P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods. Subband coding of images using vector quantization. IEEE Trans. Commun., 36:713--719, June 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....VQ [HH88] which during prediction aim to reduce the maximum, would reduce the codebook storage further. Alternatively, mean residual VQ can be viewed as a two band sub band coder where the high band is vector quantized. More than two bands are shown to be desirable in sub band coding [WBBW88] With more bands, the variance of the sub band coded image decreases. This will also make VPR compress more. For example with the maximum possible number of sub bands, the coder will be performing a discrete cosine transform of the whole image, where the variance of most coefficients is quite ....

Peter H. Westerink, Dick E. Boekee, Jan Biemond, and John W. Woods. Subband coding of images using vector quantization. IEEE Trans., COM-36:713-- 719, June 1988.

.... Subband coding was introduced in the context of speech coding in 1976 by Crochiere et al. 113] The extension of subband filtering from 1 D to 2 D was made by Vet GRAY AND NEUHOFF: QUANTIZATION 43 terli [515] and 2 D subband filtering was first applied to image coding by Woods et al. 541] [527], 540] Early wavelet coding techniques emphasized scalar or lattice vector quantization [12] 13] 130] 463] 14] 30] 185] and other vector quantization techniques have also been applied to wavelet coe#cients, including tree encoding [366] residual vector quantization [295] and other ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods, "Subband coding of images using vector quantization," IEEE Trans. Comm., vol. 36, pp. 713--719, June 1988.

....there was negligible interband correlation and except for the lowest frequency subband, the intraband correlation was small. Thus, DPCM is commonly used to code the lowest frequency band [2, 3] while zero memory scalar or vector quantization have usually been applied to code the higher subbands [1, 2, 4]. However, inspection of the subbands reveals that most of the energy is concentrated in areas which corresponds to edge activity in the original image. Some of the approaches which exist to exploit the local characteristics of the imagery are spatially adapting filter banks [5] spatially ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods, "Subband coding of images using vector quantization, " IEEE Trans. Commun., vol. COM-36, pp. 713--719, June 1988.

.... Subband coding was introduced in the context of speech coding in 1976 by Crochiere et al. 113] The extension of subband filtering from 1 D to 2 D was made by Vet GRAY AND NEUHOFF: QUANTIZATION 43 terli [515] and 2 D subband filtering was first applied to image coding by Woods et al. 541] [527], 540] Early wavelet coding techniques emphasized scalar or lattice vector quantization [12] 13] 130] 463] 14] 30] 185] and other vector quantization techniques have also been applied to wavelet coefficients, including tree encoding [366] residual vector quantization [295] and ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods, "Subband coding of images using vector quantization," IEEE Trans. Comm., vol. 36, pp. 713--719, June 1988.

.... on wavelets and subband coding [365] Subband coding was introduced in the context of speech coding in 1976 by Crochiere et al. 89] The extension of subband filtering from 1 D to 2 D was made by Vetterli [364] and 2 D subband filtering was first applied to image coding by Woods et al. [380, 371, 381]. Early wavelet coding techniques emphasized scalar or lattice vector quantization [10, 11, 100, 321, 12, 27, 140] and other vector quantization techniques have also been applied to wavelet coefficients, including tree encoding [261] residual vector quantization [212] and other methods [85] A ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods, "Subband coding of images using vector quantization," IEEE Trans. Comm., vol. COM-36, pp. 713--719, June 1988.

.... This quantization choice might seem oversimplified, but as was evidenced by our experiments and also stated in [37] using more sophisticated quantization and optimization schemes such as optimal bit allocation among bands [38] optimal non uniform scalar quantization [39] or vector quantization [40, 41] was not justifiable when the aforementioned advanced data representation and organization strategies were used since the performance gain, if any, was only marginal. 2.1.2 Connected Component Analysis For natural images, the majority of pixels belong to either homogeneous regions or texture ....

P.H. Westernik, D.E. Boekee, J. Biemond, and J.W. Woods, "Subband coding of images using vector quantization," IEEE Transactions on Communications, vol. 36, pp. 713--719, June 1988.

....of good filter banks for image subband coders has been done. Largely, the filter bank portion of the coder has been postulated and the optimization of the total coder has focused on the scheme for the bit efficient representation of the subband signals. Examples of this approach can be found in [67, 64, 56, 19, 10, 21, 29]. In [56] an empirically derived perceptual noise masking model was applied to estimate a target noise level for each subband pel to be quantized. This was then used to set the quantization levels in a DPCM scheme. It was reported that nearly transparent image quality resulted at bit rates ....

....the time and frequency responses of the obtained filter banks. 2.3 Classification of image defects At high compression ratios, the image transform subband data are coarsely approximated. As a result of the associated quantization noise, the decoded image will contain distortions. Westerink et al. [64] made a quantitative investigation of quantization noise for a 2 channel, QMF filter bank system. Using the gain plus additive noise quantizer model [28] they defined 4 noise types: Signal related noise (i.e. a filtered version of the original signal) aliasing noise due to nonperfect ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods. Subband coding of images using vector quantization. IEEE Trans. Commun., COM-36(6):713-- 719, June 1988.

....band subimage to achieve good visual reconstruction quality. This technique is free from blocking artifacts, mosquito noise which exist in block transform coding schemes like JPEG when high compression ratio is required. So far there are lots of algorithms proposed for wavelet transform coding[1,3,4]. They can be divided into two generations. In the first generation[1,3] the wavelet based coding were designed to exploit the decomposition of image into statistical independent subimages. The coding gain was obtained by optimizing the quantizer with efficient bit allocation. However, the ....

....is free from blocking artifacts, mosquito noise which exist in block transform coding schemes like JPEG when high compression ratio is required. So far there are lots of algorithms proposed for wavelet transform coding[1,3,4] They can be divided into two generations. In the first generation[1,3], the wavelet based coding were designed to exploit the decomposition of image into statistical independent subimages. The coding gain was obtained by optimizing the quantizer with efficient bit allocation. However, the inter band dependencies among subbands were not examined. The new generation ....

[Article contains additional citation context not shown here]

P.H.Westerink, D.E.Boekee, J.Biemond, J.W.Woods, "Subband Coding of Image Using Vector Quantization", IEEE Trans. Comm., Vol. 36, No. 6, pp. 713-719, June 1988.

....to the conventional DCT with scalar quantization, the wavelet and WP image decompositions call for new powerful quantization strategies to exploit the underlining space frequency image characterizations. Early efforts in wavelet image coding, which was disguised under the name of subband coding [22, 23, 24, 25, 26], failed to respond to this new need by only recognizing the frequency grouping of coefficients in the same band, and invoking the conventional bit allocation schemes [22] with marginal coding gain over DCT based methods, until Lewis and Knowles [27] came up with the zerotree data structure, ....

P. Westerink, D. Boekee, J. Biemond, and J. W. Wood, "Subband coding of images using vector quantization," IEEE Trans. Communications, vol. 36, pp. 713-719, June 1988.

....costs. 7 Compression of Wavelet decomposed Images Lossy image compression is a real world information processing problem which is well suited to study the influence of different complexity measures on the performance of vector quantization algorithms. Image compression based on orthogonal [45, 36, 15] or nonorthogonal [14, 44] wavelet decompositions has witnessed increasing popularity in recent years. The wavelet data format is supposedly optimal for natural image representation since the coefficients are statistically independent if we average over a large set of natural images [21, 44] ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J.W. Woods. Subband coding of images using vector quantization. IEEE Transactions on Communications, 36:713--719, 1988.

....there was negligible interband correlation and except for the lowest frequency subband, the intraband correlation was small. Thus, DPCM is commonly used to code the lowest frequency band [2, 3] while zero memory scalar or vector quantization have usually been applied to code the higher subbands [1, 2, 4]. However, inspection of the subbands reveals that most of the energy is concentrated in areas which corresponds to edge activity in the original image. Some of the approaches which exist to exploit the local characteristics of the imagery are spatially adapting filter banks [5] spatially ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods, "Subband coding of images using vector quantization, " IEEE Trans. Commun., vol. COM-36, pp. 713--719, June 1988.

....used an extension of Chen and Pratt s [49] combined Huffman and run length coding scheme to code QMF pyramids. Adelson et al. have used both separable and hexagonal QMF pyramids for image coding [27, 36, 39] Mallat [3] used filters derived from wavelet theory to code images. Westerink et al. [50] have used vector quantization for subband coding of images. In figures 4.22 and 4.23, we give examples of data compression of the 256 Theta256 Lena image using a separable 9 tap QMF bank, a 3 tap asymmetric filter bank (described in section 4.4.2) and a hexagonal 4 ring QMF bank. In all ....

Peter H. Westerink, Dick E. Boekee, Jan Biemond, and John W. Woods. Subband coding of images using vector quantization. IEEE Trans. COM, COM-36(6):713--719, June 1988.

....such subband histograms are consistent with probability densities having the form C exp u .7 , where the exponent .7 would be 2 if the Gaussian case applied. In fact such Generalized Gaussian models have been long used to model subband coe#cients in the compression literature (e.g. [100]) Field [36] investigated the fourth order cumulant structure of images and showed that it was significantly nonzero. This is far out of line with the Gaussian model, in which all cumulants of order three and higher vanish. In later work, Field [37] proposed that wavelet transforms of images ....

P. H. Westerink and J. Biemond and D. E. Boekee and J. W. Woods, "Subband Coding of Images Using Vector Quantization", IEEE Tr. on Comm., Vol. 36, No. pp. 713-719, June 1988.

....band subimage to achieve good visual reconstruction quality. This technique is free from blocking artifacts, mosquito noise which exist in block transform coding schemes like JPEG when high compression ratio is required. So far there are lots of algorithms proposed for wavelet transform coding[1,2,3,4].They can be divided into two generations. In the first generation[1,2] the wavelet based coding were designed to exploit the decomposition of image into statistical independent subimages. The coding gain was obtained by optimizing the quantizer with efficient bit allocation. However, the ....

....is free from blocking artifacts, mosquito noise which exist in block transform coding schemes like JPEG when high compression ratio is required. So far there are lots of algorithms proposed for wavelet transform coding[1,2,3,4] They can be divided into two generations. In the first generation[1,2], the wavelet based coding were designed to exploit the decomposition of image into statistical independent subimages. The coding gain was obtained by optimizing the quantizer with efficient bit allocation. However, the inter band dependencies among subbands were not examined. The new generation ....

P.H.Westerink, D.E.Boekee, J.Biemond, J.W.Woods, "Subband Coding of Image Using Vector Quantization ", IEEE Trans. Comm., Vol. 36, No. 6, pp. 713-719, June 1988.

....transform coding [5] composite source models (e.g. classified VQ [6, 7] image segmentation models, etc. The performance of each of these approaches have been optimized through careful consideration of quantization strategy (e.g. optimal bit allocation among bands [8] vector quantizers [9, 10], optimal non uniform scalar quantizers [11] trellis coded quantization [12] etc. It has been a general trend that more complex models are needed to achieve improved coding efficiency. One could easily draw the conclusion that the promising directions in image coding involve higher modeling ....

P. H. Westerink, D. E. Boekee, J. Biemond and J. W. Wood, "Subband Coding of Images Using Vector Quantization", IEEE Trans. on Communications, Vol. 36, pp. 713-719, June, 1988.

....to be of less perceptual importance than the horizontal or vertical. F NaN P F NaN Pq F NaN P F NaN P F NaN P F NaN Pq Figure 8: Same orientation crossband vector from octave band subbands Crossband vectors from uniform subbands: The first study of crossband VQs was by Westerink et al. [155]. After splitting the image into 16 uniform subbands by 32 tap separable QMFs, 16 D vectors were formed as in Figure 7. Full search unstructured VQs were designed using the GLA. The paper also analyzed the coding gain of VQ over SQ, and briefly examined the effect of training sequence size on VQ ....

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods. Subband coding of images using vector quantization. IEEE Trans. Comm., COM-36:713--719, June 1988.

No context found.

P. H. Westerink, D. E. Boekee, J. Biemond, and J. W. Woods. Subband coding of images using vector quantization. IEEE Trans. Commun., 36:713--719, June 1988.

No context found.

P. H. Westerink, D.E. Boekee, J. Biemond, and J.W. Woods. Subband coding of images using vector quantization. IEEE Trans. on Comm., 36:713--719, 1989.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC