JP5529537B2 - Method and apparatus for multi-path video encoding and decoding - Google Patents

Description

[Display related applications]
This application is filed on Sep. 22, 2006, and is entitled “METHOD AND APPARATUS FOR MULTIPLE PASS VIDEO VIDEO CODING AND DECODING”. The benefit of PCT / US2006 / 037139 is claimed, and this PCT International Application No. PCT / US2006 / 037139 is included in this application.

  The present invention relates to video encoding and decoding, and more particularly to a method and apparatus for multi-path video encoding and decoding.

  Currently, International Organization for Standardization / International Electrotechnical Commission (ISO / IEC) Moving Picture Experts Group-4 (MPEG-4) Part 10 Advanced Video Coding (AVC) Standard / International Telecommunication Union Telecommunication Division (ITU-T) H. The H.264 standard (hereinafter referred to as “MPEG4 / H.264 standard” or simply “H.264 standard”) is the most powerful and latest video coding standard. As with all other video coding standards, H.264. The H.264 standard uses block-based motion compensation and transform coding such as discrete cosine transform (DCT). It is well known that DCT is efficient for video coding and is suitable for high end applications such as broadcast high definition television (HDTV). However, the DCT algorithm is not suitable for applications that require very low bit rates, such as dedicated video cell phones. At very low bit rates, the DCT transform introduces blocking artifacts even when using a deblocking filter. This is because only very few coefficients are encoded at a very low bit rate, and each coefficient tends to have a very coarse quantization step.

  Matching pursuits (MP) is a greedy algorithm that decomposes an arbitrary signal into a linear range of waveforms selected from a redundant dictionary of functions. These waveforms are selected to best match the signal configuration / structure.

  Now suppose that we have 1-D signals f (t) and we want to decompose this signal using basis vectors from an over-complete dictionary set G. Individual dictionary functions can be expressed as follows:

  Here, γ is a subscript parameter associated with a specific dictionary element. The decomposition begins by selecting γ and maximizes the absolute value of the inner product as follows:

  The residual signal is then calculated as follows:

This residual signal is then developed in the same way as the original signal. Such a procedure continues iteratively until a determined number of expansion factors are generated or several energy thresholds for the residual are reached. At each stage n, a dictionary function γ _n is generated. After all M stages, the signal is approximated and acquired by a linear function of the dictionary as follows:

It has been found that the complexity of the matching tracking decomposition of an n-sample signal is on the order of k · N · d · nlog ₂ n. Here, d is dependent on the size of the dictionary that does not consider conversion, N is the number of selected expansion coefficients, and the constant k depends on the policy for selecting the dictionary function. Given an advanced overcomplete dictionary, matching tracking It is more computationally intensive than the 8 × 8 and 4 × 4 DCT integer transforms used in the H.264 standard, and the complexity is defined as O (nlog ₂ n).

  Usually, the matching tracking algorithm corresponds to an arbitrary set of redundant base shapes. It has been proposed to develop the signal using an overcomplete basis of the Gabor function. The 2-D Gabor dictionary is very redundant and each shape may be at an integer pixel location in the encoded residual image. Matching tracking has a larger dictionary set and each encoded basis function matches well with the structure in the residual signal, so the frame-based Gabor dictionary does not contain an artificial block structure.

  Gabor Redundant Dictionary Set is very low based on matching tracking for the proposed video coding system using matching tracking algorithm (hereinafter referred to as “the traditional matching tracking video coding method based on Gabor”) Adopted for bit rate video coding. The proposed system is based on a simulation model for very low bit rate image coding, or a framework of a low bit rate hybrid DCT system, referred to as “SIM3” for short, with a DCT residual encoder , Replaced with a matching tracking encoder. This encoder uses matching tracking and decomposes motion residual images into 2-D Gabor functions that can be lexicographically separated. The proposed system has been shown to perform well with low motion sequences at low bit rates.

  A smooth 16 × 16 sine-squared window was applied to the predicted image for the 8 × 8 partition in the traditional Gabor-based matching tracking video coding approach. The matching tracking video codec in the conventional tracking tracking video coding method based on Gabor is ITU-TH. Based on H.263 codec. However, H.C. The H.264 standard allows variable block size motion compensation with a small block size that may be as small as 4 × 4 for luminance motion compensation. Further, H.C. The H.264 standard is primarily based on transforms such as 4x4 DCT for the baseline and main profiles, rather than 8x8 as most other prominent conventional video coding techniques. Directional space prediction for intra coding improves the quality of the predicted signal. All these highlighted features / configurations are H.264 standard is more efficient, but matching tracking is It requires handling more complex situations when applying to the H.264 standard. A smooth 16 × 16 sine square window is shown as follows:

  A hybrid coding scheme (hereinafter referred to as “conventional hybrid coding scheme”) is used for H.264 for motion estimation. It was proposed to benefit from some of the features and structures introduced by the H.264 standard and to replace the transformation in the spatial domain. Prediction errors are encoded using a matching tracking algorithm that decomposes the signal with an appropriately designed two-dimensional anisotropic redundancy dictionary. In addition, rapid atom search technology was introduced. However, the proposed conventional hybrid coding scheme does not deal with whether the proposed conventional hybrid coding scheme uses a one-path scheme or a two-path scheme. Further, in the proposed conventional hybrid coding system, the motion estimation unit is H.264. H.264 standard, but whether any deblocking filter is used for the encoding scheme, or any other method smooths the blocking artifacts caused by the predicted image at very low bit rates. Does not deal with whether or not used.

  The above and other disadvantages and disadvantages of the prior art are overcome by the present invention. The present invention relates to a method and apparatus for multi-path video encoding and decoding.

  According to one aspect of the present invention, a video encoder is provided for encoding video signal data using a multi-path video encoding scheme. The video encoder includes a motion estimator and a decomposition module. The motion estimator performs motion estimation on the video signal data and obtains a motion residual corresponding to the video signal data in the first coding path. The decomposition module is in signal communication with the motion estimator and decomposes the motion residual in a later coding path.

  In accordance with another aspect of the present invention, a method for encoding video signal data using a multi-path video encoding scheme is provided. The method performs motion estimation on the video signal data to obtain a motion residual corresponding to the video signal data in the first coding path, and decomposes the motion residual in a later coding path. Is included.

  According to another aspect of the invention, a video decoder is provided for decoding a video bitstream. The video decoder includes an entropy decoder, an atom decoder, an inverse transformer, a motion compensator, a deblocking filter, and a combiner. The entropy decoder decodes the video bitstream and obtains a decompressed video bitstream. The atom decoder is capable of signal communication with the entropy decoder, decodes the decompressed atom corresponding to the decompressed bitstream, and obtains the decrypted atom. The inverse transformer is capable of signal communication with the atom decoder and applies the inverse transform to the decoded atom to form a reconstructed residual image. The motion compensator is in signal communication with the entropy decoder and performs motion compensation using a motion vector corresponding to the decompressed bitstream to form a reconstructed predicted image. The deblocking filter is capable of signal communication with the motion compensator, performs deblocking filtering on the reconstructed prediction image, and smoothes the reconstructed prediction image. The combiner is in signal communication with the inverse transformer and the overlap block motion compensator, and combines the reconstructed predicted image and the residual image to obtain a reconstructed image.

  According to another aspect of the present invention, a method for decoding a video bitstream is provided. The method decodes a video bitstream and obtains a decompressed video bitstream; decodes a decompressed atom corresponding to the decompressed bitstream and obtains a decoded atom; Applying the inverse transform to the decoded atom to form a reconstructed residual image, and performing motion compensation using the motion vector corresponding to the decompressed bitstream to reconstructed prediction A step of forming an image, a step of performing deblocking filtering on the reconstructed prediction image, smoothing the reconstructed prediction image, and a reconstructed image by combining the reconstructed prediction image and the residual image. And obtaining a step.

The above and other aspects, features, configurations, and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention, and the detailed description of the present invention will be made with reference to the accompanying drawings. Should be read while.
The invention is better understood with reference to the following exemplary drawings.

In accordance with an embodiment of the principles of the present invention, a two-path H.E. 2 is a diagram of an exemplary first path of an exemplary encoder in a matching tracking encoder / decoder (CODEC) based on the H.264 standard. In accordance with an embodiment of the principles of the present invention, a two-path H.E. 2 is a diagram of an exemplary second path of an exemplary encoder in a matching tracking encoder / decoder (CODEC) based on the H.264 standard. In accordance with an embodiment of the principles of the present invention, two paths H. 2 is a diagram of an exemplary decoder in a matching tracking encoder / decoder (CODEC) based on the H.264 standard. FIG. 4 is a diagram for an exemplary method for encoding an input video stream in accordance with an embodiment of the present principles. FIG. 4 is a diagram for an exemplary method for decoding an input video stream in accordance with an embodiment of the present principles.

  The present invention relates to a method and apparatus for multi-path video encoding and decoding. Advantageously, the present invention is useful for H.264 in very low bit rate applications, for example. Corrects blocking artifacts introduced by the DCT transform used in the H.264 standard. Furthermore, the present invention is not limited to just low bit rate applications, but may be used for other (large) bit rates while maintaining the scope of the present invention.

  This specification illustrates the principles of the invention. Those skilled in the art may practice the principles of the present invention and make various modifications and changes that fall within the spirit and scope of the invention, although not explicitly described or shown herein.

  All examples and conditional explanations listed here are for educational purposes and help the reader understand the principles and concepts of the present invention contributed by the inventor to promote technology. It should be construed that the invention is not limited to such specific recited examples and conditions.

  Furthermore, all descriptions herein which enumerate the principles, forms, embodiments and specific examples of the present invention are intended to encompass these structural and functional equivalents. . In addition, such equivalents include presently known equivalents and equivalents developed in the future, ie, any element developed to perform the same function regardless of structure.

  Thus, for example, it will be appreciated by those skilled in the art that the block diagrams provided herein show conceptual diagrams of exemplary circuits that implement the principles of the invention. Similarly, any flowcharts, flow diagrams, state transition diagrams, pseudocode, etc. are substantially represented on a computer-readable medium, whether or not the computer or processor is explicitly shown. The various processes performed by are shown.

  The functionality of the various elements shown in the figures may be provided by the use of dedicated hardware and hardware capable of executing software in combination with appropriate software. When provided by a processor, the functionality may be provided by one dedicated processor, one shared processor, or multiple individual processors that may be partially shared. Furthermore, the explicit use of the terms “processor” or “controller” should not be construed to refer exclusively to hardware capable of executing software, but digital signal processor (“DSP”) hardware. Hardware, read only memory (“ROM”) storing software, random access memory (“RAM”), and non-volatile storage means may be implicitly included, but are not limited thereto.

  Conventional and / or specially configured separate hardware may be included. Similarly, any switches shown in the drawings are merely conceptual. These functions may be performed by program logic operations, dedicated logic, interaction between program control and dedicated logic, or manually, and by the implementer so that the specific technique is more fully understood from the context. Selectable.

  In the claims, any element indicated as a means for performing a specified function is intended to include any way of performing that function, for example: a) A combination of circuit elements that perform a function; b) includes any form of software, the software of any form including firmware, microcode, etc., and an appropriate circuit that activates the software to perform the function Combined with. The invention as defined by such claims is such that the functions provided by the various recited means are combined and brought together in the form required by the claims. Any means that can provide the above functionality is considered equivalent to the means shown herein.

  In accordance with the principles of the present invention, a multi-path video encoding and decoding scheme is provided. Multi-path video encoding and decoding schemes may be used with matching tracking. In an illustrative embodiment, the two-path H.264. An encoding scheme based on H.264 is disclosed for matching tracking video encoding.

  H. The H.264 standard utilizes motion compensation based on blocks similar to other video compression standards and transforms such as DCT. At very low bit rates, the DCT transform incorporates blocking artifacts even when using a deblocking filter. This is because only very few coefficients can be encoded at a very low bit rate, and each coefficient tends to have a very coarse quantization step. In accordance with the principles of the present invention, matching tracking using overcomplete bases is utilized to encode the residual image. The motion compensation and mode decision unit is H.264. H.264 standard. Overlap block motion compensation (OBMC) is used to smooth the predicted image. In addition, a new approach is provided for selecting bases other than matching tracking.

  In accordance with the principles of the present invention, the video encoder and / or decoder applies OBMC to the predicted image to reduce blocking artifacts caused by the prediction model. A matching tracking algorithm is used to encode the residual image. The advantage of matching tracking is that the matching tracking is not based on blocks, is based on frames, and there are no blocking artifacts caused by differences in coding residuals.

  Referring to FIG. 1A and FIG. Typical first and second path portions of the encoder in a matching tracking encoder / decoder (CODEC) based on the H.264 standard are indicated by reference numerals 110 and 160. The encoder is indicated by reference numeral 190 and the decoder portion is indicated by reference numeral 191.

  Referring to FIG. 1A, the input of the first path unit 110 can be in signal communication with the non-inverting input of the combiner 112, the input of the encoder control module 114, and the first input of the motion estimator 116. It is connected to the. The first output portion of the coupler 112 is connected to the first input portion of the buffer 118 so as to allow signal communication. The second output of the combiner 112 is connected to the input of the integer transform / scaling / quantization module 120 so as to be in signal communication. The output unit of the integer transform / scaling / quantization module 120 is connected to the first input unit of the scaling / inverse transform module 122 so that signal communication is possible.

  The first output unit of the encoder control module 114 is connected to the first input unit of the inter-frame predictor 126 so as to allow signal communication. The second output unit of the encoder control module 114 is connected to the first input unit of the motion compensator 124 so as to be capable of signal communication. The third output unit of the encoder control module 114 is connected to the second input unit of the motion estimator 116 so as to be capable of signal communication. The fourth output unit of the encoder control module 114 is connected to the second input unit of the scaling / inverse conversion module 122 so as to allow signal communication. The fifth output unit of the encoder control module 114 is connected to the first input unit of the buffer 118 so as to allow signal communication.

  The output unit of the motion estimator 116 is connected to the second input unit of the motion compensator 124 and the second input unit of the buffer 128 so as to allow signal communication. The inverting input unit of the combiner 112 is selectively connected to the signal communication means with the output unit of the motion compensator 124 or the output unit of the inter-frame predictor 126 so as to be capable of signal communication. The selected output of the motion compensator 124 or the inter-frame predictor 126 is connected in signal communication with the first input of the combiner 128. The output unit of the scaling / inverse conversion module 122 is connected to the second input unit of the coupler 128 so as to be able to perform signal communication. The output unit of the combiner 128 is connected to the second input unit of the inter-frame predictor 126, the third input unit of the motion estimator 116, and the input / output unit of the motion compensator 124 so as to be capable of signal communication. The output unit of the buffer 118 can be used as the output unit of the first path unit 110.

  For the first path unit 110, an encoder control module 114, an integer transform / scaling / quantization module 120, a buffer 118, and a motion estimator 116 are included in the encoder 190. Further, for the first path portion, a scaling / inverse transform module 122, an inter-frame predictor 126, and a motion compensator 124 are included in the decoder portion 191.

  The input unit of the first path unit 110 receives the input video 111 and stores control data (eg, motion vectors, mode selections, predicted images, etc.) in the buffer 118 for use in the second path unit 160. To do.

  Referring to FIG. 1B, the first input unit of the second path unit 160 is connected to the input unit of the entropy encoder 166 so as to be capable of signal communication. The first input unit receives control data 162 (for example, mode selection or the like) and motion vector 164 from the first path unit 110. The second input unit of the second path unit 160 is connected to the non-inverting input unit of the coupler 168 so as to allow signal communication. The third input unit of the second path unit 160 is connected to the input unit of the overlap block motion compensation (OBMC) / deblocking module 170 so as to be capable of signal communication. The second input unit of the second path unit 160 receives the input video 111, and the third input unit of the second path unit receives the predicted image 187 from the first path unit 110.

  An output unit of the combiner 168 that provides the residual 172 is connected to an input unit of an atom detector 174 so as to be capable of signal communication. The output of the atom detector 174 that provides the encoded residual 178 is connected in signal communication with the input of the atom encoder 176 and the first non-inverting input of the combiner 180. The output portion of the OBMC / deblocking module 170 is connected to the inverting input portion of the combiner 168 and the second non-inverting input portion of the combiner 180 so as to be capable of signal communication. An output unit of the combiner 180 that provides output video is connected to an input unit of the reference buffer 182 so as to be capable of signal communication. The output unit of the atom encoder 176 is connected to the input unit of the entropy encoder 166 so that signal communication is possible. The output unit of the entropy encoder 166 can be used as the output unit of the second path unit 160 and provides an output bit stream.

  For the second path section 160, the entropy encoder is included in the encoder 190, and the combiner 168, OBMC module 170, atom detector 174, atom encoder 176, and reference buffer 182 are included in the decoder section 191. include.

  Referring to FIG. An exemplary decoder in a matching tracking encoder / decoder (CODEC) based on the H.264 standard is indicated by reference numeral 200.

  The input unit of the decoder 200 is connected to the input unit of the entropy decoder 210 so that signal communication is possible. The output unit of the entropy decoder is connected to the input unit of the atom decoder 220 and the input unit of the motion compensator 250 so as to be capable of signal communication. The output unit of the inverse transform module 230 that provides the residual is connected to the first non-inverting input unit of the combiner 270 so as to be capable of signal communication. The output unit of the motion compensator 250 is connected to the input unit of the OBMC / deblocking module 260 so that signal communication is possible. The output part of the OBMC / deblocking module 260 is connected to the second non-inverting input part of the coupler 270 so as to allow signal communication. The output unit of the combiner can be used as the output unit of the decoder 200.

  H. Unlike the matching tracking video codec in the conventional tracking tracking video coding method based on Gabor based on the H.263 codec, the principle of the present invention is based on the ITU-TH. It can be applied to H.264 / AVC encoding system. Because of frame-based residual coding, we The OBMC that is not implemented in the H.264 / AVC codec is applied to the predicted image.

  In an embodiment based on the principles of the present invention, the first path in the video coding scheme is H.264. H.264 standard. In the first path, no encoding is actually performed. For example, all control data such as mode selections, predicted images, and motion vectors are stored in a buffer for the second path. A DCT transform has been applied to the first path for motion compensation and mode selection using rate distortion optimization (RDO). Instead of encoding the residual image using DCT coefficients, all residual images are saved for the second path. In embodiments based on the principles of the present invention, 16 × 16 constrained intra coding or H.264. It is proposed to apply H.264 standard compliant forced intra coding and in particular to handle the boundary portion between intra-coded and inter-coded macroblocks.

  In the second path, the motion vector and the control data may be encoded by entropy encoding. The residual image may be encoded by matching tracking. Atom search and parameter encoding may be performed based on, for example, conventional Gabor based matching tracking video encoding techniques. The reconstructed image is saved for the reference frame.

  One advantage of matching tracking video coding is that matching tracking is not block based and there are no blocking artifacts. However, if motion prediction is performed on a block basis and is inaccurate, this motion prediction generates some blocking artifacts at very low bit rates. Simulations show that atoms appear in moving contours and regions, where the motion vector (MV) is not very accurate. By improving motion prediction, atoms can better represent residuals.

  In order to remove artifacts from motion prediction, one method is described in H.264. H.264 or the like, or using an improved deblocking filter to smooth the block boundaries in the predicted image. In another approach, a smoothed motion model using overlapping blocks (OBMC) is used. The conventional Gabor-based matching tracking video coding technique employs a 16 × 16 sine square window. The N × N sine square window may be defined, for example, according to a conventional hybrid coding scheme. A 16 × 16 sine square window is formed for the 8 × 8 block, and the 16 × 16 block is processed as four 8 × 8 blocks.

  However, H.C. In the H.264 standard, partitions having luma block sizes of 16 × 16, 16 × 8, 8 × 16, and 8 × 8 samples are supported. If a partition with 8 × 8 samples is selected, the 8 × 8 partition is further divided into 8 × 4, 4 × 8, or 4 × 4 luminance sample and corresponding chromaticity sample partitions. Here, four approaches are proposed to accommodate more partition types. In the first approach, an 8 × 8 sine square window is used for 4 × 4 partitions. For all other partitions larger than 4x4, these partitions are divided into several 4x4 partitions. The second approach uses a 16 × 16 sine square window for 8 × 8 or larger partitions, but this second approach does not affect partitions smaller than 8 × 8. The third technique uses an adaptive OBMC for all partitions. All of the above three methods implement only OBMC and do not implement a deblocking filter. In the fourth method, OBMB and a deblocking filter are combined.

  In addition to the redundant Gabor dictionary set in the traditional Gabor-based matching tracking video coding scheme implemented for residual coding, we propose to utilize more overcomplete bases. At low bit rates, the transformed motion model cannot accurately represent the natural motion of relevant visual features / constructions such as moving edges. Therefore, most of the residual error energy is located in these regions. Therefore, it is meaningful to represent an error image using an edge detection redundant dictionary. Discrete wavelet transforms (e.g., 2-D Dual-Tree Discrete Wavelet Transform (DDWT)) are more likely than 2-D Gabor dictionaries are used or some other edge detection dictionaries are used. Has less redundancy. 2-DDDWT has more subbands / directivity than 2-DDWT. Each subband indicates a certain direction, and each subband is for edge detection. After noise shaping, 2-DDDDWT achieves a higher PSNR with the same retained coefficients compared to standard 2-DDWT. Therefore, 2-DDDWT is more suitable for encoding edge information. After applying OBMC to the predicted image, the error image has smooth edges. A parametric overcomplete 2-D dictionary may be used to provide smooth edges.

  With reference to FIG. 3, an exemplary method for encoding an input video stream is indicated by reference numeral 300. The method 300 includes a start block 305 and proceeds to decision block 310. Decision block 310 determines whether the current frame is an I frame. If the current frame is an I frame, proceed to function block 355. If the current frame is not an I frame, proceed to function block 315.

  The function block 355 includes the H.264 function. H.264 standard compliant frame encoding is performed, an output bitstream is provided, and the process proceeds to end block 370.

  The function block 315 is an H.264 function. H.264 standard compliant motion compensation is performed and processing proceeds to function block 320. The function block 320 stores the motion vector (MV), control data, and prediction block, and proceeds to decision block 325. Decision block 325 determines whether the end of the frame has been reached. If the end of the frame has been reached, proceed to function block 330. If the end of the frame has not been reached, return to function block 315.

  The function block 330 performs OBMC and / or deblocking filtering on the predicted image and proceeds to a function block 335. The function block 335 obtains a residual image from the original image and the predicted image and proceeds to the function block 340. The function block 340 encodes the residual using matching tracking and proceeds to a function block 345. The function block 345 performs entropy encoding to provide an output bitstream and proceeds to an end block 370.

  Referring to FIG. 4, an exemplary method for decoding an input video sequence is indicated by reference numeral 400. The method 400 includes a start block 400 and proceeds to decision block 410. Decision block 410 determines whether the current frame is an I frame. If the current frame is an I frame, proceed to function block 435. If the current frame is not an I frame, proceed to function block 415.

  The function block 435 includes H.264. H.264 standard compliant decoding is performed to provide a reconstructed image and processing proceeds to end block 470.

  The function block 415 decodes the motion vector, control data, and matching tracking atom and proceeds to the function block 420 and the function block 425. The function block 420 reconstructs the residual image using the decoded atom and proceeds to a function block 430. The function block 425 reconstructs the predicted image by decoding the motion vectors and other control data and applying OBMC / and / or deblocking filtering and proceeds to a function block 430. The function block 430 combines the reconstructed residual image and the reconstructed predicted image to provide a reconstructed image and proceeds to an end block 470.

  Some of the many attendant advantages / features of the present invention have been described, some of which have been described above. For example, one advantage / feature is a video encoder that encodes video signal data using a multi-path video encoding scheme, where the video encoder includes a motion estimator and a decomposition module. The motion estimator performs motion estimation on the video signal data and obtains a motion residual corresponding to the video signal data in the first coding path. The decomposition module is in signal communication with the motion estimator and decomposes motion residuals in later coding paths.

  Another advantage / feature is the video encoder as described above, and the multi-path video encoding scheme is a two-path video encoding scheme. The video encoder further includes a buffer, wherein the buffer is in signal communication with the motion estimator and the decomposition module and is obtained in the first coding path for use in the second coding path. Remember the difference. The decomposition module decomposes the motion residual using the redundant Gabor dictionary set in the second coding path.

  Another advantage / feature is a video encoder that uses a two-path video encoding scheme as described above, wherein the motion estimator is in the first encoding path in the International Telecommunication Union Telecommunication Sector (ITU-T). ) H. Motion estimation and coding mode selection according to the H.264 standard are executed.

  Another advantage / feature is a video encoder that uses a two-pass video encoding scheme as described above, which further includes a prediction module and an overlap block motion compensator. The prediction module is in signal communication with the buffer and forms a prediction image corresponding to the video signal data in the first coding path. The overlap block motion compensator is in signal communication with the buffer, performs overlap block motion compensation (OBMC) on the predicted image using a 16 × 16 sine square window, and predicts in the second coding path. Smooth the image. The buffer stores the predicted image in the first path for use in the second coding path.

  Yet another advantage / feature is a video encoder that uses a two-pass video encoding scheme as described above, the video encoder further including a prediction module and an overlap block motion compensator. The prediction module is in signal communication with the buffer and forms a prediction image corresponding to the video signal data in the first coding path. The overlap block motion compensator is in signal communication with the buffer and performs overlap block motion compensation (OBMC) only on partitions larger than 8 × 8 of the predicted image in the second coding path. The buffer stores a predicted image in the first coding path for use in the second coding path.

  Yet another advantage / feature is a video encoder that uses a two-pass video encoding scheme as described above, the video encoder further including a prediction module and an overlap block motion compensator. The prediction module is in signal communication with the buffer and forms a prediction image corresponding to the video signal data in the first coding path. The overlap block motion compensator is in signal communication with the buffer and uses an 8 × 8 sine square window for the 4 × 4 partition of the predicted image in the second coding path to overlap block motion compensation (OBMC). ). When OBMC is performed in the second coding path, all partitions of the predicted image are divided into 4 × 4 partitions. The buffer stores a predicted image in the first coding path for use in the second coding path.

  Another advantage / feature is a video encoder that uses a two-pass video coding scheme as described above, the video encoder further including a prediction module and an overlap block motion compensator. The prediction module is in signal communication with the buffer and forms a prediction image corresponding to the video signal data in the first coding path. The overlap block motion compensator is in signal communication with the buffer and performs overlap block motion compensation (OBMC) that is adaptive for all partitions of the predicted image in the second coding path. The buffer stores a predicted image in the first coding path for use in the second coding path.

  Yet another advantage / feature is a video encoder that uses a two-pass video encoding scheme as described above, the video encoder further including a prediction module and a deblocking filter. The prediction module is in signal communication with the buffer and forms a prediction image corresponding to the video signal data in the first coding path. The deblocking filter is in signal communication with the buffer and performs a deblocking operation on the predicted image in the second coding path. The buffer stores a predicted image in the first coding path for use in the second coding path.

  Yet another advantage / feature is a video encoder that uses a two-pass video coding scheme as described above, and the decomposition module performs a dual tree wavelet transform to decompose the motion residual.

  In addition, another advantage / feature is a video encoder that uses a two-path video coding scheme and a dual tree wavelet transform as described above, where the decomposition module uses noise shaping to determine the coefficients of the dual tree wavelet transform. select.

  Yet another advantage / feature is a video encoder that uses a two-pass video encoding scheme as described above, where the decomposition module utilizes a parametric overcomplete 2-D dictionary and in the second encoding path. Decompose the motion residual.

  Yet another advantage / feature is a video decoder that decodes a video bitstream, the video decoder comprising an entropy decoder, an atom decoder, an inverse transformer, a motion compensator, a deblocking filter, and a combiner Is included. The entropy decoder decodes the video bitstream and obtains a decompressed video bitstream. The atom decoder is capable of signal communication with the entropy decoder, decodes the decompressed atom corresponding to the decompressed bitstream, and obtains the decoded atom. The inverse transformer is capable of signal communication with the atom decoder and applies the inverse transform to the decoded atom to form a reconstructed residual image. The motion compensator is in signal communication with the entropy decoder and performs motion compensation using a motion vector corresponding to the decompressed bitstream to form a reconstructed predicted image. The deblocking filter is capable of signal communication with the motion compensator, performs deblocking filtering on the reconstructed prediction image, and smoothes the reconstructed prediction image. The combiner is in signal communication with the inverse transformer and the overlap block motion compensator, and combines the reconstructed predicted image and the residual image to obtain a reconstructed image.

  These and other features / configurations and advantages of the present invention will be readily ascertainable by one of ordinary skill in the art based on the teachings herein. The teachings of the present invention may be implemented in various forms of hardware, software, firmware, special purpose processors, or combinations thereof.

  Most preferably, the teachings of the present invention are implemented as a combination of hardware and software. Furthermore, the software may be implemented as an application program realized in a specific form in the program storage unit. The application program may be uploaded to and executed by a machine having any suitable architecture. Preferably, the machine is on a computer platform having hardware such as one or more central processing units (“CPU”), random access memory (“RAM”), and input / output (“I / O”) interfaces. Implemented. The computer platform may include an operating system and microinstruction code. The various processes and functions described herein may be part of microinstruction code or part of an application program that may be executed by the CPU, or any combination thereof. In addition, various other peripheral devices may be connected to the computer platform, for example, the various other peripheral devices may be other data storage devices and printing devices.

  Preferably, some of the constituent system elements and methods shown in the accompanying drawings are implemented in software, so that the actual connection between system elements or processing function blocks is the way in which the present invention is programmed. It may be different depending on. Given the teachings herein, one of ordinary skill in the related art will be able to contemplate these and similar implementations or configurations of the present invention.

While illustrative embodiments have been described herein with reference to the accompanying drawings, the invention is not limited to the precise embodiments described above, and various modifications and changes can be made without departing from the scope or spirit of the invention. It may be made by those skilled in the art without. All such variations and modifications are intended to be included within the scope of the present invention as set forth in the appended claims.
[Appendix 1]
A video encoder that encodes video signal data using a multi-path video encoding scheme, the video encoder comprising:
A motion estimator (116) that performs motion estimation on the video signal data to obtain a motion residual corresponding to the video signal data in a first coding path;
A decomposition module (174) that is in signal communication with the motion estimator and decomposes the motion residual in a later coding path;
Including a video encoder.
[Appendix 2]
The multi-path video encoding scheme is a two-path video encoding scheme, and the video encoder further includes a buffer (118), and the buffer (118) is capable of signal communication with the motion estimator and the decomposition module. And storing the motion residual obtained in the first coding path for use in the second coding path, the decomposition module (174) being a redundant in the second coding path. The video encoder of claim 1, wherein the motion residual is decomposed using a Gabor dictionary set.
[Appendix 3]
The motion estimator (116) is connected to the International Telecommunication Union Telecommunication Division (ITU-T) H.264 in the first coding path. The video encoder according to appendix 2, which performs motion estimation and encoding mode selection according to the H.264 standard.
[Appendix 4]
A prediction module (124, 126) that is in signal communication with the buffer and forms a predicted image corresponding to the video signal data in the first encoding path;
Signal communication with the buffer is performed, and an overlap block motion compensation (OBMC) is performed on the prediction image using a 16 × 16 sine square window, and the prediction image in the second coding path is smoothed. An overlap block motion compensator (170);
The video encoder of claim 2, further comprising: the buffer storing the predicted image in the first coding path for subsequent use in the second coding path.
[Appendix 5]
A prediction module (124, 126) that is in signal communication with the buffer and forms a predicted image corresponding to the video signal data in the first encoding path;
An overlap block motion compensator (170) capable of signal communication with the buffer and performing overlap block motion compensation (OBMC) only on 8 × 8 or more partitions of the predicted image in the second coding path; ,
The video encoder of claim 2, further comprising: the buffer storing the predicted image in the first coding path for subsequent use in the second coding path.
[Appendix 6]
A prediction module (124, 126) that is in signal communication with the buffer and forms a predicted image corresponding to the video signal data in the first encoding path;
Overrun that is in signal communication with the buffer and performs overlap block motion compensation (OBMC) using an 8 × 8 sine square window for 4 × 4 partitions of the predicted image in the second coding path. A wrap block motion compensator (170);
And when OBMC is performed in the second coding path, all partitions of the predicted image are divided into 4 × 4 partitions, and the buffer follows in the second coding path. The video encoder of claim 2, storing the predicted image in the first encoding path for use.
[Appendix 7]
A prediction module (124, 126) that is in signal communication with the buffer and forms a predicted image corresponding to the video signal data in the first encoding path;
Overlapping block motion compensator (170) that is in signal communication with the buffer and that performs adaptive overlapping block motion compensation (OBMC) for all partitions of the predicted image in the second coding path When,
The video encoder of claim 2, further comprising: the buffer storing the predicted image in the first coding path for subsequent use in the second coding path.
[Appendix 8]
A prediction module (124, 126) that is in signal communication with the buffer and forms a predicted image corresponding to the video signal data in the first encoding path;
A deblocking filter (170) that is in signal communication with the buffer and that performs a deblocking operation on the predicted image in the second coding path;
The video encoder of claim 2, further comprising: the buffer storing the predicted image in the first coding path for subsequent use in the second coding path.
[Appendix 9]
The video encoder of claim 2, wherein the decomposition module (174) performs a dual tree wavelet transform to decompose the motion residual.
[Appendix 10]
The video encoder of claim 9, wherein the decomposition module (174) uses noise shaping to select coefficients for the dual-tree wavelet transform.
[Appendix 11]
The video encoder of claim 2, wherein the decomposition module (174) uses a parametric overcomplete 2-D dictionary to decompose the motion residual in the second encoding path.
[Appendix 12]
A method of encoding video signal data using a multipath video encoding scheme, comprising:
Performing motion estimation on the video signal data to obtain a motion residual corresponding to the video signal data in a first coding path (315);
Decomposing the motion residual in a later coding path (340);
Including a method.
[Appendix 13]
The multi-path video coding scheme is a two-path video coding scheme, and the method includes the motion residual acquired in the first coding path for subsequent use in a second coding path. 13. The method of claim 12, further comprising the step of storing (320), wherein the decomposing step (340) decomposes the motion residual using a redundant Gabor dictionary set in the second encoding path. .
[Appendix 14]
The motion estimation and coding mode selection may be performed according to International Telecommunication Union Telecommunication Sector (ITU-T) H.264 in the first coding path. The method according to appendix 13, which is performed according to the H.264 standard.
[Appendix 15]
Forming a predicted image corresponding to the video signal data in the first encoding path (315);
Storing the predicted image in the first encoding path (320);
Performing overlap block motion compensation (OBMC) on the predicted image using a 16 × 16 sine square window to smooth the predicted image in the second coding path (330);
The method according to appendix 13, further comprising:
[Appendix 16]
Forming a predicted image corresponding to the video signal data in the first encoding path (315);
Storing the predicted image in the first encoding path (320);
Performing overlap block motion compensation (OBMC) on only 8 × 8 or more partitions of the predicted image in the second coding path (330);
The method according to appendix 13, further comprising:
[Appendix 17]
Forming a predicted image corresponding to the video signal data in the first encoding path (315);
Storing the predicted image in the first encoding path (320);
Performing overlap block motion compensation (OBMC) using an 8 × 8 sine square window for 4 × 4 partitions of the predicted image in the second coding path;
The method of claim 13, further comprising: when OBMC is performed on the second coding path, all partitions of the predicted image are divided into 4 × 4 partitions.
[Appendix 18]
Forming a predicted image corresponding to the video signal data in the first encoding path (315);
Storing the predicted image in the first encoding path (320);
Performing adaptive block motion compensation (OBMC) adaptive for all partitions of the predicted image in the second coding path (330);
The method according to appendix 13, further comprising:
[Appendix 19]
Forming a predicted image corresponding to the video signal data in the first encoding path (315);
Storing the predicted image in the first encoding path (320);
Performing a deblocking operation on the predicted image in the second coding path (330);
The method according to appendix 13, further comprising:
[Appendix 20]
The method of claim 13, wherein the decomposing step (340) performs a dual tree wavelet transform to decompose the motion residual.
[Appendix 21]
The method of claim 20, wherein the decomposing step (340) uses noise shaping and selects coefficients for the dual-tree wavelet transform.
[Appendix 22]
The method of claim 13, wherein the decomposing step (340) uses a parametric overcomplete 2-D dictionary to decompose the motion residual in the second coding path.
[Appendix 23]
A video decoder for decoding a video bitstream,
An entropy decoder (210) for decoding the video bitstream to obtain a decompressed video stream;
An atom decoder (220) that is in signal communication with the entropy decoder and decodes a decompressed atom corresponding to the decompressed bitstream to obtain a decoded atom;
An inverse transformer (230) that is in signal communication with the atom decoder and applies an inverse transform to the decoded atom to form a reconstructed residual image;
A motion compensator (250) that is in signal communication with the entropy decoder and performs motion compensation using a motion vector corresponding to the decompressed bitstream to form a reconstructed prediction image; ,
A deblocking filter (260) that is in signal communication with the motion compensator and performs deblocking filtering on the reconstructed prediction image to smooth the reconstructed prediction image;
A combiner (270) that is in signal communication with the inverse transformer and the overlap block motion compensator and combines the reconstructed prediction image and the residual image to obtain a reconstructed image;
Including a decoder.
[Appendix 24]
A method for decoding a video bitstream, comprising:
Decoding the bitstream and obtaining a decompressed video bitstream (405);
Decoding a decompressed atom corresponding to the decompressed bitstream to obtain a decrypted atom (415);
Applying an inverse transform to the decoded atom to form a reconstructed residual image (420);
Performing motion compensation using a motion vector corresponding to the decompressed bitstream to form a reconstructed predicted image (425);
Performing deblocking filtering on the reconstructed predicted image and smoothing the reconstructed predicted image (425);
Combining the reconstructed predicted image and the residual image to obtain a reconstructed image (430);
Including a method.

Claims (6)

A video encoder that encodes a video signal using multiple paths,
A motion estimator for performing motion estimation on the video signal ;
A motion compensator for performing motion compensation on the video signal;
A transform quantizer that performs DCT transform, scaling, and quantization on the residual signal of the first path;
A circuit for reconstructing a predicted video signal from the residual signal of the first path;
A circuit for storing the predicted image and control data and transferring to the encoder of the second path;
An adder circuit that calculates a second path residual by finding a difference between the original video signal and the reconstructed predicted video signal;
A filter that performs deblocking or overlap block motion compensation on the predicted image from the first path ;
A matching tracking encoder that is in signal communication with the motion estimator and that represents the motion residual of the second path by decomposing and representing a waveform expansion selected from an edge detection redundancy dictionary of linear functions;
Including a video encoder.   The multi-path video encoding scheme is a two-path video encoding scheme, the video encoder further includes a buffer, and the buffer is capable of signal communication with the motion estimator and the decomposition module; Storing the motion residual obtained in the first coding path for subsequent use in the coding path, the decomposition module using a redundant Gabor dictionary set in the second coding path; The video encoder of claim 1, wherein the motion residual is decomposed.   The motion estimator is connected to the International Telecommunication Union Telecommunication Sector (ITU-T) H.264 in the first coding path. The video encoder of claim 2, wherein the video encoder performs motion estimation and coding mode selection according to the H.264 standard. A method of encoding a video signal using multiple paths, comprising:
Performing mode determination and motion estimation on the video signal ;
Motion compensating the video signal;
Performing DCT transform, scaling, and quantization on the residual signal of the first path;
Reconstructing a predicted video signal from the residual signal of the first path;
Storing the predicted image and control data and transferring to the encoder of the second path;
Calculating a second path residual by finding a difference between the original video signal and the reconstructed predicted video signal;
Performing deblocking filtering or overlap block motion compensation on the predicted image from the first path ;
Representing the motion residual of the second path in a later coding path as a development of a waveform selected from an edge detection redundancy dictionary of linear functions;
Including a method.   The multi-path video coding scheme is a two-path video coding scheme, and the method includes the motion residual acquired in the first coding path for subsequent use in a second coding path. 5. The method of claim 4, further comprising: storing the motion residual using a redundant Gabor dictionary set in the second coding path.   The motion estimation and coding mode selection may be performed according to International Telecommunication Union Telecommunication Sector (ITU-T) H.264 in the first coding path. The method of claim 5, wherein the method is performed according to the H.264 standard.

Images