Academia.eduAcademia.edu

Outline

Title
Abstract
Figures
Introduction
Experimental Results
Conclusions
References
All Topics
Computer Science
Biomedical Informatics

Subband coding of image sequences at low bit rates

Profile image of Jan BiemondJan Biemond

1990, Signal Processing: Image Communication

https://doi.org/10.1016/0923-5965(90)90029-H
visibility

…

description

8 pages

link

1 file

Abstract

In this paper a low bit rate subband coding scheme for image sequences is described. Typically, the scheme is based on temporal DPCM in combination with an intraframe subband coder. In contrast to previous work, however, the subbands are divided into blocks onto which conditional replenishment is applied, while a bit allocation algorithm divides the bits among the blocks assigned for replenishment. A solution is given for the "dirty window' effect by setting blocks to zero that were assigned to be replenished but received no bits. The effect of motion compensation and the extension to color images are discussed as well. Finally, several image sequence coding results are given for a bit rate of 300 kbit/s.

Figures (10)
Fig. 1. General hybrid image sequence coding scheme with motion compensation.
Fig. 1. General hybrid image sequence coding scheme with motion compensation.
“ig. 3. SNR2s5; values for each frame of Trevor, for using CR block sizes of (a) 4x4 and (b) 8x8 samples.
“ig. 3. SNR2s5; values for each frame of Trevor, for using CR block sizes of (a) 4x4 and (b) 8x8 samples.
Fig. 4. Image sequence encoding scheme including motion compensation and Zero Block Correction. P.H. Westerink et al. / Subband coding of image sequences
Fig. 4. Image sequence encoding scheme including motion compensation and Zero Block Correction. P.H. Westerink et al. / Subband coding of image sequences
Fig. 6. Subband image sequence coding scheme without motion compensation. Fig. 5. The effect on the SNRs; for (a) with and (b) without motion compensation.
Fig. 6. Subband image sequence coding scheme without motion compensation. Fig. 5. The effect on the SNRs; for (a) with and (b) without motion compensation.
Fig. 7. Image sequence encoding scheme using subband coding for the chrominance components U and V. subbands and then only the low-pass sudbdand needs to be encoded. As with the luminance signal the chrominance components are encoded using temporal DPCM, while the frame prediction error does not benefit from the application of motion compensation. The resulting FD subbands can be encoded quite well using the same type of Lloyd-Max quantizers as was used for the luminance FD subbands. Applica- tion of conditional replenishment to the chromin- ance subbands did not improve the reconstructed sequence. For the overall bit allocation the bits are now distributed over the 2 chrominance subbands and the remaining blocks of the 12 luminance subbands. As a color distortion measure for the bit allocation algorithm the sum of the mean- squared error of each color component is taken. To assure a color image, the chrominance sub- bands always received at least one bit/sample. Finally, the encoder for the chrominance signals is shown in Fig. 7. It can be seen that the frame memory is in the subband domain, holding only the low-pass subband. memory is in the subband domain, holding only is Shown in Fig. 7. It can be seen that the frame
Fig. 7. Image sequence encoding scheme using subband coding for the chrominance components U and V. subbands and then only the low-pass sudbdand needs to be encoded. As with the luminance signal the chrominance components are encoded using temporal DPCM, while the frame prediction error does not benefit from the application of motion compensation. The resulting FD subbands can be encoded quite well using the same type of Lloyd-Max quantizers as was used for the luminance FD subbands. Applica- tion of conditional replenishment to the chromin- ance subbands did not improve the reconstructed sequence. For the overall bit allocation the bits are now distributed over the 2 chrominance subbands and the remaining blocks of the 12 luminance subbands. As a color distortion measure for the bit allocation algorithm the sum of the mean- squared error of each color component is taken. To assure a color image, the chrominance sub- bands always received at least one bit/sample. Finally, the encoder for the chrominance signals is shown in Fig. 7. It can be seen that the frame memory is in the subband domain, holding only the low-pass subband. memory is in the subband domain, holding only is Shown in Fig. 7. It can be seen that the frame
Fig. 8. Uncoded and decoded subbands of the frame difference between frame 1 and 3 of Trevor. it can be seen that here the low-pass subband looks As an example, in Fig. 8(a) the subbands of the frame difference between frame 1 and 3 of the COST 211-bis sequence Trevor are shown. Also shown in Fig. 8(b) are the decoded subbands, that is, after application of conditional replenishment, bit allocation and = scalar quantization for 300 kbit/s. With the image dimensions being 256 x 256 pixels, the equivalent bit rate for each frame is therefore always fixed to approximately 0.366 bits/ pixel. The frame rate is 12.5 frames per second. Compared to the subbands of still pictures it can be seen that here the low-pass subband looks s therefore always fixed to approximately
Fig. 8. Uncoded and decoded subbands of the frame difference between frame 1 and 3 of Trevor. it can be seen that here the low-pass subband looks As an example, in Fig. 8(a) the subbands of the frame difference between frame 1 and 3 of the COST 211-bis sequence Trevor are shown. Also shown in Fig. 8(b) are the decoded subbands, that is, after application of conditional replenishment, bit allocation and = scalar quantization for 300 kbit/s. With the image dimensions being 256 x 256 pixels, the equivalent bit rate for each frame is therefore always fixed to approximately 0.366 bits/ pixel. The frame rate is 12.5 frames per second. Compared to the subbands of still pictures it can be seen that here the low-pass subband looks s therefore always fixed to approximately
Fig. 9. The effect of applying Zero Block Correction (ZBC) or the SNR,;,: (a) with ZBC and (b) without ZBC. ANU LIGA Li sal etary pay 8eYUAaIIY IO INET’ VU. The effect of Zero Block Correction is shown in Fig. 9, where the values of the SNR); are plotted for each frame coded at 300 kbit/s. It can be seen that starting at approximately frame 30 the applied correction becomes favorable, because at this point the motion becomes more severe. Finally, in Fig. 10(a, b) frame 51 of the Trevor sequence is shown with and without applying ZBC, respectively. It can be seen that artefacts such as those above the shoulders are largely removed. As a penalty, the corrected frame appears to be slightly less sharp. Note, for instance, that the stripes in the shirt have disappeared. However, in observing the real time sequence the lack of sharpness seems to be con- cealed by the amount of motion present. Therefore a large visual improvement can be noticed when the correction is applied. Compared to the more commonly used algorithms where DCT and MC are applied, there are no blocking effects visible because there is no block transform involved. Instead, the coding degradation manifests itself as a low-pass effect, which in general can be con-
Fig. 9. The effect of applying Zero Block Correction (ZBC) or the SNR,;,: (a) with ZBC and (b) without ZBC. ANU LIGA Li sal etary pay 8eYUAaIIY IO INET’ VU. The effect of Zero Block Correction is shown in Fig. 9, where the values of the SNR); are plotted for each frame coded at 300 kbit/s. It can be seen that starting at approximately frame 30 the applied correction becomes favorable, because at this point the motion becomes more severe. Finally, in Fig. 10(a, b) frame 51 of the Trevor sequence is shown with and without applying ZBC, respectively. It can be seen that artefacts such as those above the shoulders are largely removed. As a penalty, the corrected frame appears to be slightly less sharp. Note, for instance, that the stripes in the shirt have disappeared. However, in observing the real time sequence the lack of sharpness seems to be con- cealed by the amount of motion present. Therefore a large visual improvement can be noticed when the correction is applied. Compared to the more commonly used algorithms where DCT and MC are applied, there are no blocking effects visible because there is no block transform involved. Instead, the coding degradation manifests itself as a low-pass effect, which in general can be con-
The experiments were also performed on the first 75 frames of the ‘Miss America’ sequence at 64 kbit/s. The frame rate was 7.5 frames per second. The average SNR;>;; was 33.5 dB. Even though there is less severe motion, the coding results clearly benefit from the ZBC, while also in this case the low-pass error effect is favourable over the DCT blocking errors.
The experiments were also performed on the first 75 frames of the ‘Miss America’ sequence at 64 kbit/s. The frame rate was 7.5 frames per second. The average SNR;>;; was 33.5 dB. Even though there is less severe motion, the coding results clearly benefit from the ZBC, while also in this case the low-pass error effect is favourable over the DCT blocking errors.

Related papers

Motion-compensated 3-D subband coding of video
John Woods

IEEE Transactions on Image Processing, 1999

This paper describes a video coding system based on motion-compensated three-dimensional (3-D) subband/wavelet coding (MC-3DSBC), which can overcome the limits of both 3-D SBC and MC prediction-based coding. In this new system, spatiotemporal subbands are generated by MC temporal analysis and a spatial wavelet transform, and then encoded by 3-D subbandfinite state scalar quantization (3DSB-FSSQ). The rate allocation from the GOP level to each class of subbands is optimized by utilizing the structural property of MC-3DSBC that additive superposition approximately holds for both rate and distortion. The proposed video coding system is applied to several test video clips. Its performance exceeds that of both a known MPEG-1 implementation and a similar subband MC predictive coder while maintaining modest computational complexity and memory size.

View PDFchevron_right
A Variable Resolution Video Codec for Low Bit-Rate Applications
Franco Tommasi

IEEE Journal on Selected Areas in Communications, 1987

Demand for interpersonal video and audio communication at bit rates available on basic and subprimary ISDN accesses is growing and it is expected that service implementation will require large numbers of low-cost tiigh-quality equipment. This paper describes the concept and simulation results of a coding scheme of the hybrid (intraframe transform and temporal DPCM) type whose main feature is the possibility of retaining high resolution for slowly changing pictures. Methods for controlling the spatioltemporal resolutions and simulation results are given. A byproduct of the algorithm allows the reduction of the transmission buffer to negligible'values.

View PDFchevron_right
Video Compression by Three-dimensional Motion-Compensated Subband Coding
Tor Ramstad

1970

Three-dimensional subband coding is an alternative approach to hybrid coding schemes which are used in todays video compression standards. In this work, two possible models for threedimensional subband coding have been tested. In model I the temporal ltering is performed directly on the input sequence. In model II the temporal ltering is performed along the trajectories of the motion. In both models the resulting lowpass and highpass sequences are coded using interframe and intraframe coding. Motion compensation is performed using blockmatching. The coding results are compared to the H263 standard.

View PDFchevron_right
Very low bit-rate video coding with DFD segmentation
Murat Kunt

Signal Processing: Image Communication, 1995

A video coding system for applications requiring very low bit-rate is presented. This coding scheme uses an intraframe coder for the initial frame in the video sequence and subsequent frames are coded using an interframe coding method. A wavelet-based technique is used for intraframe coding. For interframe coding, displaced frame differences (DFD) are computed and coded using a segmentation-based method wherein the displaced frame difference is segmented into active and inactive regions using morphological operators. To meet the very low bit-rate requirements, the motion vectors are processed so as to reduce their contribution to the overall bit-rate. To reduce coding artifacts, a post-processing technique is developed for use at the decoder. Coding performance of the proposed coding scheme is evaluated at 16 kbit/s and 32 kbit/s using luminance component of several typical test sequences at QCIF resolution with a frame rate 8.3 frame/s.

View PDFchevron_right
<title>Subband coding of image sequences using multiple vector quantizers</title>
Vitor Silva

Visual Communications and Image Processing '97, 1997

One efficient way to compress digital images is subband coding. Subband coding using vector quantization could be a competitor to DCT-like image compression schemes. In this paper we will describe an image sequence compression algorithm based on difference image coding techniques, with block motion compensation, difference image segmentation in rectangles using quadtrees, decomposition of rectangles in subbands and vector quantization of the subbands. The vector quantization scheme uses multiple vector quantizers, which yields a better bitrate allocation. The quantization of each subband is performed by 3 different tree structured vector quantizers (TSVQ) at variable tree depths. The Rate-Distortion (R-D) curves of all the rectangles are scanned to get the best global R-D combination. The best combination parameters are coded and used to quantize the subbands of all the rectangles. The results show a slightly better performance of the this scheme in relation to the optimal scalar quantization of subbands (Entropy Limit). The coding speed of this VQ scheme is only 3 times slower than 1 single vector quantization per vector.

View PDFchevron_right
Hierarchical transform and subband coding of video signals
Luc Vandendorpe

Signal Processing: Image Communication, 1992

The new coding scheme introduced here allows the coding of different resolution video signals. The coding is performed in a way that makes it possible to decode either the whole or only some parts of the transmitted format. Transform and subband coding methods are considered. Transform coding is described in the light of the multirate filter bank theory. Coding systems of increasing efficiency are studied. The case of pure intra coding is quite easy to handle. More ett~cient systems include a temporal predictor. Some emphasis will be put on those systems and especially on motion compensated schemes. The method proposed in this paper is the coding of the subbands using for prediction only the information of the same subbands or of the bands of lower order in the previous decoded picture.

View PDFchevron_right
Very Low Bit Rate Coding of Visual Information - A Review
Murat Kunt

In this paper the compression of visual information at very low data rates is discussed. High compression coding of still pictures can be split into three categories: waveform, second generation and fractal coding techniques. Each coding approach i n troduces a di erent artifact at the target bit rates. The objective o f t h e ongoing research is to mask these artifacts as much as possible to the human visual system. Video compression techniques have to deal with data enriched by one more component which is the temporal coordinate. Compression techniques developed for still images can be either generalized for 3 D signals, or a hybrid approach can be de ned based on motion compensation. The video compression techniques can then be classi ed into the following four classes: waveform, object based, model based and fractal coding techniques. The aim of this paper is to provide the readers with an overview of these approaches as well as with a list of references for further information of the details of each technique.

View PDFchevron_right
Content-based rate control scheme for very low bit-rate video coding
Ja-Ling Wu

IEEE Transactions on Consumer Electronics, 1997

In t h e applications of very low bit-rate video coding, t h e quality of t h e encoded pictures always suffers from serious degradation d u e t o limited bandwidth. T h e quantization levels, in typical video coding systems, are adjusted depending only on t h e fullness of t h e o u t p u t buffer while t h e content of t h e video is not taken into account. This explains why t h e picture quality varies saliently in very low bitrate applications.

View PDFchevron_right
Efficient algorithm for very low bit rate embedded image coding
Dr.mohammed Ghanbari

IET Image Processing, 2008

The authors propose an embedded wavelet-based image coding algorithm that exploits both the inter-and intra-subband correlations among the wavelet coefficients. The proposed coding algorithm is based on spatial orientation trees (SOT) in which the basic unit is a block of m  n coefficients in contrast to a single coefficient in the set partitioning in hierarchical trees (SPIHT) algorithm. Each SOT has a root node (a block of m  n coefficients) in the LL-subband with the child and descendent blocks in the high frequency subbands. Thus it fuses the features of both block-and tree-based coding algorithms into a single algorithm. Performance of the proposed method is compared (in terms of rate-distortion performance) with the other state-of-the-art coding algorithms including the JPEG2000 for popular test images. Simulation results show that the proposed algorithm has a better coding efficiency over the other coders at very low bit rates. Also, compared with SPIHT it reduces the elements of the auxiliary lists, thereby reducing the memory requirements. In addition, the encoder of the proposed algorithm is significantly faster than that of the SPIHT, but with a slight increase in its decoder complexity.

View PDFchevron_right
Progressive very low bit rate image coding
Athanassios Skodras

1995

Abstract: This paper describes progressive very low bit rate coding using segmented images. A segmentation algorithm which gives good results for coding purposes is given and it is compared to existing segmentation methods used for the same purpose. Progressive image coding is achieved by a recently proposed set of weakly separable basis functions.

View PDFchevron_right

References (15)

  1. R. Ansari, H.P. Gaggioni and D.J. Le Gall," H DTV coding using a nonrectangular subband decomposition", Proc. SPIE, Vol. 1001, Visual Communications and Image Pro- cessing, November 1988, pp. 821-824.
  2. J.A. Bellisio and Kou-Hu Tzou, "HDTV and the emerging broadband ISDN network", Proc. SPIE, Vol. 1001, Visual Communications and Image Processing, November 1988, pp. 772-786.
  3. J. Biemond, F. Bosveld and R.L. Lagendijk, "Hierarchical subband coding of HDTV in BISDN", Proc. IEEE Inter- nat. Conf. Acoust. Speech Signal Process., Albuquerque, NM, April 1990, pp. 2113-2116.
  4. J. Biemond, B.P. Thieme and D.E. Boekee, "Subband coding of moving video using hierarchical motion estima- tion and vector quantization", Proc. Internat. Workshop on 64 kbit/ s Coding of Moving Video, Hannover, Germany, June 1988.
  5. M. Bierling, "Displacement estimation by hierarchical blockmatching", Proc. SPIE, Vol. 1001, Visual Communi- cations and Image Processing, November 1988, pp. 942- 951.
  6. A. von Brandt, '~Subband coding of videoconference sig- nals using quadrature mirror filters", Proc. lASTED Conf. Appl. Signal Process. Dig. Filt., Paris, France, June 1985, pp. 212-215.
  7. A. yon Brandt, "Subband coding of motion pictures at 2 Mbit/s", Frequenz (in German), Vol. 40, No. 8, August 1986, pp. 190-197.
  8. A. von Brandt, "Motion estimation and subband coding using quadrature mirror filters", Proc. European Signal Process. Conf. ( EUSIPCO ), The Hague, The Netherlands, September 1986, pp. 829-832.
  9. H. Gharavi, "Differential subband coding of video sig- nals", Proc. lnternat. Conf. Acoust. Speech Signal Process. (ICASSP), Glasgow, Scotland, May 1989, pp. 1819-1822.
  10. B. Girod, "The efficiency of motion-compensating predic- tion for hybrid coding of video sequences", IEEE J. Sel. Areas Commun., Vol. SAC-5, No. 7, August 1987, pp. 1140-1154.
  11. J.D. Johnston, "A filter family designed for use in quad- rature mirror filter banks", Proc. lnternat. Conf. Acoust. Speech Signal Process. (ICASSP), Denver, U.S.A., April 1987, pp. 291-294.
  12. H.G. Musmann, P. Pirsch and H.J. Grallert, "Advances in picture coding", Proc. IEEE, Vol. 73, No. 4, April 1985, pp. 523-547.
  13. P.H. Westerink, J. Biemond and D.E. Boekee, "An optimal bit allocation algorithm for subband coding", Proc. Inter- nat. Conf. Acoust. Speech Signal Process. ( ICASSP), New York, April 1988, pp. 757-760.
  14. P.H. Westerink, J. Biemond and D.E. Boekee, "Evaluation of image subband coding schemes", Proc. European Signal Process. Conf. ( EUSIPCO), Grenoble, France, September 1988, pp. 1149-1152.
  15. J.W. Woods and S.D. O'Neil, "Subband coding of images", IEEE Trans. Acoust. Speech Signal Process., Vol. ASSP-34, No. 5, October 1986, pp. 1278-1288. Signal Processing: Image Communication

Related papers

for Subband Coding of Moving Images
Mark Polomski

1993

With emergence of multimedia, CD-ROMS, and high speed networks there is a need for reducing the data rate for transmission and storage of images without severe loss of quality. An image compression system is presented based on motion compensated interpolation using multiframe matching techniques and subband analysis of keyframes and residual (error) frames, followed by vector quantization and arithmetic coding. Multiframe matching differs from conventional block matching in that the intermediate frames are used in addition to the keyframes for motion estimation. Multiframe matching significantly improves the efficiency of motion compensated interpolation. Keyframes and residuals are coded using a quad-split pyramid. The subbands are then coded using variance-normalized, error limited vector quantization. The vector fields, vector quantization codebooks, and look up tables are losslessly coded using arithmetic coding. In computer simulations, bit rates below 2.5 megabits per second are achieved with high quality reconstruction at the receiver. All simulation computations are carried out using integer precision arithmetic.

View PDFchevron_right
Image sequence coding at very low bit rates: a review
Astrid Lundmark

IEEE Transactions on Image Processing, 1994

This paper presents a review of promising techniques for very low bit-rate, below 64 kb/s, image sequence coding. Image sequence coding at such low rates will be a crucial technique in forthcoming visual services, e.g., visual information transmission and storage. A typical application is to transmit moving videophone scenes through the existing analog telephone lines or via a mobile channel. Two types of potential coding techniques are addressed: waveform-based image sequence coding and model-based image sequence coding.

View PDFchevron_right
Motion compensated interpolation for subband coding of moving images
Mark Polomski

1993

With emergence of multimedia, CD-ROMS, and high speed networks there is a need for reducing the data rate for transmission and storage of images without severe loss of quality. An image compression system is presented based on motion compensated interpolation using multiframe matching techniques and subband analysis of keyframes and residual (error) frames, followed by vector quantization and arithmetic coding. Multiframe matching differs from conventional block matching in that the intermediate frames are used in addition to the keyframes for motion estimation. Multiframe matching significantly improves the efficiency of motion compensated interpolation. Keyframes and residuals are coded using a quad-split pyramid. The subbands are then coded using variance-normalized, error limited vector quantization. The vector fields, vector quantization codebooks, and look up tables are losslessly coded using arithmetic coding. In computer simulations, bit rates below 2.5 megabits per second a...

View PDFchevron_right
Subband Coding of Color Images with Dynamic Arithmetic Coder Allocation
Tor Ramstad

This paper presents a subband coder for compression of color still images. The red, green, and blue (RGB) color components are rst transformed to a luminance-chrominance representation, and then each of the three image components are decomposed by an FIR analysis lter bank. The two chrominance components are decimated by a factor of 10:1 or 20:1, depending on the bit rate. The coder is based on dynamic classi cation of blocks of subband samples into 5 classes according to the blockwise mean square value. Each subband sample is quantized in a uniform threshold quantizer, and the quantizer index coded in an arithmetic coder matched to the statistics of the corresponding class. The dynamic classi cation scheme gives a signi cant rate-distortion gain as compared to a scheme without classi cation using the same lter banks and coding methods.

View PDFchevron_right
Subband coding of images
John Woods

IEEE Transactions on Acoustics, Speech, and Signal Processing, 1986

Subband coding has become quite popular for the source encoding of speech. This paper presents a simple yet efficient extension of this concept to the source coding of images. We specify the constraints for a set o f two-dimensional quadrature mirror filters (QMF's) for a particular frequency-domain partition, and show that these constraints are satisfied by a separable combination of one-dimensional QMF's. Bits are then optimally allocated among the subbands to minimize the mean-squared error for DPCM coding of the subbands. Also, an adaptive technique is developed to allocate the bits within each subband by means of a local variance mask. Optimum quantization is employed with quantizers matched to the Laplacian distribution. Subband coded images are presented along with their signal-to-noise ratios (SNR's). The SNR performance of the subband coder is compared to that of the adaptive discrete cosine transform (DCT), vector quantization,-and differential vector quantization for bit rates of 0.67, 1.0, and 2.0 bits per pixel for 256 X 256 monochrome images. The adaptive subband coder has the best SNR performance.

View PDFchevron_right

Related topics

  • Biomedical signal and image proc...
  • Image Sequence
  • Electrical and Electronic Engine...
    • 580 California St., Suite 400
    San Francisco, CA, 94104
    © 2025 Academia. All rights reserved