KR20050049296A - Apparatus and method for interframe wavelet coding for adjustability of computing complexity - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

The present invention relates to an interframe wavelet coding apparatus and a method thereof.

2. Technical problem to be solved by the invention

The present invention provides an interframe wavelet coding apparatus and method for reducing computational complexity of a decoding apparatus by controlling a split level and a filter coefficient according to a frame information amount during wavelet transformation.

3. Summary of Solution to Invention

An MCTF unit for filtering the input GOP image on the time axis, a wavelet transform unit for spatial wavelet transforming the filtered frame, a quantizer for quantizing wavelet coefficients output from the wavelet transform unit, a motion vector calculated by the MCTF unit, and the And an entropy coding unit for entropy coding the wavelet coefficients quantized by the quantization unit and a wavelet filter manager for selecting a segmentation level and a filter length of the wavelet transform unit based on the motion prediction information of the GOP image calculated by the MCTF unit.

4. Important uses of the present invention

Used for video coding and decoding.

Description

Apparatus and Method for Interframe Wavelet Coding for Adjustability of Computing Complexity

The present invention relates to a video coding apparatus and a method thereof. More particularly, the present invention relates to an interframe wavelet coding apparatus and a method for controlling a computational complexity of a decoding apparatus by controlling a segmentation level and a filter length during wavelet transformation.

In general, the interframe wavelet coding method is a coding method suitable for various applications requiring scalability functions because of its excellent compression rate and scalability. In particular, in heterogeneous network environments such as convergence networks of broadcasting and communication, scalability of video is the most important function of guaranteeing quality of service (QoS).

The conventional DCT-based video coding method is optimized for supporting a single scalability function and thus is not suitable for a service requiring multiple scalability functions such as a heterogeneous network environment.

Interframe wavelet coding is also referred to as 3-D sub-band coding because it performs 2D image wavelet transform on an image obtained by subband filtering on the time axis.

When filtering on the time axis, filtering is applied in consideration of the motion of the image to improve compression efficiency, not just filtering. Therefore, such filtering is referred to as Motion Compensated Temporal Filtering (MCTF).

Interframe wavelet coding starts with MCTF of an image input in units of a group of pictures (GOP).

The MCTF includes a method of time-base filtering a plurality of frames and a method of time-base filtering two frames. In the time-base filtering method of a plurality of frames, original frames are left as reference frames and high-pass filtered frames are generated based on adjacent frames.

In the time-base filtering method of two frames, two adjacent frames are cross-referenced to generate lowpass and highpass filtered frames.

In this case, a motion vector is calculated for each input block before the filtering and the motion vector is applied to the filtering.

The Ohm and Woods methods are known to apply motion vectors to MCTF. In general, Woods method, which is more efficient for calculating motion vectors, is widely used.

On the other hand, Haar filters are widely used in MCTF. The Haar filter lowpass and highpass filter two adjacent frames as a unit.

1 is a diagram for explaining a process of MCTF using a Haar filter. For convenience of explanation, the case where the GOP is 8 will be described as an example.

Referring to the drawings, when the MCTF is performed on two adjacent frames, four tL frames low-pass filtered and four tH frames 111-112, 113, and 114 high-pass filtered on the time axis are generated, and the first temporal level (1st temporal level) is generated. , 110).

When the MCTF is performed on the four low-pass filtered tL frames again, two low-pass filtered t-LL frames and two t-LH frames 121 and 122 are generated, and the second temporal level (2nd temporal level, 120).

When the MCTF is performed on the two t-LL frames low-pass filtered in the same manner, one t-LLL frame 131 and one t-LLH frame 132 are generated, and the third time level ( 3rd temporal level (130).

As such, the eight frames 111, 112, 113, 114, 121, 122, 131, and 132 on which the MCTF is performed undergo spatial wavelet transform, quantization, and entropy coding steps, and finally a bit frame is generated and transmitted to the receiving end. do.

In the conventional wavelet transform step, the 9/7 filter or the 5/3 filter are used in the same manner regardless of the frame.

However, if the wavelet transform is performed using the same filter regardless of the frame, there is a problem in that the amount of calculation is unnecessarily large during the inverse wavelet transform of the decoding step. The amount of calculation of the inverse wavelet transform requires a larger amount of divisions of the wavelet transform, and a larger amount of calculation requires a longer filter length.

As described above, when the amount of computation of the inverse wavelet transform is large, there is a problem in that an interframe video coding method cannot be used in a terminal having low computing power, such as a PDA.

An object of the present invention is to provide an interframe wavelet coding apparatus and method for reducing the computational complexity of a decoding apparatus by controlling a segmentation level and a filter length according to a frame information amount during wavelet transformation. have.

In addition, an object of the present invention is to provide an interframe wavelet coding apparatus and method which can decode even a terminal having low computing power by reducing inverse wavelet computational complexity.

Those skilled in the art to which the present invention pertains can easily recognize other objects and advantages of the present invention from the drawings, the detailed description of the invention, and the claims.

The present invention for achieving the above object is an interframe wavelet coding apparatus, the MCTF unit for filtering the input GOP image with respect to the time axis, the wavelet transform unit for spatial wavelet transform the filtered frame, the wavelet output from the wavelet transform unit The wavelet is based on a quantization unit for quantizing a coefficient, an entropy coding unit for entropy coding a motion vector computed by the MCTF unit, and a wavelet coefficient quantized in the quantization unit, and motion prediction information of the GOP image computed by the MCTF unit. And a wavelet filter manager for selecting a dividing level and a filter length of the converter.

In addition, the present invention provides an interframe wavelet coding apparatus, the MCTF unit for performing low pass and high pass filtering of the input GOP image on the time axis and recursively performing low pass and high pass filtering on the low pass filtered frame. A first wavelet transformer having a split level and a maximum filter length, the wavelet transforming the last low-pass filtered one frame, a split level and a shorter filter length that is less than or equal to the first wavelet transformer, and a high pass filtered remaining GOP A motion vector calculated by the MCTF unit and a quantization unit quantizing the wavelet coefficients output from the first wavelet transformer and the second wavelet transformer, and a quantized wavelet coefficient Entropy coding unit for entropy coding It should.

The present invention also relates to an interframe decoding apparatus, which includes an entropy decoding unit for entropy decoding a bitstream including split level and filter length information for inverse wavelet transform, and an inverse quantization of quantized wavelet coefficients output from the entropy decoding unit. A quantization unit, an inverse wavelet transform unit for inverse wavelet transforming the wavelet coefficients output from the inverse quantizer based on the partition level and the filter length, and an MCTF synthesis unit for performing MCTF synthesis based on a motion vector of the entropy decoding unit. .

In addition, the present invention provides an interframe wavelet coding method, the MCTF step of filtering the input GOP image on the time axis, the wavelet transform step of spatial wavelet transforming the filtered frame, and the quantization step of quantizing the wavelet coefficients generated in the wavelet transform step. An entropy coding step of entropy coding the motion vector calculated in the MCTF step and the wavelet coefficients quantized in the quantization step, and a segmentation level and a filter of the wavelet transform step based on the motion prediction information of the GOP image calculated in the MCTF step A wavelet transform selecting step of selecting a length and including the splitting level and the filter length information in the entropy coded bitstream.

In addition, the present invention provides an interframe wavelet decoding method for entropy decoding a bitstream including split level and filter length information for inverse wavelet transform, and inversely quantizing quantized wavelet coefficients generated from the entropy decoding step. MCTF synthesis step of performing inverse quantization, inverse wavelet transform of wavelet coefficients generated from the inverse quantization step based on the partition level and filter length and MCTF synthesis based on the motion vector of the entropy decoding step It includes.

The following merely illustrates the principles of the invention. Therefore, those skilled in the art, although not explicitly described or illustrated herein, can embody the principles of the present invention and invent various devices that fall within the spirit and scope of the present invention. In addition, all conditional terms and embodiments listed herein are in principle clearly intended to be understood solely for the purpose of understanding the concept of the invention and are not to be limited to the specifically listed embodiments and states. do.

In addition, it is to be understood that all detailed descriptions, including the principles, aspects, and embodiments of the present invention, as well as listing specific embodiments, are intended to include structural and functional equivalents of these matters. In addition, these equivalents should be understood to include not only equivalents now known, but also equivalents to be developed in the future, that is, all devices invented to perform the same function regardless of structure.

Thus, for example, the block diagrams herein should be understood to represent a conceptual view of example circuitry embodying the principles of the invention. Similarly, all flowcharts, state transitions, pseudocodes, and the like are understood to represent various processes performed by a computer or processor, whether or not the computer or processor is substantially illustrated on a computer readable medium and whether the computer or processor is clearly shown. Should be.

The functionality of the various elements shown in the figures, including functional blocks represented by a processor or similar concept, can be provided by the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functionality may be provided by a single dedicated processor, by a single shared processor or by a plurality of individual processors, some of which may be shared.

In addition, the explicit use of terms presented in terms of processor, control, or similar concept should not be interpreted exclusively as a citation to hardware capable of running software, and without limitation, ROM for storing digital signal processor (DSP) hardware, software. (ROM), RAM, and non-volatile memory are to be understood to implicitly include. Other hardware for the governor may also be included.

In the claims of this specification, components expressed as means for performing the functions described in the detailed description include all types of software including, for example, a combination of circuit elements or firmware / microcode, etc. that perform the functions. It is intended to include all methods of performing a function which are combined with appropriate circuitry for executing the software to perform the function. The invention, as defined by these claims, is equivalent to what is understood from this specification, as any means capable of providing such functionality, as the functionality provided by the various enumerated means are combined, and in any manner required by the claims. It should be understood that.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings. First of all, in adding reference numerals to the components of each drawing, it should be noted that the same components have the same number as much as possible even if displayed on different drawings. In addition, in describing the present invention, when it is determined that the detailed description of the related known technology may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

2 is a block diagram of a video coding apparatus according to the present invention.

The conventional interframe wavelet coding method incurs a large amount of computation during inverse wavelet transform during the decoding process. The decoding device may operate in various terminals having very different computing power, such as PDAs, notebook computers, PCs, set-top boxes, and the like.

Therefore, if video is coded in consideration of the computing power of each terminal, an optimized service may be provided for each terminal. To this end, the present invention provides a coding method capable of adjusting the amount of computation in the inverse wavelet transform process, which takes a large amount of computation in the decoding process. That is, the present invention performs wavelet transform based on the discrimination level and the filter length which are differentiated according to the characteristics of the MCTF frames.

Referring to the drawings, the coding apparatus according to the present invention calculates a motion vector from an input GOP image, and uses the MCTF unit 210 to perform filtering on a time axis, and to perform spatial wavelet transform on the filtered frame. The wavelet transform unit 220, a quantization unit 230 for quantizing the wavelet transformed wavelet coefficients, an entropy coding unit 240 for coding the motion vector and the quantized wavelet coefficients, and a wavelet transform unit according to the information amount of the filtered frame ( And a wavelet filter manager 250 for selecting an appropriate filter length and division level of 220.

The MCTF unit 210 lowpass and highpass filter the input GOP images using the motion vector as described above. The MCTF unit 210 generally uses a Haar filter and repeatedly applies low pass filtering and high pass filtering until the number of low pass filtered frames becomes one. Thus, when the GOP is 2 ^N , the MCTF is total Will be performed once.

On the other hand, when the MCTF is performed, most of the image information is present in the low-pass filtered frame, and the amount of the image information in the high-pass filtered frame is proportional to the degree of change of the frame.

That is, when the change in the frame is small, most of the image information exists in the low-pass filtered frame on the time axis, and the high-pass filtered frame has little information. On the other hand, when there are many changes in the frame, a lot of information exists in the high-pass filtered frame.

Frames that do not contain much information do not affect the compression efficiency of the coding device even when the resolution of the wavelet transforming unit 220 is low. Compression efficiency is hardly affected

Accordingly, the coding apparatus according to the present invention includes a wavelet filter manager 250 to select an appropriate filter length and a split level of the wavelet transform unit 220 based on motion prediction information, which is information about a change in frames generated during MCTF. do. The wavelet transform unit 220 performs spatial wavelet transform of each frame based on the selected division level and the filter length.

3 illustrates in detail the coding process according to the present invention. For convenience of explanation, it is assumed that the input frame has a GOP of 4.

As shown, when an image frame is input, the MCTF unit 210 performs subband filtering on a time axis. The filtering process of the MCTF unit 210 is as described with reference to FIG. 1.

That is, the MCTF unit 210 generates two high pass filtered t-H frames 311 and 313 and two low pass filtered t-L frames 312 and 314 for the first time level.

The wavelet filter manager 250 selects an appropriate filter length and a split level of the wavelet transform unit 220 to spatially wavelet transform the high pass filtered tH frames 311 and 313 based on the motion prediction information obtained during the MCTF. do.

The wavelet transform unit 220 spatially transforms the high-pass filtered t-H frames 311 and 313 using a filter having the selected split level and the selected filter length.

Meanwhile, the MCTF unit 210 performs subband filtering on the low-pass filtered tL frames 312 and 314 again on the time axis, and one high-pass filtered t-LH frame 315 on the second time level. And generate one low-pass filtered t-LL frame 316.

In addition, the wavelet filter management unit 250 selects an appropriate partition level and the selected filter of the wavelet transform unit 220 to wavelet the high-pass filtered t-LH frame 315 based on the motion prediction information obtained in the MCTF process. After selecting the length, the wavelet transform unit 220 performs a spatial wavelet transform on the high-pass filtered t-LH frame 315 using a filter having the selected split level and the filter length.

In addition, the low-pass filtered T-LL frame 316 is spatially wavelet-transformed by the wavelet transform unit 220 having the maximum dividing level and the maximum filter length.

4 is a block diagram of a video coding apparatus according to the present invention when a GOP (group of picture) is 8. Referring to the drawings, the GOP image input by the MCTF unit 210 includes tH frames 111, 112, 113, and 114 of the first time level, t-LH frames 121 and 122 of the second time level, and t-LLH frames of the third time level. 132, filtered into t-LLL frame 131.

In general, since the t-LLL frame 131 has a large amount of information present, the t-LLL frame 131 is wavelet-converted by the wavelet converter 410 having the maximum dividing level and the maximum filter length. Preferably, the wavelet converter 410 may use a 9/7 filter or a 5/3 filter having three or four division levels.

In addition, the t-LLH frame 132 of the third temporal level is wavelet transformed by the wavelet converter 412, and the t-LH frames 121 and 122 of the second temporal level are wavelet transformed by the wavelet converter 414.

Further, the t-H frames 111, 112, 113, and 114 of the first temporal level are wavelet transformed by the wavelet converter 416.

The wavelet filter manager 250 selects the dividing level and the filter length of the wavelet transformers 412, 414, 416 based on the amount of information included in each frame based on the motion prediction information obtained during the MCTF, and the wavelet transform unit 220. To control. In this case, the wavelet transducers 412, 414, and 416 may select to have a dividing level and a filter length smaller than or equal to the wavelet transducer 410.

In addition, the wavelet filter manager 250 may control the split level and the filter length to have an optimal amount of calculation in consideration of the computing capability of the decoding terminal. In particular, in order to minimize the decoding operation of the decoding terminal, each wavelet converter 412, 414, 416 may use a Haar filter having a one-step partitioning level.

In addition, the wavelet filter manager 250 includes information on the selected split level and the filter length in the coded bitstream so that the wavelet filter manager 250 can be used during inverse wavelet transform.

When the wavelet transform is completed, the wavelet transform coefficients of each temporal level are input to the quantization unit 105 and quantized, and the quantized wavelet coefficients and the motion vector are input to the entropy coding unit 240 to generate a bitstream.

5 is a block diagram of a video decoding apparatus according to the present invention. Referring to the figure, the bitstream received by the decoding apparatus includes information on the dividing level and filter length of the wavelet transformer selected in the coding step.

The entropy decoding unit 510 receives the bitstream and performs entropy decoding to output quantized wavelet coefficients and a motion vector, and the quantized wavelet coefficients are input to the inverse quantizer 520 and dequantized.

The wavelet coefficients output from the inverse quantization unit 520 are input to the inverse wavelet transform unit 530 together with the split level and the filter length information included in the bitstream.

The inverse wavelet transform unit 530 performs inverse wavelet transform on the input wavelet coefficients based on the dividing level and the filter length of the filter used in the wavelet transform in the coding apparatus.

The MCTF synthesizing unit 540 reconstructs the GOP image by MCTF synthesizing a frame of each subband in which inverse wavelet transform is completed using the motion vector decoded by the entropy decoding unit 510.

As mentioned above, although this invention was demonstrated by the limited embodiment and drawing, this invention is not limited by this, The person of ordinary skill in the art to which this invention belongs, Of course, various modifications and variations are possible within the scope of equivalent claims.

According to the present invention, the operation amount of the inverse wavelet transform in the decoding process can be adjusted by controlling the segmentation level and the filter length of the wavelet transform filter according to the frame information amount in the video coding process.

Accordingly, the interframe wavelet decoding is possible even in a PDA terminal having a low computational capability by controlling the division level and the filter length of the wavelet transform filter in consideration of the computational capability of the decoding terminal.

1 is a view for explaining a process of motion compensation temporal filtering (MCTF);

2 is a block diagram of a video coding apparatus according to the present invention;

3 is a view for explaining a coding process of FIG.

4 is a block diagram of a video coding apparatus according to the present invention when a group of picture (GOP) is 8;

5 is a block diagram of a video decoding apparatus according to the present invention.

<Description of main reference numerals in the drawings>

210: MCTF unit 220: wavelet converting unit

230: quantization unit 240: entropy coding unit

250: wavelet filter management unit

510: entropy decoding unit 520: inverse quantization unit

530: inverse wavelet transform unit 540: MCTF synthesis unit

Claims (18)

An MCTF unit for filtering the input GOP image with respect to the time axis; A wavelet transform unit for spatial wavelet transforming the filtered frame; A quantizer for quantizing wavelet coefficients output from the wavelet transform unit; An entropy coding unit for entropy coding a motion vector computed by the MCTF unit and a wavelet coefficient quantized by the quantization unit; And A wavelet filter manager for selecting a segmentation level and a filter length of the wavelet transform unit based on the motion prediction information of the GOP image calculated by the MCTF unit. Including, The partition level and filter length information are included in the entropy coded bitstream. Interframe Wavelet Coding Device The method of claim 1, The MCTF unit Recursively perform lowpass and highpass filtering on lowpass filtered frames, The wavelet filter management unit Selecting the dividing level and the filter length of the wavelet transform unit according to the information amount of the filtered frame Interframe Wavelet Coding Device. The method of claim 2, The wavelet transform unit A first wavelet converter having a maximum division level and a maximum filter length and wavelet transforming the last low pass filtered one frame, and A second wavelet transformer having a split level and a filter length less than or equal to the first wavelet transformer and wavelet transforming the remaining high-pass filtered GOP frames; Containing Interframe Wavelet Coding Device. The method of claim 3, wherein The wavelet filter management unit Selecting a 9/7 filter having a dividing level of 3 or 4 with the first wavelet transducer; Interframe Wavelet Coding Device. The method of claim 3, wherein The wavelet filter management unit Selecting a 5/3 filter having three or four division levels with the first wavelet transducer Interframe Wavelet Coding Device. The method of claim 3, wherein The wavelet filter management unit The second wavelet transformer selects a partition level and a Haar filter in one step. Interframe Wavelet Coding Device. An MCTF unit performing low pass and high pass filtering of the input GOP image on the time axis and recursively performing low pass and high pass filtering on the low pass filtered frame; A first wavelet converter having a maximum division level and a maximum filter length and wavelet transforming the last low pass filtered one frame; A second wavelet converter having a split level and a filter length less than or equal to the first wavelet converter and wavelet transforming the remaining high pass filtered GOP frame; A quantizer for quantizing wavelet coefficients output from the first wavelet transformer and the second wavelet transformer; And An entropy coding unit for entropy coding a motion vector calculated by the MCTF unit and a wavelet coefficient quantized by the quantization unit; Interframe Wavelet Coding Device. The method of claim 7, wherein The first wavelet converter 9/7 filter with 3 or 4 levels of division Containing Interframe Wavelet Coding Device. The method of claim 8, The first wavelet converter 5/3 filter with 3 or 4 levels of division Containing Interframe Wavelet Coding Device. The method of claim 8, The second wavelet converter is Haar filter with one level of split Containing Interframe Wavelet Coding Device. An entropy decoding unit for entropy decoding a bitstream including a split level and filter length information for inverse wavelet transform; An inverse quantizer for inversely quantizing the quantized wavelet coefficients output from the entropy decoding unit; An inverse wavelet transform unit for inverse wavelet transforming the wavelet coefficients output from the inverse quantizer based on the division level and the filter length; And An MCTF synthesis unit configured to perform MCTF synthesis based on the motion vector of the entropy decoding unit; Containing Interframe wavelet video decoding device. An MCTF step of filtering the input GOP image on the time axis; Wavelet transforming the filtered frame with a spatial wavelet transform; A quantization step of quantizing the wavelet coefficients generated in the wavelet transform step; An entropy coding step of entropy coding a motion vector calculated in the MCTF step and a wavelet coefficient quantized in the quantization step; A wavelet transform selection step of selecting a segmentation level and a filter length of the wavelet transform step based on the motion prediction information of the GOP image calculated in the MCTF step; And Including the partition level and filter length information in the entropy coded bitstream. Containing Interframe Wavelet Coding Method. The method of claim 12, The MCTF stage Recursively performing lowpass and highpass filtering on the lowpass filtered frame, The wavelet transform selection step is Selecting the division level and the filter length according to the amount of information that the filtered frame has Interframe Wavelet Coding Method. The method of claim 13, The wavelet transform step is Wavelet transforming the last low pass filtered one frame using a first wavelet transformer having a maximum division level and a maximum filter length, and Wavelet transforming the remaining GOP frames using a second wavelet transformer having a split level and a filter length less than or equal to the first wavelet transformer Containing Interframe Wavelet Coding Method. The method of claim 14, The wavelet transform selection step is Selecting a 9/7 filter having a dividing level of 3 or 4 with the first wavelet transducer; Interframe Wavelet Coding Method. The method of claim 14, The wavelet transform selection step is Selecting a 5/3 filter having three or four division levels with the first wavelet transducer Interframe Wavelet Coding Method. The method of claim 14, The wavelet transform selection step is Selecting a Haar filter having a split level of one step by the second wavelet converter Interframe Wavelet Coding Method. An entropy decoding step of entropy decoding a bitstream including split level and filter length information for inverse wavelet transform; An inverse quantization step of inverse quantizing the quantized wavelet coefficients generated from the entropy decoding step; An inverse wavelet transform step of inverse wavelet transforming the wavelet coefficients generated from the inverse quantization step based on the division level and the filter length; And An MCTF synthesis step of performing MCTF synthesis based on the motion vector of the entropy decoding step; Containing Interframe Wavelet Decoding Method.