WO2009052742A1 - An interframe prediction encoding/decoding method and apparatus - Google Patents

An interframe prediction encoding/decoding method and apparatus Download PDF

Info

Publication number
WO2009052742A1
WO2009052742A1 PCT/CN2008/072690 CN2008072690W WO2009052742A1 WO 2009052742 A1 WO2009052742 A1 WO 2009052742A1 CN 2008072690 W CN2008072690 W CN 2008072690W WO 2009052742 A1 WO2009052742 A1 WO 2009052742A1
Authority
WO
WIPO (PCT)
Prior art keywords
motion vector
decoding
template matching
current block
encoding
Prior art date
Application number
PCT/CN2008/072690
Other languages
English (en)
French (fr)
Inventor
Lianhuan Xiong
Steffen Kamp
Michael Evertz
Mathias Wien
Original Assignee
Huawei Technologies Co., Ltd.
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 Huawei Technologies Co., Ltd. filed Critical Huawei Technologies Co., Ltd.
Priority to EP08841850.4A priority Critical patent/EP2202985B1/en
Priority to JP2010529217A priority patent/JP5081305B2/ja
Publication of WO2009052742A1 publication Critical patent/WO2009052742A1/zh
Priority to US12/761,229 priority patent/US20100208814A1/en

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51Motion estimation or motion compensation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/103Selection of coding mode or of prediction mode
    • H04N19/109Selection of coding mode or of prediction mode among a plurality of temporal predictive coding modes
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/146Data rate or code amount at the encoder output
    • H04N19/147Data rate or code amount at the encoder output according to rate distortion criteria
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/176Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51Motion estimation or motion compensation
    • H04N19/513Processing of motion vectors
    • H04N19/517Processing of motion vectors by encoding
    • H04N19/52Processing of motion vectors by encoding by predictive encoding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/70Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards

Definitions

  • the present invention relates to video coding techniques, and in particular, to an interframe prediction coding and decoding method and apparatus.
  • interframe prediction techniques are widely used.
  • the interframe prediction refers to a technique of performing image prediction based on adjacent frames by utilizing temporal redundancy of data, that is, correlation of pixels existing between adjacent frames in a moving image sequence. According to statistics, in the moving images of the two frames before and after, only the brightness value of no more than 10% of the pixels changes more than 2%, and the chromaticity value changes less. The difference between the test direction, pixel precision, and block size. Its main technical points include:
  • motion matching blocks such as 16x16 macroblocks, or further divide 16x16 macroblocks into smaller blocks, such as blocks of size 16x8, 8x16, 8x8, and 4x4.
  • Motion matching search 2.
  • Motion vector precision and pixel interpolation motion estimation includes integer pixel motion estimation and non-integral pixel motion estimation, and non-integral pixel motion estimation includes half-pixel and quarter-pixel motion estimation;
  • Reference frame Can be divided into a forward reference frame and a backward reference frame, or a single reference frame and multiple reference frames; 4, motion vector (MV) prediction, using the encoded motion vector to predict the current motion vector, and then transmit the current vector and the prediction vector The difference.
  • MV motion vector
  • inter-frame prediction technology Based on the inter-frame prediction technology, currently, the major international video coding standards and Coding standards being developed, such as H.264, Digital Audio Video Codec Standard Working Group (AVS), H.264 Layered Video Coding (SVC), and H.264 Multi-View Video Coding (MVC), etc.
  • An inter-frame prediction technique based on motion compensation is proposed, which can greatly improve coding efficiency.
  • motion vector information for motion compensation is obtained by using motion estimation technology, and the motion vector information is written into the coded code stream and transmitted to the decoding end; wherein, the coding
  • the code stream sent by the terminal further includes a macroblock type and residual information; the decoding end performs motion compensation using the motion vector information decoded from the code stream, thereby decoding the image.
  • motion vector information occupies a large portion of the code stream.
  • FIG. 1 is a schematic diagram of an implementation of an existing inter-frame prediction technique based on motion compensation.
  • the input frame is stored in the frame memory through transformation calculation, quantization, inverse quantization, and inverse transformation; at a later time, the system performs motion estimation on the currently input frame according to the previous frame stored in the frame memory, and obtains a motion vector, wherein the motion estimation process can be performed by using any one of the motion estimation algorithms in the prior art; performing motion compensation according to the motion estimation result; performing transformation calculation, quantization, and encoding on the motion compensated result, and outputting Decoding end.
  • inter-frame prediction techniques based on motion compensation are defined in the prior art. For example, in inter-prediction of existing H.264/Advanced Video Coding (AVC), based on motion vector information decoded from a code stream, The decoding end generates a prediction signal corresponding to the position in the reference frame, and obtains the brightness value of the pixel corresponding to the decoded position according to the obtained prediction signal and the residual information carried in the code stream, that is, the transform coefficient information; In the motion vector information of the block, the motion vector prediction of the current block is performed using the motion vector of the block adjacent to the current block to reduce the code stream required to deliver the motion vector information of the current block.
  • AVC Advanced Video Coding
  • FIG. 2 is a schematic diagram of a conventional motion vector prediction method.
  • the motion vector predictor (MVP) of the current block can be derived from the motion vector of the neighboring blocks VIII, B, C, and D.
  • the current block mentioned here can be a macroblock, a block or a partition. According to the segmentation of the current block, the prediction of MVP can be divided into median prediction and non-median prediction.
  • the non-median prediction can be used, specifically: If the segmentation is 8x16 The block, when the left block is the current block, its MVP is equal to MVA (the motion vector of the neighboring block A), when the right block is the current block, its MVP is equal to MVC; if it is divided into 16x8 blocks, then the MVP of the upper block Equal to MVB, the MVP of the lower block is equal to MVA.
  • the median prediction can be used, that is, the MVP of the current block is equal to the median of MVA, MVB, and MVC.
  • MV is the motion vector of the current block estimated by any existing motion estimation algorithm.
  • the MVD is entropy encoded, written into the code stream, and sent to the decoder.
  • a Skip mode is also provided.
  • the code stream corresponding to the mode carries only the macroblock mode information in the macroblock type, and there is no motion vector information and residual information.
  • the decoding end can learn the motion vector of the block adjacent to the macroblock in the current frame according to the macroblock mode information decoded from the received code stream, and then according to the motion vector of the adjacent block.
  • the information is derived from the motion vector of the current block; and the reconstructed value of the current block can be replaced by the predicted value at the corresponding position of the reference frame after determining the corresponding position of the current block in the reference frame according to the motion vector information of the current block.
  • FIG. 4 is a schematic diagram of a conventional template definition manner.
  • an L-shaped area is selected as a template in the upper and left areas of the target area.
  • the template size M is defined as the number of pixels in the horizontal direction from the left side of the target area left boundary covered by the template, and may of course be the number of pixels in the vertical direction from the upper side of the upper boundary of the target area covered by the template.
  • the number of pixels in the template coverage area is 2 x NX M+MXM.
  • templates of other shapes can be used.
  • different weights can be set for different regions in the template to improve the accuracy of the subsequent calculation of the cost function.
  • the execution process of template matching is similar to the matching method in traditional motion estimation, that is, the cost function corresponding to different positions when the template is searched in different regions of the reference frame.
  • the cost function referred to herein may refer to the absolute sum of the pixels in the region of the template region that correspond to the pixels in the region where the search is performed in the reference frame.
  • the cost function here can also be a variance and other cost functions that contain flat constraints on the motion vector field.
  • the above-mentioned template matching search process can use different search ranges or different search methods as needed, for example, a combination of full pixel and half pixel can be used to reduce the complexity of the template matching process.
  • the code stream sent by the encoding end to the decoding end includes macroblock type information of the current block, and may further include residual information; after receiving the code stream, the decoding end matches by template
  • the method finds a motion vector corresponding to the current block, and further finds a corresponding position of the current block in the reference frame, and uses the pixel value at the corresponding position or the pixel value at the corresponding position to add the residual information as the pixel value of the decoded current block.
  • Embodiments of the present invention provide an interframe predictive coding method, which can better save a code rate in an interframe prediction process.
  • the embodiment of the invention provides an inter prediction decoding method, which can better save the code rate in the inter prediction process.
  • Embodiments of the present invention provide an interframe predictive coding apparatus, which can better save a code rate in an interframe prediction process.
  • Embodiments of the present invention provide an inter-prediction decoding apparatus, which can better save a code rate in an inter-frame prediction process.
  • An embodiment of the present invention provides an inter prediction encoding method, including:
  • An embodiment of the present invention further provides an inter prediction decoding method, including:
  • An embodiment of the present invention provides an inter prediction encoding device, including: an obtaining unit and a determining unit;
  • the acquiring unit is configured to acquire a template matching motion vector of the current block and a motion vector predictor
  • the determining unit is configured to compare the template matching motion vector information and the motion vector predictor, determine an encoding manner according to the comparison result, and perform encoding.
  • An embodiment of the present invention further provides an inter prediction decoding apparatus, including: a determining unit and a decoding unit;
  • the determining unit is configured to receive a code stream from the encoding end, compare the obtained template matching motion vector of the current block, and a motion vector prediction value, and determine a decoding mode according to the comparison result;
  • the decoding unit is configured to perform decoding according to the determined decoding manner.
  • An embodiment of the present invention further provides an inter-frame pre-j decoding method, where the method includes: receiving a code stream from an encoding end;
  • the embodiment of the invention has the following advantages:
  • the solution according to the embodiment of the present invention can flexibly select an optimized coding mode that is most suitable for the current situation according to the obtained template matching motion vector and motion vector prediction value information of the current block, thereby achieving the maximum saving rate.
  • FIG. 1 is a schematic diagram of an existing inter-frame prediction technology based on motion compensation
  • FIG. 2 is a schematic diagram of an existing motion vector prediction method
  • FIG. 3 is a schematic diagram of a current current block splitting situation
  • Figure 4 is a schematic diagram of a conventional template definition manner
  • Figure 5 is a flow chart of an embodiment of a method of the present invention.
  • FIG. 6 is a schematic flowchart of coding performed by a coding end according to an embodiment of the present invention
  • FIG. 7 is a schematic flowchart of decoding of a decoding end according to an embodiment of the present invention
  • FIG. 8 is a schematic structural diagram of an embodiment of an encoding apparatus according to the present invention
  • FIG. 9 is a schematic structural diagram of an embodiment of a decoding apparatus according to the present invention.
  • FIG. 10 is a schematic diagram of comparison between coding scheme performance and the prior art in coding performance according to an embodiment of the present invention. detailed description
  • the feature of the motion vector information is generated by using the template matching method, and the motion vector generated in the template matching and the existing traditional motion vector prediction value are used as context information, and the current block coding time code is determined by comparing the two.
  • the embodiments of the present invention can be based on the forward direction of the existing Skip technology Skip (P-Skip) technology implementation.
  • the solution in the embodiment of the present invention extends the existing P-Skip technology, and provides a template matching prediction method based on a conditional coding flag.
  • the macroblock can adaptively change the motion characteristics without requiring special coding motion vector information, and only requires less additional coding.
  • FIG. 5 is a flow chart of an embodiment of a method of the present invention. As shown in FIG. 5, the method includes the following steps: Step 51: The encoding end compares the obtained template matching motion vector (TM_MV) of the current block and the motion vector prediction value, and determines a coding mode for encoding the current block according to the comparison result. ⁇ Coded with a defined encoding.
  • TM_MV template matching motion vector
  • the coding end determines the coding mode and performs coding.
  • the coding end determines the code stream element of the current block coding, performs coding according to the determined code stream element, and sends the coded code stream to the decoding end.
  • the code stream element in this embodiment is a macroblock type and/or whether a template matching flag is used.
  • the manner in which the encoding end determines the code stream element of the current block encoding may be: the encoding end compares the template matching motion vector and the motion vector of the current block. Whether the predicted values are consistent. If they are consistent, the macroblock type information is only encoded in the code stream. If they are inconsistent, the macroblock type is encoded in the code stream and the template matching flag information is used to indicate that the decoder matches the template. Or the motion compensation method decodes the received code stream.
  • the method further includes: the encoding end determining that the decoding end needs to decode the code stream according to the template matching manner or the motion compensation manner, for example: encoding
  • the end-through rate-distortion optimization (RDO) algorithm compares the rate-distortion performance of the codec with the template-matched motion vector and the motion vector predictor, and indicates that the decoder decodes in a manner that the rate-distortion performance is superior.
  • Step 52 The decoding end receives the code stream from the encoding end, compares the obtained template matching motion vector and the motion vector prediction value of the current block, and determines a decoding mode according to the comparison result and performs decoding.
  • the decoding end determines, according to the macroblock type information carried in the received code stream, whether the inter prediction mode for the current block is P_Skip mode; if yes, compares Obtaining whether the template matching motion vector of the current block and the motion vector prediction value are consistent. If they are consistent, the code stream is decoded according to the motion compensation manner; if not, according to whether the template matching flag is used in the code stream. The indication is decoded according to the template matching method or the motion compensation method.
  • the decoding end obtains the matching block information of the current macroblock by decoding the pixels of the reconstructed template in the reference frame of the current slice, and the process of parsing and decoding is mixed. Implementation.
  • step 52 the decoder receives the code stream from the encoding end, presets the template matching motion vector of the current block, or determines template matching block information to obtain a template matching motion vector of the current block, and obtains a template of the current block.
  • the matching motion vector is decoded according to the template matching manner according to whether the identifier in the code stream is used in the template matching flag.
  • the decoding end determines, according to the macroblock type information carried in the received code stream, whether the inter prediction mode for the current block is a P_Skip mode, and if so, and can obtain the current block in the macroblock. If the template matches the motion vector, the template matching motion vector of the current block and the motion vector prediction value are compared. If they are consistent, the code stream is decoded according to the motion compensation mode; if not, the code stream is carried according to the code stream. Whether to use the indication in the template matching flag to decode according to the template matching method or the motion compensation method.
  • decoding is performed according to the template matching manner or the motion compensation manner according to whether the identifier in the code stream is used in the template matching flag, and the parsing and decoding are performed.
  • the process is implemented separately.
  • FIG. 6 is a schematic flowchart of coding performed by an encoding end in an embodiment of a method according to the present invention. As shown in Figure 6, the following steps are included:
  • Step 61 The encoding end uses the template matching method to calculate the template matching motion vector of the current block.
  • the allocation method calculates the template matching motion vector of the current block.
  • the specific calculation method is the same as in the prior art, that is, a pre-selected template area, such as an L-shaped template area, performs a matching search in a reference frame to find a best matching position, and then calculates a motion vector of the current block according to the position, that is, the current
  • the position of the block in the current frame is offset from the position in the reference frame.
  • Step 62 The encoding end predicts a motion vector predictor of the current block.
  • the specific prediction mode is as shown in Figures 2 and 3.
  • the coding end predicts the motion vector predictor of the current block based on the motion vectors of the blocks, B, C, and D adjacent to the current block.
  • the prediction of MVP can be specifically divided into median prediction and non-median prediction.
  • the non-median prediction can be used; in other cases, the median prediction can be used, that is, the MVP of the current block is equal to the median of MVA, MVB, and MVC.
  • Step 63 The encoder determines whether the template matching motion vector is consistent with the motion vector prediction value. If yes, step 64 is performed; otherwise, step 65 is performed.
  • Step 64 The encoding end encodes the current block, and sends the encoded code stream to the decoding end, and ends the process.
  • the encoding end since the template matching motion vector is consistent with the motion vector prediction value, the encoding end does not encode whether to use the template matching flag when encoding.
  • the code stream sent by the encoder to the decoder only includes macroblock type information.
  • Step 65 The encoding end encodes the current block, and sends a code stream that is set to use the template matching flag to the decoding end, and ends the process.
  • the encoding end Since the template matching motion vector is inconsistent with the motion vector prediction value, the encoding end needs to add whether to use the template matching flag information in the encoded code stream. Prior to this, the encoding side needs to first determine whether the decoding end is instructed to decode in accordance with the template matching method or the motion compensation method in whether or not to use the template matching flag. The encoding end determines whether the decoding end indicates whether the decoding end decodes according to the template matching manner or the motion compensation manner in the template matching flag: the encoding end encodes and decodes the template matching motion vector and the motion vector prediction value respectively, and optimizes the rate distortion. The algorithm compares which of the two codecs performs better. For example, it compares which way the reconstructed image after encoding has the smallest deviation from the original image, and determines whether to use the specific setting of the template matching flag according to the comparison result.
  • one bit in the code stream can be set to use a template. Match the flag bit. If the encoding and decoding method according to the template matching motion vector performs better, the template matching flag bit is set to 1; if the encoding and decoding method according to the motion vector prediction value performs well, whether to use the template matching The flag is set to 0.
  • the code stream sent by the encoding end to the decoding end includes a macroblock type and whether template matching flag information is used.
  • FIG. 7 is a schematic flowchart of decoding performed by a decoding end in an embodiment of a method according to the present invention. As shown in Figure 7, the following steps are included:
  • Step 71 The decoding end decodes the received code stream of the current block, and determines whether the code stream is in the P_Skip mode. If yes, step 72 is performed; otherwise, step 73 is performed.
  • the decoding end determines whether the inter-frame prediction mode corresponding to the code stream is P_Skip mode according to the macroblock type information carried in the code stream.
  • the macroblock type information may be set in a manner specified by a person to indicate that the inter-frame prediction mode corresponding to the code stream is P_Skip mode.
  • Step 72 The decoder calculates a template matching motion vector for the current block, and derives a motion vector predictor, and then performs step 74.
  • the template matching motion vector and the motion vector prediction value are obtained in the same manner as in the prior art, and are not described again.
  • the predicted value of the obtained motion vector can be used to search within the area set around it to speed up the search process.
  • Step 73 The decoding end decodes the current block according to other existing methods. Since it is not related to the present invention, the specific decoding process is not described again.
  • Step 74 The decoding end determines whether the template matching motion vector of the current block is consistent with the motion vector prediction value. If yes, step 75 is performed; otherwise, step 76 is performed.
  • Step 75 The decoding end directly uses the motion vector predictor to decode, and ends the process. If the template matching motion vector is consistent with the motion vector prediction value, then no template matching flag is used in the code stream, so the decoding end can directly decode the current vector's shipping vector prediction value.
  • the specific decoding mode is the same as the prior art, and will not be described again.
  • Step 76 The decoding end decodes according to the indication according to whether the template matching flag is used in the decoded code stream, and ends the process.
  • the decoding end selects a corresponding motion vector according to the indication of whether the template matching flag bit is used in the decoded code stream, and performs subsequent decoding, and ends the process.
  • the decoder decodes according to the template matching method. If the template matching flag is set to 0, the decoding end uses the motion vector prediction value to decode according to the motion compensation method.
  • the specific decoding process is prior art and will not be described again.
  • step 1 of the embodiment is the same as step 71 for the decoding end.
  • the decoding end determines whether the inter-frame prediction mode corresponding to the code stream is P_Skip mode according to the macroblock type information carried in the code stream.
  • the macroblock type information may be set in a manner specified by a person to indicate that the inter-frame prediction mode corresponding to the code stream is P_Skip mode.
  • Step 2 The decoder calculates a template matching motion vector for the current block, and derives a motion vector prediction value. If the template matching motion vector of the current block of the macro block is not available, step 3 is performed, otherwise step 4 is performed.
  • Step 4 - Step 6 Same as Step 74 - Step 76.
  • step 6 the decoding end selects the corresponding motion vector according to the indication of whether the template matching flag bit is used in the decoded code stream, and performs subsequent decoding, and ends the process.
  • the decoding end decodes according to the template matching manner.
  • the template matching motion vector is set to 0 in advance or the template information required for the original decoding is equal to the previous decoding template information, and is used.
  • the matching block information is found to obtain the template matching motion vector of the current block; if the template matching flag is set to 0, the decoding end uses the motion vector prediction value to perform decoding according to the motion compensation mode.
  • the specific decoding process is prior art and will not be described again.
  • Modified, modified P-Skip mode syntax is as follows:
  • the function UserTMVector() is used to perform template matching operations. The function returns true if the result of the template matching motion vector and motion vector prediction for the current block is inconsistent, otherwise it returns false.
  • the descriptions in the embodiments shown in FIG. 6 and FIG. 7 only refer to the single color component signal, that is, the luminance value.
  • the solution described in the embodiments of the invention can equally be applied to multi-component signals. However, if different components are used at different sampling rates, some restrictions may need to be met. For example, for a ratio of Y, Cb, and Cr of 4:2:0, where Y is the luminance component, Cb is the blue chrominance component, and Cr is the red chrominance component, the template region size, and the target region. The size needs to be doubled.
  • the solution described in the embodiment of the present invention is not only applicable to the single-view video coding environment shown in FIG. 6 and FIG. 7, but is also applicable to the multi-view video coding environment.
  • the motion vector generated by template matching can be used as the motion vector prediction value of the current block, and the motion vector generated by the template matching technique can be used for the subsequent current block.
  • motion vector prediction it plays a role in rate distortion optimization and MVD computational coding process, which provides more motion vector prediction applicability for the current block and improves the accuracy of motion vector prediction.
  • the template matching motion vector and the motion vector prediction value in the above embodiments shown in Figs. 6 and 7 can also be used as context information to select a context model when encoding and decoding the current block.
  • the probability relationship between the template matching motion vector and the motion vector prediction value can be determined to determine the CABAC coding probability of the current block information.
  • the specific process of coding is as follows:
  • CABAC coding is a prior art, those skilled in the art should be aware of the implementation process, and therefore will not be described again here.
  • FIG. 8 is a schematic diagram showing the structure of an embodiment of an encoding apparatus according to the present invention.
  • the device includes: an obtaining unit 81 and a determining unit 82;
  • An obtaining unit 81 configured to acquire a template matching motion vector of the current block and a motion vector predictor
  • the determining unit 82 is configured to compare the template matching motion vector information and the motion vector prediction value, and determine the encoding mode according to the comparison result and perform encoding.
  • the determining unit 82 specifically includes: a comparing subunit 821 and an encoding subunit 822;
  • the comparison subunit 821 is configured to compare the template matching motion vector of the current block and the motion vector prediction value, and send the comparison result to the encoding subunit 822;
  • the encoding sub-unit 822 is configured to perform encoding according to the comparison result. If the comparison result is consistent, the macroblock type information is only encoded in the code stream; if not, the macroblock type is encoded in the code stream and whether the template matching flag is used. Instructing the decoding end to decode the code stream according to a template matching manner or a motion compensation manner.
  • the unit may be further configured to: if the comparison result is inconsistent, compare the different performances of the template matching motion vector and the motion vector prediction value for encoding and decoding, and instruct the decoding end to perform decoding according to a better performance.
  • the determining unit 82 may further include: a selecting subunit 823, configured to select a model for encoding the current block according to the template matching motion vector information of the current block and the comparison result of the motion vector predictor, and notify the encoding subunit 822. .
  • FIG. 9 is a schematic structural diagram of an embodiment of a decoding apparatus according to the present invention. As shown in FIG. 9, the apparatus includes: a determining unit 91 and a decoding unit 92;
  • a determining unit 91 configured to receive a code stream from the encoding end, compare the obtained template matching motion vector of the current block and the motion vector prediction value, and determine a decoding mode according to the comparison result;
  • the decoding unit 92 is configured to perform decoding according to the determined decoding manner.
  • the determining unit 91 may specifically include: a comparing subunit 912, configured to receive the code stream from the encoding end, compare whether the template matching motion vector of the current block is consistent with the motion vector prediction value, and if yes, notify the decoding unit 92 to compensate according to the motion.
  • the method performs decoding on the code stream; if not, according to whether the indication in the template matching flag is carried in the code stream, the decoding unit 92 is notified to perform decoding according to a template matching manner or a motion compensation manner.
  • the determining unit 91 may further include: a mode determining subunit 911 and/or a selecting subunit 913:
  • the mode determining subunit 911 is configured to determine, according to the macroblock type information carried in the received code stream, that the inter prediction mode for the current block is P_Skip mode, and notify the comparing subunit to perform 912 to perform its own function.
  • the selecting subunit 913 is configured to select a model for decoding the code stream according to the template matching motion vector of the current block and the motion vector predictor comparison result, and notify the decoding unit 92.
  • the template matching motion vector and the motion vector prediction value are context information, and the information contained in the current code stream is determined by comparison and encoded and decoded. Since the solution according to the embodiment of the present invention is implemented based on the P_Skip mode, the code stream is not additionally added, the transmission cost used for transmitting the motion vector can be saved, and more options for motion vector prediction are provided, and The information available at the decoding end can provide context decision information such as entropy coding for the current codec, enhance the adaptability of the technology, and improve the coding efficiency.
  • FIG. 10 is a schematic diagram of comparison of coding performance in comparison with the prior art according to the embodiment of the present invention.
  • the sequence coding type is IPPP, that is, the first frame is coded according to intra prediction, and the remaining frames are coded according to inter prediction.
  • JM12.3 is the technical effect of using the existing motion compensation method on the H.264/AVC platform
  • TM Skip is the technical effect of using the scheme of the present invention on the H.264/AVC platform
  • TM 16x16 8x8 is ⁇ Technical effect with existing template matching method
  • TM Skip 16x16 8x8 is in ⁇
  • the technical effects after combining the solutions of the present invention are based on the existing template matching method.
  • the abscissa in 10 represents the code rate, and the ordinate represents the signal-to-noise ratio (PSNR), which represents the difference between the reconstructed image and the original image.
  • PSNR signal-to-noise ratio
  • the present invention can be implemented by hardware or by software plus a necessary general hardware platform. Based on such understanding, the technical solution of the present invention can be embodied in the form of a software product that can be stored in a non-volatile storage medium.
  • a computer device (may be a personal computer, server, or network device, etc.) to perform the methods described in various embodiments of the present invention.

Landscapes

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

Description

一种帧间预测编解码方法及装置 本申请要求于 2007年 10 月 15 日提交中国专利局、 申请号为 "200710181831.3"、 发明名称为 "一种帧间预测编解码方法及装置" 的中国专利申请的优先权,以及于 2008年 1月 8日提交中国专利局、 申请号为 "200810002875.X"、 发明名称为 "一种帧间预测编解码方 法及装置"的中国专利申请的优先权, 其全部内容通过引用结合在本 申请中。 技术领域
本发明涉及视频编码技术,特别涉及一种帧间预测编解码方法及 装置。
背景技术
在基于混合编码技术的视频编码中,帧间预测技术得到了广泛应 用。 所谓帧间预测是指利用数据的时间冗余, 即活动图像序列中相邻 帧之间的像素存在的相关性, 根据相邻帧进行图像预测的技术。根据 统计, 在前后两帧运动图像中, 只有不超过 10 %的像素的亮度值发 生的改变超过 2 % , 色度值改变则更少。 考方向、 像素精度以及块大小划分的不同上。 其主要技术点包括:
1、 釆用不同大小的匹配块, 如 16x 16的宏块, 或者, 将 16x 16 的宏块进一步划分为更小的块, 如大小为 16x8、 8x 16、 8x8以及 4x4 等的块, 以进行运动匹配搜索; 2、 运动向量精度和像素插值, 运动 估计包括整像素运动估计和非整像素运动估计,非整像素运动估计又 包括半像素和四分之一像素运动估计; 3、 参考帧, 可分为前向参考 帧和后向参考帧, 或者单参考帧和多参考帧; 4、 运动矢量(MV )预 测, 利用已编码的运动矢量来预测当前运动矢量, 然后传输当前矢量 和预测矢量的差值。
在帧间预测技术的基础上, 目前, 国际上主要的视频编码标准和 正在制定的编码标准, 如 H.264、 数字音视频编解码技术标准工作组 ( AVS )、 H.264 可分层视频编码(SVC ) 以及 H.264 多视角视频编 码(MVC )等中, 均提出了基于运动补偿的帧间预测技术, 该技术 能够极大地提高编码效率。
在具有运动补偿的帧间预测中, 在编码端, 利用运动估计技术, 得到用于运动补偿的运动矢量信息,并将该运动矢量信息写入到编码 码流中传送给解码端; 其中, 编码端发送的码流中还将进一步包括宏 块类型以及残差信息;解码端利用从码流中解码出来的运动矢量信息 进行运动补偿, 从而解码图像。 在利用帧间预测技术编码的图像码流 中, 运动矢量信息占用了很大一部分码流。
图 1为现有基于运动补偿的帧间预测技术实现示意图。如图 1所 示, 输入帧经过变换计算、 量化、 逆量化以及逆变换等过程, 存储到 帧存储器中; 随后时刻, 系统根据帧存储器存储的前一帧对当前输入 的帧进行运动估计, 得到运动矢量, 其中, 运动估计过程可釆用现有 技术中的任何一种运动估计算法进行;根据运动估计结果进行运动补 偿; 对运动补偿后的结果进行变换计算、 量化以及编码等过程, 输出 到解码端。
现有技术中定义了多种基于运动补偿的帧间预测技术, 比如, 在 现有 H.264/高级视频编码(AVC )的帧间预测中, 根据从码流中解码 出的运动矢量信息, 解码端生成参考帧中对应位置的预测信号, 根据 该得到的预测信号以及码流中携带的残差信息 ,即变换系数信息得到 解码后对应位置的像素的亮度值; 在编码端, 在编码当前块的运动矢 量信息时,使用与当前块相邻的块的运动矢量进行当前块的运动矢量 预测, 以减少传递当前块的运动矢量信息所需的码流。
图 2为现有运动矢量预测方式示意图。 其中, 当前块的运动矢量 预测值(MVP ), 可才艮据与其相邻的块八、 B、 C、 D的运动矢量推导 而出。 这里所提到的当前块可以是宏块、 块或分割。 根据当前块的分 割情况, MVP 的预测可分为中值预测和非中值预测。 当前块被分割 为图 3所示情况时,可釆用非中值预测,具体来说: 如果分割为 8x16 的块, 当左边块为当前块时, 其 MVP等于 MVA (相邻块 A的运动 矢量), 当右边块为当前块时, 其 MVP等于 MVC; 如果分割为 16x8 的块, 那么上边块的 MVP等于 MVB, 下边块的 MVP等于 MVA。 其余情况可釆用中值预测, 即当前块的 MVP等于 MVA、 MVB 和 MVC的中值。
得到当前块的 MVP后, 可进一步计算运动矢量差(MVD ), 即, MVD = MV-MVP。 其中, MV 为釆用现有任一运动估计算法估计出 的当前块的运动矢量。 之后, 对 MVD进行熵编码, 写入码流, 发送 到解码端。
上述方式虽然能够实现基于运动补偿的帧间预测,但该方式中必 须将运动矢量信息显式的写入到码流中进而传递给解码端,这会造成 额外的码率增加。
在现有基于运动补偿的帧间预测技术中,除了利用运动矢量预测 技术提高编码效率外, 还提供了一种 Skip (跳过)模式。 该模式对应 的编码码流中只携带有宏块类型中的宏块模式信息,而没有运动矢量 信息和残差信息。 这种情况下, 解码端可以根据从接收到的码流中解 码出的宏块模式信息, 获知在当前帧中与该宏块相邻的块的运动矢 量, 才艮据相邻块的运动矢量信息推导出当前块的运动矢量; 而当前块 的重建值则可通过根据当前块的运动矢量信息确定出该当前块在参 考帧中的对应位置后, 用参考帧对应位置上的预测值进行代替。
现有技术中还提出了一种基于模板匹配技术的帧间预测方式。模 板匹配技术用于为 N X N像素的目标区域, 即当前块推导预测信号。 由于目标区域还没有重建,所以可以在目标区域已重建的邻近区域中 定义一个模板。 图 4为现有模板定义方式示意图。 通常情况下, 在目 标区域的上边和左边区域选择一个 L形区域作为模板。 模板尺寸 M 定义为从目标区域左边界被模板覆盖的左边开始的水平方向上的像 素数, 当然, 也可以是从目标区域的上边界被模板覆盖的上边开始的 垂直方向上的像素数。 由图 4可知, 模板覆盖区域的像素数为 2 x N X M+M X M。 在实际应用中, 除 L形以外, 还可以釆用其他形状的模板。 同时 还可以为模板中的不同区域设置不同的权重,以用于提高后续计算代 价函数时的准确度。
模板匹配的执行过程与传统运动估计中的匹配方式类似,即计算 模板在参考帧的不同区域搜索时对应不同位置的代价函数。这里所提 到的代价函数可以是指位于模板区域内的像素与其在参考帧中进行 匹配搜索时的区域内对应的像素的绝对和。 当然, 这里的代价函数也 可以是方差以及包含对运动矢量场有平坦约束的其他代价函数。上述 模板匹配的搜索过程可以根据需要釆用不同的搜索范围或不同的搜 索方式, 比如可以釆用全像素和半像素相结合的搜索方式, 以降低模 板匹配过程的复杂度。
在模板匹配帧间预测方式中,编码端发送到解码端的码流中包括 当前块的宏块类型信息, 此外, 还可进一步包括残差信息; 解码端在 接收到该码流后, 通过模板匹配方式找到当前块对应的运动矢量, 进 而在参考帧中找到当前块的对应位置,并用对应位置上的像素值或对 应位置上的像素值加上残差信息作为解码后当前块的像素值。
当釆用模板匹配方式进行编码时 ,需要增加一个宏块类型来标识 当前块编码是釆用现有运动补偿方式还是釆用模板匹配方式进行编 码。 但是, 釆用这种方式后, 由于要引入特定的宏块类型来传递模板 匹配方式的标识信息, 因此, 造成额外的码率增加。
在实施本发明的过程中,发明人发现现有技术中在编码过程中需 要加入额外码流信息, 造成码率不同程度的增加。
发明内容
本发明实施例提供一种帧间预测编码方法,能够在帧间预测过程 中较好地节省码率。
本发明实施例提供一种帧间预测解码方法,能够在帧间预测过程 中较好地节省码率。 本发明实施例提供一种帧间预测编码装置,能够在帧间预测过程 中较好地节省码率。
本发明实施例提供一种帧间预测解码装置,能够在帧间预测过程 中较好地节省码率。
本发明实施例的技术方案是这样实现的:
本发明实施例提供一种帧间预测编码方法, 包括:
获取当前块的模板匹配运动矢量以及运动矢量预测值; 比较所述模板匹配运动矢量以及所述运动矢量预测值,根据比较 结果确定编码方式并进行编码。
本发明实施例还提供一种帧间预测解码方法, 包括:
接收来自编码端的码流;
获取当前块的模板匹配运动矢量以及运动矢量预测值; 比较所述模板匹配运动矢量以及所述运动矢量预测值,根据比较 结果确定解码方式并进行解码。
本发明实施例提供一种帧间预测编码装置, 包括: 获取单元以及 确定单元;
所述获取单元,用于获取当前块的模板匹配运动矢量以及运动矢 量预测值;
所述确定单元,用于比较所述模板匹配运动矢量信息以及运动矢 量预测值, 根据比较结果确定编码方式并进行编码。
本发明实施例还提供一种帧间预测解码装置, 包括: 确定单元以 及解码单元;
所述确定单元, 用于接收来自编码端的码流, 比较获取到的当前 块的模板匹配运动矢量以及运动矢量预测值,根据比较结果确定解码 方式;
所述解码单元, 用于按照所述确定的解码方式进行解码。
本发明实施例还提供一种帧间预 ' j解码方法, 该方法包括: 接收来自编码端的码流;
预设当前块的模板匹配运动矢量或确定模板匹配块信息来得到 当前块的模板匹配运动矢量;
获取所述当前块的模板匹配运动矢量,根据所述码流中携带的是 否使用模板匹配标志的指示, 按照模板匹配方式进行解码。
与现有技术相比, 本发明实施例具有如下优点:
本发明实施例所述方案能够根据获得的当前块的模板匹配运动 矢量以及运动矢量预测值信息,灵活选择最适合当前情况的优化编解 码方式, 从而达到最大程度上的节省码率的目的。 附图说明
图 1为现有基于运动补偿的帧间预测技术实现示意图; 图 2为现有运动矢量预测方式示意图;
图 3为现有当前块分割情况示意图;
图 4为现有模板定义方式示意图;
图 5为本发明方法实施例的流程图;
图 6为本发明方法实施例中编码端进行编码的流程示意图; 图 7为本发明方法实施例中解码端进行解码的流程示意图; 图 8为本发明编码装置实施例的组成结构示意图;
图 9为本发明解码装置实施例的组成结构示意图;
图 10为本发明实施例所述方案与现有技术在编码性能方面的比 较示意图。 具体实施方式
为使本发明的目的、技术方案及优点更加清楚明白, 以下参照附 图并举实施例, 对本发明作进一步地详细说明。
本发明实施方式中, 利用模板匹配方式生成运动矢量信息的特 点,以模板匹配中生成的运动矢量和现有传统运动矢量预测值作为上 下文信息,通过对两者进行比较来确定当前块编码时码流中所包含的 语法元素信息。 本发明所述实施方式可基于现有 Skip技术中的前向 跳过(P-Skip )技术实现。
本发明实施例所述方案对现有 P-Skip技术进行扩展, 提供了一 种基于条件编码标志的模板匹配预测方法。 本发明的实施方式中, 宏 块能够自适应地改变运动特性, 不需要专门的编码运动矢量信息, 只 需要艮少的额外编码即可。
图 5为本发明方法实施例的流程图。如图 5所示,包括以下步骤: 步骤 51 : 编码端对获取到的当前块的模板匹配运动矢量 ( TM_MV ) 以及运动矢量预测值进行比较, 根据比较结果确定为当 前块进行编码的编码方式并釆用确定的编码方式进行编码。
本步骤中, 编码端确定编码方式并进行编码的方式可以是: 编码 端确定为当前块编码的码流元素, 按照确定的码流元素进行编码, 并 将编码后的码流发送到解码端。本实施例中的码流元素为宏块类型和 /或是否使用模板匹配标志; 编码端确定为当前块编码的码流元素的 方式可以为:编码端比较当前块的模板匹配运动矢量与运动矢量预测 值是否一致, 如果一致, 则只在码流中编入宏块类型信息; 如果不一 致, 则在码流中编入宏块类型以及是否使用模板匹配标志信息, 以指 示解码端按照模板匹配方式或运动补偿方式对接收到的码流进行解 码。
其中,在向码流中编入宏块类型以及是否使用模板匹配标志信息 之前, 进一步包括: 编码端确定需要指示解码端按照模板匹配方式还 是运动补偿方式对所述码流进行解码, 比如: 编码端通过率失真优化 ( RDO )算法比较釆用模板匹配运动矢量以及运动矢量预测值进行编 解码的率失真性能, 指示解码端按照率失真性能较优的方式进行解 码。
步骤 52: 解码端接收来自编码端的码流, 对获取到的当前块的 模板匹配运动矢量以及运动矢量预测值进行比较,根据比较结果确定 解码方式并进行解码。
本步骤中, 解码端根据接收到的码流中携带的宏块类型信息, 确 定针对当前块的帧间预测方式是否为 P— Skip模式; 如果是, 则比较 获取到的当前块的模板匹配运动矢量以及运动矢量预测值是否一致, 如果一致, 则按照运动补偿方式对所述码流进行解码; 如果不一致, 则根据码流中携带的是否使用模板匹配标志中的指示,按照模板匹配 方式或运动补偿方式进行解码。 解码端在解析当前宏块的边界信息 时,通过解码重构模板的像素在当前片的参考帧中寻获匹配块信息来 获取到当前宏块的模板匹配运动矢量, 其解析和解码的过程混合实 施。
另一实施方式中, 步骤 52: 解码端接收来自编码端的码流, 预 设当前块的模板匹配运动矢量或确定模板匹配块信息来得到当前块 的模板匹配运动矢量, 获取所述当前块的模板匹配运动矢量, 根据所 述码流中携带的是否使用模板匹配标志中的指示,按照模板匹配方式 进行解码。
本步骤中, 解码端根据接收到的码流中携带的宏块类型信息, 确 定针对当前块的帧间预测方式是否为 P— Skip模式, 如果是, 并且可 以得到所述宏块中当前块的模板匹配运动矢量,则比较获取到的当前 块的模板匹配运动矢量以及运动矢量预测值是否一致, 如果一致, 则 按照运动补偿方式对所述码流进行解码; 如果不一致, 则根据码流中 携带的是否使用模板匹配标志中的指示,按照模板匹配方式或运动补 偿方式进行解码。若不可以得到所述宏块中当前块的模板匹配运动矢 量, 则根据码流中携带的是否使用模板匹配标志中的指示, 按照模板 匹配方式或运动补偿方式进行解码, 所述解析和解码的过程分开实 施。
下面通过具体实施例,从编码以及解码两端分别对本发明所述方 案作进一步地详细说明:
图 6为本发明方法实施例中编码端进行编码的流程示意图。如图 6所示, 包括以下步骤:
步骤 61 : 编码端利用模板匹配方式计算当前块的模板匹配运动 矢量。
本步骤中, 编码端在需要对当前块进行编码时, 首先通过模板匹 配方式计算当前块的模板匹配运动矢量。具体计算方式与现有技术中 相同, 即将预先选择的模板区域, 比如 L形模板区域在参考帧中进行 匹配搜索, 找到最佳匹配位置, 才艮据该位置计算当前块的运动矢量, 即当前块在当前帧中的位置与在参考帧中的位置偏移。
步骤 62: 编码端预测当前块的运动矢量预测值。
具体预测方式如图 2和 3所示,编码端根据与当前块相邻的块 、 B、 C、 D 的运动矢量预测当前块的运动矢量预测值。 才艮据当前块的 分割情况, MVP 的预测可具体分为中值预测和非中值预测。 当当前 块被分割为图 3所示情况时, 可釆用非中值预测; 其余情况可釆用中 值预测 , 即当前块的 MVP等于 MVA、 MVB和 MVC的中值。
步骤 63: 编码端判断模板匹配运动矢量与运动矢量预测值是否 一致, 如果一致, 则执行步骤 64; 否则, 执行步骤 65。
步骤 64: 编码端针对当前块进行编码, 并将编码后的码流发送 至解码端, 结束流程。
本步骤中,由于模板匹配运动矢量与运动矢量预测值一致,所以, 编码端在编码时不会编入是否使用模板匹配的标志。 这样, 编码端发 送到解码端的码流中只包括宏块类型信息。
步骤 65: 编码端针对当前块进行编码, 并将设置了是否使用模 板匹配标志的码流发送至解码端, 结束流程。
由于模板匹配运动矢量与运动矢量预测值不一致, 所以, 编码端 需要在编码的码流中加入是否使用模板匹配标志信息。 在这之前, 编 码端需要首先确定,在是否使用模板匹配标志中指示解码端按照模板 匹配方式还是运动补偿方式进行解码。 其中, 编码端确定在是否使用 模板匹配标志中指示解码端按照模板匹配方式还是运动补偿方式进 行解码的方式为:编码端对模板匹配运动矢量以及运动矢量预测值分 别进行编解码, 通过率失真优化算法比较两次编解码中哪次性能更 好, 比如, 比较釆用哪种方式编码后的重建图像与原始图像的偏差最 小, 根据比较结果确定该是否使用模板匹配标志的具体设置。
在具体实现中,可以将码流中的一个比特位设置为是否使用模板 匹配标志位。 如果根据模板匹配运动矢量进行编解码的方式性能较 好, 则将该是否使用模板匹配标志位设置为 1 ; 如果根据运动矢量预 测值进行编解码的方式性能较好,则将该是否使用模板匹配标志设置 为 0。
本步骤中,编码端发送到解码端的码流中包括宏块类型以及是否 使用模板匹配标志信息。
后续过程中, 解码端即可根据接收到的码流进行解码。 图 7为本 发明方法实施例中解码端进行解码的流程示意图。 如图 7所示, 包括 以下步骤:
步骤 71 : 解码端对接收到的当前块的码流进行解码, 判断该码 流是否为 P— Skip模式, 如果是, 则执行步骤 72, 否则, 执行步骤 73。
本步骤中,解码端才艮据码流中携带的宏块类型信息确定该码流对 应的帧间预测方式是否为 P— Skip模式。 其中, 宏块类型信息可以通 过人为规定的方式 ,设置特定的标识来表示该码流对应的帧间预测方 式为 P— Skip模式。
步骤 72: 解码端计算针对当前块的模板匹配运动矢量, 并推导 运动矢量预测值, 然后执行步骤 74。
模板匹配运动矢量与运动矢量预测值的获取方式与现有技术中 相同, 不再赘述。
其中, 计算模板匹配运动矢量的过程中, 可以以获取到的运动矢 量预测值为中心,在其周围设定的区域内进行搜索,以加快搜索过程。
步骤 73 : 解码端按照其它现有方式对当前块进行解码, 由于与 本发明无关, 所以具体解码过程不再赘述。
步骤 74: 解码端判断当前块的模板匹配运动矢量与运动矢量预 测值是否一致, 如果一致, 则执行步骤 75; 否则, 执行步骤 76。
步骤 75: 解码端直接利用运动矢量预测值进行解码, 结束流程。 如果模板匹配运动矢量与运动矢量预测值一致,那么码流中没有 是否使用模板匹配标志,所以解码端可直接利用当前块的运送矢量预 测值进行解码。 具体解码方式与现有技术相同, 不再赘述。 步骤 76: 解码端根据解码出的码流中的是否使用模板匹配标志 的指示, 按照指示的方式进行解码, 结束流程。
本步骤中,解码端根据解码出的码流中的是否使用模板匹配标志 位的指示, 选择对应的运动矢量, 并进行后续解码, 结束流程。
比如, 如果是否使用模板匹配标志位设置为 1 , 则解码端按照模 板匹配方式进行解码; 如果是否使用模板匹配标志位设置为 0, 则解 码端利用运动矢量预测值, 按照运动补偿方式进行解码。 具体解码过 程为现有技术, 不再赘述。
需要说明的是, 本实施例中, 由于编码端发送的码流中不包括残 差信息, 所以后续过程中解码端进行解码时, 只需要将参考帧中与当 前块对应位置的像素值作为当前块解码后的像素值。
在另一实施例中, 对于解码端, 所述实施方式步骤 1同步骤 71。 本步骤中,解码端才艮据码流中携带的宏块类型信息确定该码流对 应的帧间预测方式是否为 P— Skip模式。 其中, 宏块类型信息可以通 过人为规定的方式 ,设置特定的标识来表示该码流对应的帧间预测方 式为 P— Skip模式。
步骤 2: 解码端计算针对当前块的模板匹配运动矢量, 并推导运 动矢量预测值, 如果所述宏块的当前块的模板匹配运动矢量不可获 得, 执行步骤 3 , 否则执行步骤 4。
步骤 4 -步骤 6同步骤 74 -步骤 76。
步骤 6中,解码端根据解码出的码流中的是否使用模板匹配标志 位的指示, 选择对应的运动矢量, 并进行后续解码, 结束流程。
比如, 如果是否使用模板匹配标志位设置为 1 , 则解码端按照模 板匹配方式进行解码,此时模板匹配运动矢量预先设 0或者原先通过 解码所需的模板信息等于前一个解码模板信息,用来寻获匹配块信息 来获取到当前块的模板匹配运动矢量;如果是否使用模板匹配标志位 设置为 0, 则解码端利用运动矢量预测值, 按照运动补偿方式进行解 码。 具体解码过程为现有技术, 不再赘述。
为实现图 6和 7所示流程, 需要对现有 P— Skip模式的语法进行 修改, 修改后的 P— Skip模式语法如下:
Figure imgf000014_0001
函数 UserTMVector()用于执行模板匹配操作。如果针对当前块的 模板匹配运动矢量和运动矢量预测值的结果不一致,则函数返回 true , 否则返回 false。
需要说明的是,图 6和图 7所示实施例中的描述只涉及到单颜色 分量信号, 即亮度值, 在实际应用中, 通过扩展目标和模板区域到所 有分量的相同位置的像素,本发明实施例所述方案可同样运用到多分 量信号上。 但是, 如果不同的分量釆用不同的釆样率, 可能需要满足 一些限制条件。 比如, 对于 Y、 Cb、 Cr之比为 4: 2: 0的颜色子 釆样, 其中 Y是指亮度分量, Cb指蓝色色度分量, 而 Cr指红色色度 分量, 模板区域大小和目标区域的大小需要成 2倍关系。 而且, 本发 明实施例所述方案不仅仅适用于图 6和图 7所示的单视角视频编码环 境中, 也同样适用于多视角视频编码环境中。
另外, 上述图 6和图 7所示实施例中, 利用模板匹配生成的运动 矢量可作为当前块的运动矢量预测值,即可将本次利用模板匹配技术 生成的运动矢量用于后续当前块的运动矢量预测中,在率失真优化和 MVD计算编码过程中起到作用, 从而为当前块提供更多的运动矢量 预测适用性, 提高运动矢量预测的准确程度。
上述图 6和图 7所示实施例中的模板匹配运动矢量和运动矢量预 测值还可以作为上下文信息,来选择对当前块进行编解码时的上下文 模型。 比如, 对于基于上下文的算数编码 CABAC来说, 可以利用模 板匹配运动矢量和运动矢量预测值之间的比较关系,决定对当前块信 息进行 CABAC编码的概率模型。 其中, 编码的具体过程如下:
计算当前块的模板匹配运动矢量并获取运动矢量预测值; 将模板匹配运动矢量与运动矢量预测值进行比较, 当两者一致 时, 选择概率模型 P1对当前块进行编码; 如果不一致, 选择概率模 型 P2进行编码。
解码的具体过程如下:
计算当前块的模板匹配运动矢量并获取运动矢量预测值; 将模板匹配运动矢量与运动矢量预测值进行比较, 当两者一致 时, 选择概率模型 PI对当前块进行解码; 如果不一致, 选择概率模 型 P2进行解码。
由于 CABAC编码为现有技术,本领域技术人员应该清楚其实现 过程, 故此处不再赘述。
基于上述方法, 图 8 为本发明编码装置实施例的组成结构示意 图。 如图 8所示, 该装置包括: 获取单元 81以及确定单元 82;
获取单元 81 , 用于获取当前块的模板匹配运动矢量以及运动矢 量预测值;
确定单元 82 , 用于比较模板匹配运动矢量信息以及运动矢量预 测值是否一致, 根据比较结果确定编码方式并进行编码。
其中, 确定单元 82中具体包括: 比较子单元 821以及编码子单 元 822;
比较子单元 821 , 用于比较当前块的模板匹配运动矢量以及运动 矢量预测值是否一致, 并将比较结果发送给编码子单元 822;
编码子单元 822, 用于根据比较结果进行编码, 如果比较结果一 致, 则只在码流中编入宏块类型信息; 如果不一致, 则在码流中编入 宏块类型以及是否使用模板匹配标志,以指示解码端按照模板匹配方 式或运动补偿方式对码流进行解码。 该单元还可进一步用于, 如果比 较结果不一致,比较釆用模板匹配运动矢量以及运动矢量预测值进行 编解码的不同性能, 指示解码端按照性能较优的方式进行解码。
上述确定单元 82 中还可进一步包括: 选择子单元 823 , 用于根 据当前块的模板匹配运动矢量信息以及运动矢量预测值的比较结果, 选择对当前块进行编码的模型, 并通知编码子单元 822。
图 9为本发明解码装置实施例的组成结构示意图。 如图 9所示, 该装置包括: 确定单元 91以及解码单元 92;
确定单元 91 , 用于接收来自编码端的码流, 比较获取到的当前 块的模板匹配运动矢量以及运动矢量预测值是否一致,根据比较结果 确定釆用的解码方式;
解码单元 92 , 用于按照确定的解码方式进行解码。 其中, 确定单元 91可具体包括: 比较子单元 912, 用于接收来 自编码端的码流,比较当前块的模板匹配运动矢量与运动矢量预测值 是否一致, 如果一致, 则通知解码单元 92按照运动补偿方式对该码 流进行解码; 如果不一致, 则根据该码流中携带的是否使用模板匹配 标志中的指示, 通知解码单元 92按照模板匹配方式或运动补偿方式 进行解码。
此外, 确定单元 91中还可进一步包括: 模式确定子单元 911和 / 或选择子单元 913:
模式确定子单元 911; 用于根据接收到的码流中携带的宏块类型 信息, 确定针对当前块的帧间预测方式为 P— Skip模式后, 通知比较 子单元执行 912执行其自身功能。
选择子单元 913 , 用于根据当前块的模板匹配运动矢量以及运动 矢量预测值比较结果, 选择对该码流进行解码的模型, 并通知解码单 元 92。
图 8和图 9所示实施例的具体工作流程请参照方法相应部分的说 明, 此处不再赘述。
可见, 釆用本发明实施例的技术方案, 以模板匹配运动矢量和运 动矢量预测值为上下文信息,通过比较来确定当前码流中包含的信息 并进行编解码。 由于本发明实施例所述方案基于 P— Skip模式实现, 因此不会额外的增加码流, 可以节省传递运动矢量所用的传递代价, 并为运动矢量预测提供更多的选择余地,同时通过作为编解码端都可 以获得的信息, 可以为当前编解码提供熵编码等上下文决策信息, 增 强技术的自适应性, 提高编码效率。
图 10为本发明实施例所述方案与现有技术相比在编码性能方面 的比较示意图。 如图 10所示, 假设序列编码类型为 IPPP, 即第一帧 按照帧内预测方式编码,其余帧按照帧间预测方式编码。其中, JM12.3 为在 H.264/AVC平台上使用现有运动补偿方式的技术效果; TM Skip 为在 H.264/AVC平台上使用本发明所述方案的技术效果; TM 16x16 8x8为釆用现有模板匹配方式的技术效果; TM Skip 16x16 8x8为在釆 用现有模板匹配方式的基础上结合本发明所述方案后的技术效果。图
10 中的横坐标表示码率, 纵坐标表示信噪比 (PSNR ), 即表示重建 图像与原始图像的差异。 从图 10可以看出, 本发明实施例所述方案 能够达到较好的节省码率的效果。
通过以上的实施方式的描述,本领域的技术人员可以清楚地了解 到本发明, 可以通过硬件实现, 也可以借助软件加必要的通用硬件平 台的方式来实现。基于这样的理解, 本发明的技术方案可以以软件产 品的形式体现出来, 该软件产品可以存储在一个非易失性存储介质
(可以是 CD-ROM, U盘, 移动硬盘等) 中, 包括若干指令用以使 得一台计算机设备(可以是个人计算机, 服务器, 或者网络设备等) 执行本发明各个实施例所述的方法。
总之, 以上所述仅为本发明的较佳实施例而已, 并非用于限定本 发明的保护范围。 凡在本发明的精神和原则之内, 所作的任何修改、 等同替换、 改进等, 均应包含在本发明的保护范围之内。

Claims

权利要求
1、 一种帧间预测编码方法, 其特征在于, 包括:
获取当前块的模板匹配运动矢量以及运动矢量预测值; 比较所述模板匹配运动矢量以及所述运动矢量预测值,根据比较 结果确定编码方式;
釆用确定的编码方法进行编码。
2、 根据权利要求 1所述的方法, 其特征在于, 所述根据比较结 果确定编码方式并进行编码的步骤包括:
根据比较结果确定为所述当前块进行编码时所需的码流元素,根 据所述确定的码流元素按照不同的编码方式进行编码。
3、 根据权利要求 2所述的方法, 其特征在于, 所述码流元素为 宏块类型和 /或是否使用模板匹配标志; 所述根据比较结果确定为所 述当前块进行编码时所需的码流元素的步骤包括:
如果所述当前块的模板匹配运动矢量和运动矢量预测值一致,则 只在码流中编入宏块类型信息; 如果不一致, 则在所述码流中编入宏 块类型以及是否使用模板匹配标志信息,以指示所述解码端按照模板 匹配方式或运动补偿方式对所述码流进行解码。
4、 根据权利要求 3所述的方法, 其特征在于, 所述在所述码流 中编入宏块类型以及是否使用模板匹配标志信息之前, 进一步包括: 确定需要指示所述解码端按照模板匹配方式还是运动补偿方式对所 述码流进行解码, 包括:
比较釆用所述模板匹配运动矢量以及所述运动矢量预测值进行 编解码的不同性能,指示所述解码端按照所述性能较优的方式进行解 码。
5、 根据权利要求 1所述的方法, 其特征在于, 在获取所述当前 块的模板匹配运动矢量过程中,以所述当前块的运动矢量预测值为中 心, 在其周边预定范围内进行匹配搜索。
6、 根据权利要求 1所述的方法, 其特征在于, 所述根据比较结 果确定编码方式的步骤包括:根据比较结果选择对所述当前块进行编 码的模型, 按照所选择的编码模型进行编码。
7、 根据权利要求 1所述的方法, 其特征在于, 所述根据比较结 果确定编码方式并进行编码之后, 进一步包括:
将所述获取到的模板匹配运动矢量用于为后续当前块进行的运 动矢量预测中。
8、 一种帧间预测解码方法, 其特征在于, 包括:
接收来自编码端的码流;
获取当前块的模板匹配运动矢量以及运动矢量预测值; 比较所述模板匹配运动矢量以及所述运动矢量预测值,根据比较 结果确定解码方式;
釆用确定的解码方式进行解码。
9、 根据权利要求 8所述的方法, 其特征在于, 所述根据比较结 果确定解码方式并进行解码的步骤包括:
如果所述当前块的模板匹配运动矢量与运动矢量预测值一致,则 按照运动补偿方式对所述码流进行解码; 如果不一致, 则根据所述码 流中携带的是否使用模板匹配标志中的指示,按照模板匹配方式或运 动补偿方式进行解码。
10、 根据权利要求 8所述的方法, 其特征在于, 所述比较所述模 板匹配运动矢量以及所述运动矢量预测值之前, 进一步包括:
根据所述码流中携带的宏块类型信息 ,确定针对所述当前块的帧 间预测方式为前向跳过 P— Skip模式。
11、 根据权利要求 8所述的方法, 其特征在于, 在获取所述当前 块的模板匹配运动矢量过程中,以所述当前块的运动矢量预测值为中 心, 在其周边预定范围内进行匹配搜索。
12、 根据权利要求 8所述的方法, 其特征在于, 所述根据比较结 果确定解码方式的步骤包括:根据比较结果选择对所述当前块进行解 码的模型, 按照所选择的解码模型进行解码。
13、 根据权利要求 8所述的方法, 其特征在于, 所述根据比较结 果确定解码方式并进行解码之后, 进一步包括:
将所述获取到的模板匹配运动矢量用于为后续当前块进行的运 动矢量预测中。
14、 一种帧间预测编码装置, 其特征在于, 该装置包括: 获取单 元以及确定单元;
所述获取单元,用于获取当前块的模板匹配运动矢量以及运动矢 量预测值;
所述确定单元,用于比较所述模板匹配运动矢量信息以及运动矢 量预测值, 根据比较结果确定编码方式并进行编码。
15、 根据权利要求 14所述的装置, 其特征在于, 所述确定单元 中具体包括: 比较子单元以及编码子单元;
所述比较子单元,用于比较当前块的模板匹配运动矢量以及运动 矢量预测值, 并将比较结果发送给所述编码子单元;
所述编码子单元, 用于根据所述比较结果进行编码, 如果比较结 果一致, 则只在码流中编入宏块类型信息; 如果不一致, 则在所述码 流中编入宏块类型以及是否使用模板匹配标志信息,以指示所述解码 端按照模板匹配方式或运动补偿方式对所述码流进行解码。
16、 根据权利要求 15所述的装置, 其特征在于, 所述编码子单 元进一步用于, 如果比较结果不一致, 比较釆用所述模板匹配运动矢 量以及所述运动矢量预测值进行编解码的不同性能,指示所述解码端 按照所述性能较优的方式进行解码。
17、 根据权利要求 15所述的装置, 其特征在于, 所述确定单元 中进一步包括: 选择子单元, 用于根据所述比较子单元中对当前块的 模板匹配运动矢量以及运动矢量预测值的比较结果,选择对所述当前 块进行编码的模型, 并通知所述编码子单元。
18、 一种帧间预测解码装置, 其特征在于, 该装置包括: 确定单 元以及解码单元;
所述确定单元, 用于接收来自编码端的码流, 比较获取到的当前 块的模板匹配运动矢量以及运动矢量预测值,根据比较结果确定解码 方式;
所述解码单元, 用于按照所述确定的解码方式进行解码。
19、 根据权利要求 18所述的装置, 其特征在于, 所述确定单元 具体包括: 比较子单元, 用于接收来自编码端的码流, 比较当前块的 模板匹配运动矢量与运动矢量预测值, 如果一致, 则通知所述解码单 元按照运动补偿方式对所述码流进行解码; 如果不一致, 则根据所述 码流中携带的是否使用模板匹配标志中的指示,通知所述解码单元按 照模板匹配方式或运动补偿方式进行解码。
20、 根据权利要求 19所述的装置, 其特征在于, 所述确定单元 中进一步包括: 模式确定子单元, 用于根据所述码流中携带的宏块类 型信息, 确定针对所述当前块的帧间预测方式为 P— Skip模式后, 通 知所述比较子单元执行自身功能。
21、 根据权利要求 19所述的装置, 其特征在于, 所述确定单元 中进一步包括: 选择子单元, 用于根据所述比较子单元对当前块的模 板匹配运动矢量与运动矢量预测值的比较结果,选择对所述码流进行 解码的模型, 并通知所述解码单元。
22、 一种帧间预测解码方法, 其特征在于, 该方法包括: 接收来自编码端的码流;
预设当前块的模板匹配运动矢量或确定模板匹配块信息来得到 当前块的模板匹配运动矢量;
获取所述当前块的模板匹配运动矢量,根据所述码流中携带的是 否使用模板匹配标志的指示, 按照模板匹配方式进行解码。
23、 根据权利要求 22所述的方法, 其特征在于, 所述得到当前 块的模板匹配运动矢量之前, 进一步包括:
根据所述码流中携带的宏块类型信息 ,确定针对所述当前块的帧 间预测方式为前向跳过 P— Skip模式。
24、 根据根据权利要求 22所述的方法, 其特征在于, 所述预设 前块的模板匹配运动矢量为 0 或通过解码所需的模板信息等于前一 个解码模板信息,用来寻获匹配块信息来获取到当前块的模板匹配运 动矢量。
PCT/CN2008/072690 2007-10-15 2008-10-15 An interframe prediction encoding/decoding method and apparatus WO2009052742A1 (en)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP08841850.4A EP2202985B1 (en) 2007-10-15 2008-10-15 An interframe prediction encoding/decoding method and apparatus
JP2010529217A JP5081305B2 (ja) 2007-10-15 2008-10-15 フレーム間予測符号化の方法および装置
US12/761,229 US20100208814A1 (en) 2007-10-15 2010-04-15 Inter-frame prediction coding method and device

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN200710181831.3 2007-10-15
CN200710181831 2007-10-15
CN200810002875.X 2008-01-08
CN200810002875XA CN101415122B (zh) 2007-10-15 2008-01-08 一种帧间预测编解码方法及装置

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US12/761,229 Continuation US20100208814A1 (en) 2007-10-15 2010-04-15 Inter-frame prediction coding method and device

Publications (1)

Publication Number Publication Date
WO2009052742A1 true WO2009052742A1 (en) 2009-04-30

Family

ID=40579084

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2008/072690 WO2009052742A1 (en) 2007-10-15 2008-10-15 An interframe prediction encoding/decoding method and apparatus

Country Status (6)

Country Link
US (1) US20100208814A1 (zh)
EP (1) EP2202985B1 (zh)
JP (1) JP5081305B2 (zh)
KR (1) KR101174758B1 (zh)
CN (1) CN101415122B (zh)
WO (1) WO2009052742A1 (zh)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110176611A1 (en) * 2010-01-15 2011-07-21 Yu-Wen Huang Methods for decoder-side motion vector derivation
JP2012129756A (ja) * 2010-12-14 2012-07-05 Nippon Telegr & Teleph Corp <Ntt> 符号化装置、復号装置、符号化方法、復号方法、符号化プログラム及び復号プログラム
JP2013502141A (ja) * 2009-08-13 2013-01-17 サムスン エレクトロニクス カンパニー リミテッド 動きベクトルを符号化/復号化する方法及びその装置
KR20130105821A (ko) * 2010-09-10 2013-09-26 톰슨 라이센싱 이미지 컨텐츠에 기초한 검색 및 랭킹에 의한, 비디오 압축 시 레퍼런스 블록으로의 링크를 인코딩
US9338477B2 (en) 2010-09-10 2016-05-10 Thomson Licensing Recovering a pruned version of a picture in a video sequence for example-based data pruning using intra-frame patch similarity
US9544598B2 (en) 2010-09-10 2017-01-10 Thomson Licensing Methods and apparatus for pruning decision optimization in example-based data pruning compression
US9602814B2 (en) 2010-01-22 2017-03-21 Thomson Licensing Methods and apparatus for sampling-based super resolution video encoding and decoding
US9813707B2 (en) 2010-01-22 2017-11-07 Thomson Licensing Dtv Data pruning for video compression using example-based super-resolution
CN111770336A (zh) * 2014-03-28 2020-10-13 索尼公司 编码和解码数据的方法、存储介质及数据编码和解码设备

Families Citing this family (28)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101557514B (zh) * 2008-04-11 2011-02-09 华为技术有限公司 一种帧间预测编解码方法、装置及系统
EP2510694B1 (en) 2009-12-10 2017-07-26 Thomson Licensing DTV Method and apparatus for coding and decoding an image block
KR101522850B1 (ko) 2010-01-14 2015-05-26 삼성전자주식회사 움직임 벡터를 부호화, 복호화하는 방법 및 장치
KR101529992B1 (ko) * 2010-04-05 2015-06-18 삼성전자주식회사 픽셀 그룹별 픽셀값 보상을 위한 비디오 부호화 방법과 그 장치, 및 픽셀 그룹별 픽셀값 보상을 위한 비디오 복호화 방법과 그 장치
KR101454133B1 (ko) * 2010-09-02 2014-10-23 엘지전자 주식회사 영상 부호화 및 복호화 방법과 이를 이용한 장치
US8787459B2 (en) * 2010-11-09 2014-07-22 Sony Computer Entertainment Inc. Video coding methods and apparatus
JP2012257198A (ja) 2011-05-17 2012-12-27 Canon Inc 立体画像符号化装置、その方法、および立体画像符号化装置を有する撮像装置
CN102801995B (zh) * 2012-06-25 2016-12-21 北京大学深圳研究生院 一种基于模板匹配的多视点视频运动和视差矢量预测方法
CN103533376B (zh) * 2012-07-02 2017-04-12 华为技术有限公司 帧间预测编码运动信息的处理方法、装置和编解码系统
US20150208083A1 (en) * 2012-07-10 2015-07-23 Lg Electronics Inc. Method and device for processing video signal
US9374597B2 (en) 2012-09-26 2016-06-21 Thomson Licensing Inter-image prediction method and device and corresponding encoding method and device
US9674543B2 (en) * 2012-11-14 2017-06-06 Samsung Electronics Co., Ltd. Method for selecting a matching block
US20150172669A1 (en) * 2013-03-15 2015-06-18 KASAH Technology, Inc. System and method for processing compressed images and video for improved data communication
CN103237224A (zh) * 2013-04-01 2013-08-07 深圳市云宙多媒体技术有限公司 一种基于门限的P_skip块提前判决方法
CN104853209B (zh) * 2014-02-16 2020-09-29 同济大学 图像编码、解码方法及装置
CN103974069B (zh) * 2014-05-22 2017-06-20 华为技术有限公司 可重用的视频编码方法和编码器
CN105338351B (zh) 2014-05-28 2019-11-12 华为技术有限公司 基于模板匹配的帧内预测编、解码、阵列扫描方法及装置
WO2016008157A1 (en) 2014-07-18 2016-01-21 Mediatek Singapore Pte. Ltd. Methods for motion compensation using high order motion model
CN105282558B (zh) * 2014-07-18 2018-06-15 清华大学 帧内像素预测方法、编码方法、解码方法及其装置
CN104244009B (zh) * 2014-08-26 2017-07-07 南京邮电大学 一种分布式视频编码中码率控制方法
WO2017124298A1 (zh) * 2016-01-19 2017-07-27 北京大学深圳研究生院 视频编码、解码方法及其帧间预测方法、装置和系统
US11638027B2 (en) 2016-08-08 2023-04-25 Hfi Innovation, Inc. Pattern-based motion vector derivation for video coding
US12063387B2 (en) 2017-01-05 2024-08-13 Hfi Innovation Inc. Decoder-side motion vector restoration for video coding
CN109089119B (zh) 2017-06-13 2021-08-13 浙江大学 一种运动矢量预测的方法及设备
SG11201913273XA (en) * 2017-06-30 2020-01-30 Huawei Tech Co Ltd Error resilience and parallel processing for decoder side motion vector derivation
EP3709650A4 (en) 2017-11-09 2021-07-14 Samsung Electronics Co., Ltd. APPARATUS AND METHOD FOR CODING MOVEMENT INFORMATION, AND APPARATUS AND METHOD FOR DECODING
US11109041B2 (en) * 2019-05-16 2021-08-31 Tencent America LLC Method and apparatus for video coding
CN110610527B (zh) * 2019-08-15 2023-09-22 苏州瑞派宁科技有限公司 计算suv的方法、装置、设备、系统及计算机存储介质

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1893652A (zh) * 2002-03-29 2007-01-10 株式会社东芝 视频编码方法和装置,以及视频解码方法和装置
CN1976476A (zh) * 2006-12-22 2007-06-06 上海广电(集团)有限公司中央研究院 融合多参考帧选择和运动估计的快速帧间模式判决方法

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH11346365A (ja) * 1998-04-03 1999-12-14 Matsushita Electric Ind Co Ltd 符号化圧縮方法、および符号化圧縮装置
KR100325253B1 (ko) * 1998-05-19 2002-03-04 미야즈 준이치롯 움직임벡터 검색방법 및 장치
US20050105621A1 (en) * 2003-11-04 2005-05-19 Ju Chi-Cheng Apparatus capable of performing both block-matching motion compensation and global motion compensation and method thereof
JP4213646B2 (ja) * 2003-12-26 2009-01-21 株式会社エヌ・ティ・ティ・ドコモ 画像符号化装置、画像符号化方法、画像符号化プログラム、画像復号装置、画像復号方法、及び画像復号プログラム。
JP3879741B2 (ja) * 2004-02-25 2007-02-14 ソニー株式会社 画像情報符号化装置および画像情報符号化方法
JP4145275B2 (ja) * 2004-07-27 2008-09-03 富士通株式会社 動きベクトル検出・補償装置
US9667999B2 (en) * 2005-04-25 2017-05-30 Avago Technologies General Ip (Singapore) Pte. Ltd. Method and system for encoding video data
JP2007043651A (ja) * 2005-07-05 2007-02-15 Ntt Docomo Inc 動画像符号化装置、動画像符号化方法、動画像符号化プログラム、動画像復号装置、動画像復号方法及び動画像復号プログラム
CN1791224A (zh) * 2005-12-19 2006-06-21 宁波大学 一种基于h.264的自适应块搜索范围快速运动估计方法

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1893652A (zh) * 2002-03-29 2007-01-10 株式会社东芝 视频编码方法和装置,以及视频解码方法和装置
CN1976476A (zh) * 2006-12-22 2007-06-06 上海广电(集团)有限公司中央研究院 融合多参考帧选择和运动估计的快速帧间模式判决方法

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10110902B2 (en) 2009-08-13 2018-10-23 Samsung Electronics Co., Ltd. Method and apparatus for encoding/decoding motion vector
JP2013502141A (ja) * 2009-08-13 2013-01-17 サムスン エレクトロニクス カンパニー リミテッド 動きベクトルを符号化/復号化する方法及びその装置
US9883186B2 (en) 2009-08-13 2018-01-30 Samsung Electronics Co., Ltd. Method and apparatus for encoding/decoding motion vector
US8787463B2 (en) 2009-08-13 2014-07-22 Samsung Electronics Co., Ltd. Method and apparatus for encoding/decoding motion vector
US8792558B2 (en) 2009-08-13 2014-07-29 Samsung Electronics Co., Ltd. Method and apparatus for encoding/decoding motion vector
US8811488B2 (en) 2009-08-13 2014-08-19 Samsung Electronics Co., Ltd. Method and apparatus for encoding/decoding motion vector
US9544588B2 (en) 2009-08-13 2017-01-10 Samsung Electronics Co., Ltd. Method and apparatus for encoding/decoding motion vector
US20110176611A1 (en) * 2010-01-15 2011-07-21 Yu-Wen Huang Methods for decoder-side motion vector derivation
US9602814B2 (en) 2010-01-22 2017-03-21 Thomson Licensing Methods and apparatus for sampling-based super resolution video encoding and decoding
US9813707B2 (en) 2010-01-22 2017-11-07 Thomson Licensing Dtv Data pruning for video compression using example-based super-resolution
JP2013539935A (ja) * 2010-09-10 2013-10-28 トムソン ライセンシング 画像コンテンツベースの検索及びランキングによるビデオ圧縮における参照ブロックへのリンクの符号化
US9544598B2 (en) 2010-09-10 2017-01-10 Thomson Licensing Methods and apparatus for pruning decision optimization in example-based data pruning compression
US9338477B2 (en) 2010-09-10 2016-05-10 Thomson Licensing Recovering a pruned version of a picture in a video sequence for example-based data pruning using intra-frame patch similarity
KR20130105821A (ko) * 2010-09-10 2013-09-26 톰슨 라이센싱 이미지 컨텐츠에 기초한 검색 및 랭킹에 의한, 비디오 압축 시 레퍼런스 블록으로의 링크를 인코딩
KR101892019B1 (ko) 2010-09-10 2018-09-28 톰슨 라이센싱 이미지 컨텐츠에 기초한 검색 및 랭킹에 의한, 비디오 압축 시 레퍼런스 블록으로의 링크를 인코딩
JP2012129756A (ja) * 2010-12-14 2012-07-05 Nippon Telegr & Teleph Corp <Ntt> 符号化装置、復号装置、符号化方法、復号方法、符号化プログラム及び復号プログラム
CN111770336A (zh) * 2014-03-28 2020-10-13 索尼公司 编码和解码数据的方法、存储介质及数据编码和解码设备

Also Published As

Publication number Publication date
EP2202985B1 (en) 2017-01-18
JP2011501542A (ja) 2011-01-06
JP5081305B2 (ja) 2012-11-28
EP2202985A1 (en) 2010-06-30
EP2202985A4 (en) 2011-11-23
CN101415122B (zh) 2011-11-16
CN101415122A (zh) 2009-04-22
KR101174758B1 (ko) 2012-08-17
US20100208814A1 (en) 2010-08-19
KR20100083824A (ko) 2010-07-22

Similar Documents

Publication Publication Date Title
WO2009052742A1 (en) An interframe prediction encoding/decoding method and apparatus
US11910014B2 (en) Image encoding method using a skip mode, and a device using the method
CN106488249B (zh) 运动图像编码装置及其方法、运动图像解码装置及其方法
JP5452584B2 (ja) ビデオ符号化及び復号化におけるテンプレート・マッチング予測(tmp)のための方法及び装置
EP1618744B1 (en) Video transcoding
CN101946516A (zh) 快速宏块增量量化参数的决定
JP2015156653A (ja) ビデオコーディングのための適応動き解像度
WO2021239160A1 (zh) 视频解码方法、视频编码方法、电子设备及存储介质
US20090016443A1 (en) Inter mode determination method for video encoding
TWI806212B (zh) 視訊編碼器、視訊解碼器及相應方法
KR20110073263A (ko) 인트라 예측 부호화 방법 및 부호화 방법, 그리고 상기 방법을 수행하는 인트라 예측 부호화 장치 및 인트라 예측 복호화 장치
CN110351552B (zh) 视频编码中一种快速编码方法
CN111937389B (zh) 用于视频编解码的设备和方法
WO2021190518A1 (zh) 解码、编码、编解码方法、装置及其设备
TW202236852A (zh) 高效視訊編碼器架構
TWI737141B (zh) 視訊編碼中用於畫面內分區之轉換類型分配方法和裝置
CN113301335A (zh) 视频信号的编码和解码方法

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: 08841850

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2010529217

Country of ref document: JP

NENP Non-entry into the national phase

Ref country code: DE

REEP Request for entry into the european phase

Ref document number: 2008841850

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2008841850

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 20107010509

Country of ref document: KR

Kind code of ref document: A