Network Vector Quantization

A Wyner-Ziv Quantizer design method is introduced when the indices at the output of the encoder are transmitted over a noisy channel. The source encoder is considered as a scalar Lloyd quantizer followed by a binning and an index assignment (BIA) mapping. A modified simulated annealing based algorithm is used for BIA mapping design. A minimax solution for Wyner-Ziv problem under channel mismatch condition is also suggested when the channel is assumed to be binary symmetric channel and no information about the statistic of channel is available except the range of bit error rate. Finally the simulation results are presented which show the effectiveness of the proposed algorithm over other common alternative approaches. These results approve the proposed minimax solution too.

show abstract

“…[2]- [6]. In [5], [6], a Lagrangian cost function of rate and distortion is considered and minimized using the generalized Lloyd algorithm (GLA).…”

Section: Introductionmentioning

confidence: 99%

“…[2]- [6]. In [5], [6], a Lagrangian cost function of rate and distortion is considered and minimized using the generalized Lloyd algorithm (GLA). Also quantizer design for two correlated sources is considered in [7]- [9].…”

Section: Introductionmentioning

confidence: 99%

Scalar Quantizer Design for Noisy Channel with Decoder Side Information

Shamaie

Lahouti

2010

2010 Data Compression Conference

View full text Add to dashboard Cite

show abstract

“…The optimal design of MDSQ was pioneered by Vaishampayan in [3] for the case of two descriptions, through a generalized Lloyd type algorithm [2] which alternatively optimizes the decoder, respectively the encoder while the other component is fixed. The optimal design of multiple description vector quantizers was also considered in [1].…”

Section: Introductionmentioning

confidence: 99%

Speed-Up of Encoder Optimization Step in Multiple Description Scalar Quantizer Design

Dumitrescu

2008

Data Compression Conference (Dcc 2008)

View full text Add to dashboard Cite

The design of optimal multiple description scalar quantizers was pioneered by Vaishampayan with a generalization of Lloyd's algorithm, which alternatively optimizes the decoder, respectively the encoder, while the other component is fixed. We propose an algorithm which speeds up the encoder optimization step from O(N 2 ) to O(N log N ) time complexity, where N is the number of cells in the central partition.

show abstract

“…This design of distributed communication can be utilized to improve the error resilience of conventional quantizers. Multiple description or networked quantization is such a technique [4], [8]- [10] of robust data communications. A multiple-description quantizer consists of two or more so-called side quantizers, each of which separately generates a "partial" or coarse description of the input signal (called side description).…”

Section: Introductionmentioning

confidence: 99%

Optimal Two-Description Scalar Quantizer Design

Dumitrescu

Wu²

2004

Algorithmica

View full text Add to dashboard Cite

Multiple description quantization is a signal compression technique for robust networked multimedia communication. In this paper we consider the problem of optimally quantizing a random variable into two descriptions, with each description being produced by a side quantizer of convex codecells. The optimization objective is to minimize the expected distortion given the probabilities of receiving either and both descriptions. The problem is formulated as one of shortest path in a weighted directed acyclic graph with constraints on the number and types of edges. An O(K 1 K 2 N 3 ) time algorithm for designing the optimal twodescription quantizer is presented, where N is the cardinality of the source alphabet, and K 1 , K 2 are the number of codewords of the two quantizers, respectively. This complexity is reduced to O(K 1 K 2 N 2 ) by exploiting the Monge property of the objective function. Furthermore, if K 1 = K 2 = K and the two descriptions are transmitted through two channels of the same statistics, then the optimal two-description quantizer design problem can be solved in O(K N 2 ) time. Introduction.Quantization is a common technique of compressing multimedia signals such as images, video and audio [5]. Data compression is achieved by quantizing signal samples from a representation of higher resolution to lower resolution so that fewer bits suffice to code each sample of the quantized signal. Optimal quantization falls into the class of resource allocation problems in operational research. The central issue is how to describe a random variable X or a random vector X to the maximum precision (or minimum distortion) possible using a given number of bits. The problem is called optimal scalar or vector quantizer design depending on whether the input is a random variable or a random vector. This paper is restricted to the treatment of scalar quantization.In conventional single-description quantization, a quantized signal is coded and transmitted in a single bit stream through a communication channel. If the channel fails then the reconstruction of the signal will be necessarily interrupted or abandoned at the receiver. Modern IP communication networks, however, offer multiple routes between any two nodes. This design of distributed communication can be utilized to improve the error resilience of conventional quantizers. Multiple description or networked quantization is such a technique [4], [8]-[10] of robust data communications.

show abstract

Network Vector Quantization

Cited by 77 publications

References 44 publications

Scalar Quantizer Design for Noisy Channel with Decoder Side Information

Scalar Quantizer Design for Noisy Channel with Decoder Side Information

Speed-Up of Encoder Optimization Step in Multiple Description Scalar Quantizer Design

Optimal Two-Description Scalar Quantizer Design

Contact Info

Product

Resources

About