应用神经网络的图像分类矢量量化编码

矢量量化作为一种有效的图像数据压缩技术，越来越受到人们的重视。设计矢量量化器的经典算法ＬＢＧ算法，由于运算复杂，从而限制了矢量量化的实用性。本文讨论了应用神经网络实现的基于边缘特征分类的矢量量化技术。它是根据人的视觉系统对图象的边缘的敏感性，应用模式识别技术，在对图像编码前，以边缘为特征对图像内容分类，然后再对每类进行矢量量化。除特征提取是采用离散余弦变换外，图像的分类和矢量量化都是由神经网络完成     

Fuzzy channel-optimized vector quantization

A novel fuzzy clustering algorithm for the design of channel-optimized source coding systems is presented in this letter. The algorithm, termed fuzzy channel-optimized vector quantizer (FCOVQ) design algorithm, optimizes the vector quantizer (VQ) design using a fuzzy clustering process in which the index crossover probabilities imposed by a noisy channel are taken into account. The fuzzy clustering process effectively enhances the robustness of the performance of VQ to channel noise without reducing the quantization accuracy. Numerical results demonstrate that the FCOVQ algorithm outperforms existing VQ algorithms under noisy channel conditions for both Gauss-Markov sources and still image data     

Competitive learning and soft competition for vector quantizerdesign

The authors provide a convergence analysis for the Kohonen learning algorithm (KLA) with respect to vector quantizer (VQ) optimality criteria and introduce a stochastic relaxation technique which produces the global minimum but is computationally expensive. By incorporating the principles of the stochastic approach into the KLA, a deterministic VQ design algorithm, the soft competition scheme (SCS), is introduced. Experimental results are presented where the SCS consistently provided better codebooks than the generalized Lloyd algorithm (GLA), even when the same computation time was used for both algorithms. The SCS may therefore prove to be a valuable alternative to the GLA for VQ design     

一种基于回归神经网络的非线性矢量量化方法

本文提出了一种两级神经网络矢量量化(NNVQ)的方法.该方案中,通过用激励矢量所驱动的网络输出来逼近所要编码的矢量.网络和激励码本都通过优化从而能克服常规算法中的一些病态情况.实验结果表明使用了回归神经网络和优化训练算法的NNVQ的性能好于一般的矢量量化算法.     

基于特征值苑配的变比特率相关矢量量化图像编码算法

该文提出了一种用于图像编码的新颖的变比特率相关矢量量化器。在编码之前,首先计算各码字的四个特征值,然后根据各特征值的升序排列得到相应的四个排序码书。在对当前输入矢量(当前处理图像块)进行编码的过程中,充分考虑当前处理图像块与其相邻图像块之间的相关性以及各码字特征值与该输入矢量特征值之间的匹配性。测试结果表明,该算法与传统矢量量化(VQ)器相比,虽然在编码质量上有少许下降,但降低了比特率并加快了编码速度。     

A fast finite-state algorithm for vector quantizer design

The Linde-Buzo-Gray (LBG) algorithm is usually used to design a codebook for encoding images in vector quantization. In each iteration of this algorithm, one must search the full codebook in order to assign the training vectors to their corresponding codewords. Therefore, the LBG algorithm needs a large computation effort to obtain a good codebook from the training set. The authors propose a finite-state LBG (FSLBG) algorithm for reducing the computation time. Instead of searching the entire codebook, they search only those codewords that are close to the codeword for a training vector in the previous iteration. In general, the number of these possible codewords can be made very small without sacrificing performance. By only searching a small part of the codebook, the computation time is reduced. In experiments, the performance of the FSLBG algorithm in terms of signal-to-noise ratio is very close to that of the LBG algorithm. However, the computation time of the FSLBG algorithm is about 10% of the time required by the LBG algorithm     

A comparison of several vector quantization codebook generationapproaches

A review and a performance comparison of several often-used vector quantization (VQ) codebook generation algorithms are presented. The codebook generation algorithms discussed include the Linde-Buzo-Gray (LBG) binary-splitting algorithm, the pairwise nearest-neighbor algorithm, the simulated annealing algorithm, and the fuzzy c-means clustering analysis algorithm. A new directed-search binary-splitting method which reduces the complexity of the LBG algorithm, is presented. Also, a new initial codebook selection method which can obtain a good initial codebook is presented. By using this initial codebook selection algorithm, the overall LBG codebook generation time can be reduced by a factor of 1.5-2.     

Sample-adaptive product quantization: asymptotic analysis andexamples

Vector quantization (VQ) is an efficient data compression technique for low bit rate applications. However the major disadvantage of VQ is that its encoding complexity increases dramatically with bit rate and vector dimension. Even though one can use a modified VQ, such as the tree-structured VQ, to reduce the encoding complexity, it is practically infeasible to implement such a VQ at a high bit rate or for large vector dimensions because of the huge memory requirement for its codebook and for the very large training sequence requirement. To overcome this difficulty, a structurally constrained VQ called the sample-adaptive product quantizer (SAPQ) has recently been proposed. We extensively study the SAPQ that is based on scalar quantizers in order to exploit the simplicity of scalar quantization. Through an asymptotic distortion result, we discuss the achievable performance and the relationship between distortion and encoding complexity. We illustrate that even when SAPQ is based on scalar quantizers, it can provide VQ-level performance. We also provide numerical results that show a 2-3 dB improvement over the Lloyd-Max (1982, 1960) quantizers for data rates above 4 b/point     

A complexity reduction technique for image vector quantization

A technique for reducing the complexity of spatial-domain image vector quantization (VQ) is proposed. The conventional spatial domain distortion measure is replaced by a transform domain subspace distortion measure. Due to the energy compaction properties of image transforms, the dimensionality of the subspace distortion measure can be reduced drastically without significantly affecting the performance of the new quantizer. A modified LBG algorithm incorporating the new distortion measure is proposed. Unlike conventional transform domain VQ, the codevector dimension is not reduced and a better image quality is guaranteed. The performance and design considerations of a real-time image encoder using the techniques are investigated. Compared with spatial domain a speed up in both codebook design time and search time is obtained for mean residual VQ, and the size of fast RAM is reduced by a factor of four. Degradation of image quality is less than 0.4 dB in PSNR.     

Variable dimension VQ encoding and codebook design

A variable dimension vector quantizer (VDVQ) has codewords of unequal dimensions. Here, a trellis-based sequential optimal VDVQ encoding algorithm is proposed. Also, a VDVQ codebook design algorithm based on splitting a node with equal or reduced dimensions is proposed that does not require any codebook parameter to be prespecified unlike known schemes. The VDVQ system is shown to outperform a few known VQ systems for AR(1) sources     

An improved vector quantization method using deep neural network

To address the challenging problem of vector quantization (VQ) for high dimensional vector using large coding bits, this work proposes a novel deep neural network (DNN) based VQ method. This method uses a k-means based vector quantizer as an encoder and a DNN as a decoder. The decoder is initialized by the decoder network of deep auto-encoder, fed with the codes provided by the k-means based vector quantizer, and trained to minimize the coding error of VQ system. Experiments on speech spectrogram coding demonstrate that, compared with the k-means based method and a recently introduced DNN-based method, the proposed method significantly reduces the coding error. Furthermore, in the experiments of coding multi-frame speech spectrogram, the proposed method achieves about 11% relative gain over the k-means based method in terms of segmental signal to noise ratio (SegSNR).     

误差敏感竞争性学习算法

本文基于等误差准则提出了一种适用于矢量量化技术的新型码书设计算法。实验表明此算法优于现存算法。为解决初始码书赋值问题，本文提出了自生成自组织神经网络方法。实验表明此算法加速了算法的收敛速度，提高了算法的性能     

Performance comparisons between PCA-EA-LBG and PCA-LBG-EA approaches in VQ codebook generation for image compression

The aim of this study is to generate vector quantisation (VQ) codebooks by integrating principle component analysis (PCA) algorithm, Linde-Buzo-Gray (LBG) algorithm, and evolutionary algorithms (EAs). The EAs include genetic algorithm (GA), particle swarm optimisation (PSO), honey bee mating optimisation (HBMO), and firefly algorithm (FF). The study is to provide performance comparisons between PCA-EA-LBG and PCA-LBG-EA approaches. The PCA-EA-LBG approaches contain PCA-GA-LBG, PCA-PSO-LBG, PCA-HBMO-LBG, and PCA-FF-LBG, while the PCA-LBG-EA approaches contain PCA-LBG, PCA-LBG-GA, PCA-LBG-PSO, PCA-LBG-HBMO, and PCA-LBG-FF. All training vectors of test images are grouped according to PCA. The PCA-EA-LBG used the vectors grouped by PCA as initial individuals, and the best solution gained by the EAs was given for LBG to discover a codebook. The PCA-LBG approach is to use the PCA to select vectors as initial individuals for LBG to find a codebook. The PCA-LBG-EA used the final result of PCA-LBG as an initial individual for EAs to find a codebook. The search schemes in PCA-EA-LBG first used global search and then applied local search skill, while in PCA-LBG-EA first used local search and then employed global search skill. The results verify that the PCA-EA-LBG indeed gain superior results compared to the PCA-LBG-EA, because the PCA-EA-LBG explores a global area to find a solution, and then exploits a better one from the local area of the solution. Furthermore the proposed PCA-EA-LBG approaches in designing VQ codebooks outperform existing approaches shown in the literature.     

矢量量化在语音波形编码中的应用

矢量量化(VQ)技术是近几年发展起来的一种高效数据压缩技术.本文介绍了VQ技术的发展历史、现状和它的基本原理,较为详细地讨论了基本矢量量化器的实用设计方法——LBG算法,并对原有的LBG算法进行了改进,给出了实验结果.     

基于改进禁止搜索算法的矢量量化码书设计

本文提出了基于改进禁止搜索(TS)算法的矢量量化(VQ)码书设计方法.禁止搜索算法的关键是如何定义一个解以及如何在当前解的基础上生成邻域解.由于码书设计的两个优化准则是最邻近条件和聚类质心条件,本文提出了两种禁止搜索算法的解描述方案,其相应算法分别叫基于码书的禁止搜索(CB-TS)算法和基于聚类划分的禁止搜索(PB-TS)算法.为了提高禁止搜索算法的性能,文中在禁止搜索算法中融入了模拟退火(SA)机制.为了进一步提高码书性能,文中还将码书设计的传统LBG算法融入禁止搜索算法中.结果表明,基于禁止搜索的两种码书设计方案所生成的码书性能都比LBG算法有明显提高.     

Design of sample adaptive product quantizers for noisy channels

Channel-optimized vector quantization (COVQ) has proven to be an effective joint source-channel coding technique that makes the underlying quantizer robust to channel noise. Unfortunately, COVQ retains the high encoding complexity of the standard vector quantizer (VQ) for medium-to-high quantization dimensions and moderate-to-good channel conditions. A technique called sample adaptive product quantization (SAPQ) was recently introduced by Kim and Shroff to reduce the complexity of the VQ while achieving comparable distortions. In this letter, we generalize the design of SAPQ for the case of memoryless noisy channels by optimizing the quantizer with respect to both source and channel statistics. Numerical results demonstrate that the channel-optimized SAPQ (COSAPQ) achieves comparable performance to the COVQ (within 0.2 dB), while maintaining considerably lower encoding complexity (up to half of that of COVQ) and storage requirements. Robustness of the COSAPQ system against channel mismatch is also examined.     

Training ratio and comparison of trained vector quantizers

The vector quantizer (VQ) codebook is usually designed by clustering a training sequence (TS) drawn from the underlying distribution function. In order to cluster a TS, we may use the K-means algorithm (generalized Lloyd (1982) algorithm) or the self-organizing map algorithm. In this paper, a survey of trained VQ performance is conducted to study the effect of the training ratio on training quantizers. The training ratio, which is defined by the ratio of the TS size to the codebook size, is dependent on the VQ structure. Hence, different VQs may show different training properties, even though the VQs are designed for the same TS. A numerical comparison of trained VQs is then conducted in conjunction with deriving their training ratios. Through the comparison, it is shown that structured VQs can achieve better performance than the full-search scheme if the codebooks are trained by a finite TS. Further, we can derive a design or comparison guideline that maintains equal training ratios in training different VQs.     

Vector quantization using tree-structured self-organizing featuremaps

In this paper, we propose a binary-tree structure neural network model suitable for structured clustering. During and after training, the centroids of the clusters in this model always form a binary tree in the input pattern space. This model is used to design tree search vector quantization codebooks for image coding. Simulation results show that the acquired codebook not only produces better-quality images but also achieves a higher compression ratio than conventional tree search vector quantization. When source coding is applied after VQ, the new model performs better than the generalized Lloyd algorithm in terms of distortion, bits per pixel, and encoding complexity for low-detail and medium-detail images     

Adaptive vector quantization using a self-development neuralnetwork

A neural network model, called SPAN (space partition network), is presented. This model differs from most of the currently seen neural networks in that it allows a network to adapt its structure by adding neurons, killing neurons, and modifying the structural relationships between neurons in the network. An adaptive vector quantization source-coding system based on SPAN is proposed. The major advantage of using SPAN as the codebook of a vector quantizer is that SPAN can capture the local context of the source signal space and map onto a lattice structure. A fast codebook-searching method utilizing the local context of the lattice is proposed, and a coding scheme, called the path coding method, for eliminating the correlation buried in the source sequence is introduced. The performance of the proposed coder is compared to an LBG (Y. Linde, A. Buzo, and R.M. Gray, 1980) coder on synthesized Gauss-Markov sources. Simulation results show that, without using the path coding method, SPAN yields performance similar to an LBG coder; however, if the path coding method is used, SPAN displays a much better performance than the LBG for highly correlated signal sources     

Adaptive fast encoding algorithm for vector quantisation

A new fast encoding algorithm is developed for vector quantisation (VQ). The proposed algorithm takes the topological structure of the codebook into account and controls the search space adaptively. Simulation results demonstrate that while the performance of the proposed algorithm is the same as that using an exhaustive search, it is more efficient than some of the other related fast VQ encoding algorithms