
帧存储器， 存储显示帧、 重建帧和预测帧， 重建帧在解码过程 中变换为预测帧， 当解码过程中预测帧的数目达到设定值时， 预测 帧将会成为显示帧。
矩阵转置模块 80根据设定的旋转角度， 对反 DCT模块 20输 出的原始图像块或所述原始图像块残差进行矩阵转置运算， 将转置 后图像块作为重建帧的一部分存入到帧存储器 50 的重建帧 502 中， 所述矩阵转置运算可以是旋转 0、 90、 180、 270度或镜像等。 图 5所示为本发明中 8 x 8矩阵经过 90度转置的实施例。
对图像块执行旋转运算操作是在构建显示帧之前， 避免了从帧 存储器中读取显示帧的操作， 消除了对帧存储器读取次数以及对存 储空间的额外要求。
由于此时该系统用于对图像块执行旋转操作， 因此， 现有技术 的系统中可以只包括熵解码模块 10、 反 DCT模块 20、 反量化模块 30和帧存储器 50。 当然， 该操作也可以 于还包括帧内 /帧间预 测模块 40的系统所完成的。
图 4所示为本发明实现图像块残差旋转的系统示意图， 在图 3 的基础上还包括运动矢量修正模块 70。 下面将结合现有技术的其 他模块来具体说明该模块执行的操作。
运动矢量修正模块 70根据设定的旋转角度， 对熵解码模块 10 中的原始运动矢量进行修正得到旋转运动矢量， 并将所述旋转运动 矢量发送给帧内 /帧间预测模块 40。
帧内 /帧间预测模块 40在预测类型为帧内预测时， 从帧存储器 重建帧 502中读取旋转运动矢量指向的一个区域的像素作为预测像 素， 该预测像素与转置后的原始图像块残差相加得到重建块， 将该 重建块写入帧存储器 50 中重建帧 502的相应位置。 现有技术与本 发明的内容在此部分读取方式是一致的， 不同之处是现有技术没有 旋转运动矢量之说， 所依据的是从码流中得到的原始运动矢量。
以 H.264标准为例， 如果帧内预测类型为 4x4亮度帧内预测模 式 1， 即垂直预测， 则帧内 /帧间预测模块 40将从已经解码完成的 重建帧 502中得到所述图像块上方图像块的最下方的 4个像素， 如 图 9 所示的 ABCD， 并将此四个像素与当前图像块残差 （a， b, c， d; e, f， g, h; i， j, k， 1； m, n， o， p )垂直相加得到重建 块为 （a+A , b+B , c+C , d+D; e+A , f+B , g+C , h+D; i+A , j+B, k+C, 1+D; m+A, n+B, o+C, p+D ), 将该重建块写入帧存 储器 50中重建帧 502的相应位置。
帧内 /帧间预测模块 40在预测类型为帧间预测时， 从帧存储器 50 中预测帧 503位置读取旋转运动矢量指向的一个区域的像素作 为预测块， 将预测块与转置后的原始图像块残差相加得到重建块， 并将重建块写入帧存储器。 该预测块将根据帧间预测类型从预测帧 503中的相应位置得到像素并经过滤波计算得到。
图 6 ( a ) 所示为旋转前相对位置的实施例示意图， 假设帧间 预测类型为 P帧， 即前向预测， 解码后的运动矢量为 （A， B )， 而 所述图像块残差在重建帧 502 的位置 （块的第一个像素位置） 为 (k， I), 则原参考图像块的位置是 （X, Y)， 其中 X=k+A, Y=b+I, 并且假设图像的大小为 （H， V)o 图 6 (b) 所示为旋转 90度后相对位置的实施例示意图。 当旋转 90度后， 运动矢量将被 修正为 (-B, A), 修正后的参考块位置为 （V-Y， X), 即 （V-B-1, k+A)， 帧内 /帧间预测模块 40将会从预测帧 503的 （V-b-l， k+A) 位置读取一个图像块， 即预测块， 该预测矩阵与转置后的图像块残 差相加得到重建块， 将该重建块写入帧存储器 50 中重建帧 502的 相应位置。
旋转其他角度的修正同理可得， 具体修正方法如下：
旋转 180度
运动矢量被修正为 （-A， -B), 修正后的参考块位置为 （H- X, V-Y), 即 （H-k-A， V-B-I ).
旋转 270度
运动矢量被修正为 （B， -A), 修正后的参考块位置为 （Y， H- X), 即 （B+l， H-A-k)。
由于才艮据重建帧 502得到的预测像素、 以及才艮据预测帧 503得 到的预测块都是经过了矩阵转置模块 80 的矩阵转置运算， 而运动 矢量修正模块 70 输出的也是进行修正得到的旋转运动矢量。 同 时， 经过矩阵转置模块 80 转置后得到的是旋转的图像块残差， 于 是， 预测像素或预测块与图像块残差相加得到的重建块， 与原始图 比已经是旋转后的图像。
本发明在不增加帧存储器 50 容量和读取次数的情况下， 在现 有技术的解码过程中增加矩阵转置模块 80， 实现对图像块的旋转 操作并显示， 进一步， 还可以增加运动矢量修正模块 70， 实现对 图像块残差的旋转操作并显示。
图 7所示为本发明实现对图像块旋转的方法流程图， 包括以下 步骤： 在步骤 701， 熵解码模块 10接收视频基本流， 将所述视频基 本流进行解码， 得到块或者宏残差的离散余弦变换 DCT 变换域矩 阵以及运动矢量。
在步骤 702， 反量化模块 30对所述 DCT变换域矩阵进行加权 反量化， 将 DCT 变换域矩阵的各个系数反变换到量化前的数值， 经加权反量化后的 DCT 变换域矩阵的系数对应于编码端图像块或 块残差经过 DCT变换后的结果。
在步骤 703， 反 DCT模块 20将加权反量化后的 DCT变换域 矩阵转换为空间域矩阵， 即转换为 4x4或 8x8的原始图像块或原始 图像块残差。
在步骤 704， 矩阵转置模块 80对反 DCT模块 20输出的原始 图像块或所述原始图像块残差进行矩阵转置运算， 将转置后图像块 作为重建帧的一部分存入到帧存储器 50的重建帧 502 中， 重建帧 502 在解码过程中变换为预测帧 503， 以供未来的解码预测所用。 当解码过程中预测帧的数目达到设定值时， 预测帧将会成为显示 帧。
所述矩阵转置运算可以是旋转 0、 90、 180、 270度或镜像等。 当解码过程中预测帧 503的数目达到设定值时， 即该解码过程不再 需要某一预测帧 503， 该预测帧 503将会成为显示帧 501， 供显示 所用。 所显示的图像即是对原始图 转后的图像。
由于对图像块执行旋转运算操作是在构建显示帧之前， 避免了 从帧存储器中读取显示帧的操作， 消除了对帧存储器读取次数以及 对存储空间的额外要求。
图 8所示为本发明通过帧内预测实现对图像块残差旋转的方法 流程图， 在图 7 的步骤 701 ~步骤 704 的基础上， 还包括以下步 在步骤 801 , 根据设定的旋转角度， 运动矢量修正模块 70 对 熵解码模块 10 中的原始运动矢量进行修正得到旋转运动矢量， 并 将所述旋转运动矢量发送给帧内 /帧间预测模块 40。 在步驟 802， 当预测类型为帧内预测时， 帧内 /帧间预测模块 40 从帧存储器 50 的重建帧 502中读取旋转运动矢量指向的一个区域的像素作为预测 像素。
在步骤 803， 将预测像素与转置后的图像块残差相加得到重建 块， 将该重建块写入帧存储器 50 中重建帧 502 的相应位置。 现有 技术与本发明的内容在此部分读取方式是一致的， 不同之处是现有 技术没有旋转运动矢量之说， 所依据的是从码流中得到的原始运动 矢量。
图 10 所示为本发明通过帧间预测实现对图像块残差旋转的方 法流程图， 在图 7的步驟 701 ~步驟 704的 上， 还包括以下步 .
在步骤 1001， 根据设定的旋转角度， 运动矢量修正模块 70对 熵解码模块 10 中的原始运动矢量进行修正得到旋转运动矢量， 并 将所述旋转运动矢量发送给帧内 /帧间预测模块 40。
在步骤 1002， 当预测类型为帧间预测时， 帧内 /帧间预测模块 40从帧存储器 50的预测帧 503位置读取旋转运动矢量指向的一个 区域的像素作为预测块。
在步骤 1003， 将预测块与转置后的原始图像块残差相加得到 重建块， 将重建块写入帧存储器 50的重建帧 502中。
由于根据重建帧 502得到的预测像素、 以及根据预测帧 503得 到的预测块都是经过了矩阵转置模块 80 的矩阵转置运算， 而运动 矢量修正模块 70 输出的也是进行修正得到的旋转运动矢量。 同 时， 经过矩阵转置模块 80 转置后得到的是旋转的图像块残差， 于 是， 预测像素或预测块与图像块残差相加得到的重建块， 与原始图 比已经是旋转后的图像。

Claims

权 利 要 求
1. 一种在数字电视接收中实现图^ Nt转的系统， 包括：
熵解码模块， 接收并解码视频基本流， 得到块或者块残差的离 散余弦变换 DCT变换域矩阵以及运动矢量；
反量化模块， 对 DCT 变换域矩阵进行加权反量化， 将 DCT 变换域矩阵的各个系数反变换到量化前的数值， 经加权反量化后的 DCT变换域矩阵的系数对应于编码端图像块或块残差经过 DCT变 换后的结果；
反 DCT模块， 将加权反量化后的 DCT 变换域矩阵转换为空 间域矩阵， 即转换为 4x4或 8x8的原始图像块或原始图像块残差； 矩阵转置模块， 对原始图像块或原始图像块残差进行矩阵转置 运算， 将转置后图像块作为重建帧的一部分存入到帧存储器；
帧存储器， 存储显示帧、 重建帧和预测帧， 重建帧在解码过程 中变换为预测帧， 当解码过程中预测帧的数目达到设定值时， 预测 帧将会成为显示帧。
2. 根据权利要求 1所述实现图 转的系统， 还包括：
运动矢量修正模块， 根据设定的旋转角度， 对熵解码模块中的 原始运动矢量进行修正得到旋转运动矢量， 并将旋转运动矢量发送 给帧内 /帧间预测模块；
帧内 /帧间预测模块， 当预测类型为帧内预测时， 从帧存储器 中读取旋转运动矢量指向的一个区域的像素作为预测像素， 将预测 像素与转置后的原始图像块残差相加得到重建块， 并将重建块写入 帧存储器。
3. 根据权利要求 2所述实现图 转的系统， 其中， 帧内 /帧间 预测模块还在预测类型为帧间预测时， 从帧存储器的预测帧位置读 取旋转运动矢量指向的一个区域的像素作为预测块， 将预测块与转 置后的原始图像块残差相加得到重建块， 并将重建块写入帧存储 器。
4. 一种在数字电视接收中实现图像旋转的方法， 包括以下步 熵解码模块接收视频基本流， 将视频基本流进行解码， 得到块 或者块残差的离散余弦变换 DCT变换域矩阵以及运动矢量；
反量化模块将对 DCT 变换域矩阵进行加权反量化， 将 DCT 变换域矩阵的各个系数反变换到量化前的数值， 经加权反量化后的 DCT变换域矩阵的系数对应于编码端图像块或块残差经过 DCT变 换后的结果；
反 DCT模块将加权反量化后的 DCT 变换域矩阵转换为空间 域矩阵， 即转换为 4x4或 8x8的原始图像块或原始图像块残差； 矩阵转置模块对原始图像块或原始图像块残差进行矩阵转置运 算， 将转置后原始图像块作为重建帧的一部分存入到帧存储器； 帧存储器存储显示帧、 重建帧和预测帧， 重建帧在解码过程中 变换为预测帧， 当解码过程中预测帧的数目达到设定值时， 预测帧 将会成为显示帧。
5. 根据权利要求 4 所述实现图^ 转的方法， 对原始图像块残 差进行矩阵转置运算后， 还包括以下步骤：
根据设定的旋转角度， 运动矢量修正模块对熵解码模块中的原 始运动矢量进行修正得到旋转运动矢量， 并将旋转运动矢量发送给 帧内 /帧间预测模块；
当预测类型为帧内预测时， 帧内 /帧间预测模块从帧存储器中 读取旋转运动矢量指向的一个区域的像素作为预测像素；
将预测像素与转置后的原始图像块残差相加得到重建块， 将重 建块写入帧存储器的重建帧中。
6. 根据权利要求 4 所述实现图 ^转的方法， 对原始图像块残 差进行矩阵转置运算后， 还包括以下步骤：
根据设定的旋转角度， 运动矢量修正模块对熵解码模块中的原 始运动矢量进行修正得到旋转运动矢量， 并将旋转运动矢量发送给 帧内 /帧间预测模块；
当预测类型为帧间预测时， 帧内 /帧间预测模块从帧存储器的 预测帧位置读取旋转运动矢量指向的一个区域的像素作为预测块； 将预测块与转置后的原始图像块残差相加得到重建块， 将重建 块写入帧存储器的重建帧中。
PCT/CN2010/001079 2009-07-21 2010-07-16 一种在数字电视接收中实现图像旋转的方法和系统 WO2011009289A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200910158216.X 2009-07-21
CN 200910158216 CN101964881A (zh) 2009-07-21 2009-07-21 一种在数字电视接收中实现图像旋转的方法和系统

Publications (1)

Publication Number Publication Date
WO2011009289A1 true WO2011009289A1 (zh) 2011-01-27

Family

ID=43498731

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2010/001079 WO2011009289A1 (zh) 2009-07-21 2010-07-16 一种在数字电视接收中实现图像旋转的方法和系统

Country Status (2)

Country Link
CN (1) CN101964881A (zh)
WO (1) WO2011009289A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101892329B1 (ko) * 2011-11-03 2018-08-27 톰슨 라이센싱 이미지 정교화에 기초한 비디오 인코딩 및 디코딩

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1170310A (zh) * 1996-05-31 1998-01-14 索尼公司 图象解码装置和方法以及图象重放装置
CN1551635A (zh) * 2003-05-08 2004-12-01 ���µ�����ҵ��ʽ���� 图像解码单元和图像编码、解码装置和编码、解码方法
US20060147122A1 (en) * 2004-12-31 2006-07-06 Kadagattur Srinidhi Method and apparatus for processing a compressed image in an order other than the order in which it was compressed
US20070147694A1 (en) * 2005-12-22 2007-06-28 Sunplus Technology Co., Ltd. Block decoding method and apparatus capable of decoding and outputting data in a longitudinal direction
CN101059952A (zh) * 2007-04-12 2007-10-24 深圳安凯微电子技术有限公司 一种图像旋转处理方法、装置及多媒体处理器

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1170310A (zh) * 1996-05-31 1998-01-14 索尼公司 图象解码装置和方法以及图象重放装置
CN1551635A (zh) * 2003-05-08 2004-12-01 ���µ�����ҵ��ʽ���� 图像解码单元和图像编码、解码装置和编码、解码方法
US20060147122A1 (en) * 2004-12-31 2006-07-06 Kadagattur Srinidhi Method and apparatus for processing a compressed image in an order other than the order in which it was compressed
US20070147694A1 (en) * 2005-12-22 2007-06-28 Sunplus Technology Co., Ltd. Block decoding method and apparatus capable of decoding and outputting data in a longitudinal direction
CN101059952A (zh) * 2007-04-12 2007-10-24 深圳安凯微电子技术有限公司 一种图像旋转处理方法、装置及多媒体处理器

Also Published As

Publication number Publication date
CN101964881A (zh) 2011-02-02

Similar Documents

Publication Publication Date Title
TWI705702B (zh) 圖像處理裝置及方法
US8009740B2 (en) Method and system for a parametrized multi-standard deblocking filter for video compression systems
NO342829B1 (no) Datamaskinlesbart lagringsmedium og apparat for koding av et flertall av videobilder ved bruk av en rekkefølgeverdi
US20030095603A1 (en) Reduced-complexity video decoding using larger pixel-grid motion compensation
JP2004140473A (ja) 画像情報符号化装置、復号化装置並びに画像情報符号化方法、復号化方法
JP6704932B2 (ja) ビデオコーディングシステムにおける残差変換及び逆変換方法
TW201032600A (en) Image processing device and method
WO2009130886A1 (ja) 動画像符号化装置、撮像装置及び動画像符号化方法
EP1759532A1 (en) Method of storing pictures in a memory using compression coding and cost function including power consumption
US20070133689A1 (en) Low-cost motion estimation apparatus and method thereof
JP5260757B2 (ja) 動画像符号化方法、動画像復号化方法、動画像符号化装置および動画像復号化装置
JP4945513B2 (ja) 可変長復号装置およびそれを用いた動画復号装置
CN103535041A (zh) 图像处理装置和方法
JP2007053554A (ja) 符号化装置、符号化方法、復号装置、復号方法およびプログラム
KR20170125154A (ko) 곡선 화면 내 예측을 사용하는 비디오 복호화 방법 및 장치
US20110110435A1 (en) Multi-standard video decoding system
JP2005223631A (ja) データ処理装置およびその方法と符号化装置および復号装置
JP2008244993A (ja) トランスコーディングのための装置および方法
CN101188765A (zh) 用于对rgb图像编码和解码的方法和设备
WO2011009289A1 (zh) 一种在数字电视接收中实现图像旋转的方法和系统
KR100636911B1 (ko) 색도 신호의 인터리빙 기반 동영상 복호화 방법 및 그 장치
JP2008289105A (ja) 画像処理装置およびそれを搭載した撮像装置
US10334262B2 (en) Moving-picture decoding processing apparatus, moving-picture coding processing apparatus, and operating method of the same
TWI433543B (zh) 高速編碼之無線分散式視訊編解碼系統
WO2005104560A1 (en) Method of processing decoded pictures.

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 10801849

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 10801849

Country of ref document: EP

Kind code of ref document: A1