Tree-structured Vector Quantization Research Articles

Landmarking for Navigational Streaming of Stored High-Dimensional Media

Modern media data such as 360° videos and light field (LF) images are typically captured in much higher dimensions than the observers’ visual displays. To efficiently browse high-dimensional media, a navigational streaming model is considered: a client navigates the media space by dictating a navigation path to a server, who in response transmits the corresponding pre-encoded media data units (MDU) to the client one-by-one in sequence. Assuming that the MDU quality is pre-chosen and fixed, the problem resides in selecting and storing redundant representations of MDUs at the server in order to best trade off storage and transmission costs, while enabling adequate user’s random access. We address this problem with a landmark-based MDU optimization framework. The media space is divided into neighborhoods, each containing one landmark (a chosen MDU). MDUs in a neighborhood use the associated landmark as a predictor for inter-coding. Thus, for any MDU transition within the same neighborhood, only one inter-coded MDU transmission is required when the landmark resides in the decoder buffer. It results in lower transmission cost and enables navigational random access. To optimize an MDU structure, we employ tree-structured vector quantizer (TSVQ) to first optimize landmark locations, then iteratively add P-MDUs as refinements using a fast branch-and-bound technique. Taking interactive LF images and viewport adaptive 360° images as illustrative applications, and I-, P- and previously proposed merge frames to intra- and inter-code MDUs, we show experimentally that landmarked MDU structures can noticeably reduce the expected transmission cost compared with MDU structures without landmarks.

Entropy-Based Semi-Fragile Watermarking of Remote Sensing Images in the Wavelet Domain

This article presents a semi-fragile image tampering detection method for multi-band images. In the proposed scheme, a mark is embedded into remote sensing images, which have multiple frequential values for each pixel, applying tree-structured vector quantization. The mark is not embedded into each frequency band separately, but all the spectral values (known as signature) are used. The mark is embedded in the signature as a means to detect if the original image has been forged. The image is partitioned into three-dimensional blocks with varying sizes. The size of these blocks and the embedded mark is determined by the entropy of each region. The image blocks contain areas that have similar pixel values and represent smooth regions in multispectral or hyperspectral images. Each block is first transformed using the discrete wavelet transform. Then, a tree-structured vector quantizer (TSVQ) is constructed from the low-frequency region of each block. An iterative algorithm is applied to the generated trees until the resulting tree fulfils a requisite criterion. More precisely, the TSVQ tree that matches a particular value of entropy and provides a near-optimal value according to Shannon’s rate-distortion function is selected. The proposed method is shown to be able to preserve the embedded mark under lossy compression (above a given threshold) but, at the same time, it detects possibly forged blocks and their positions in the whole image. Experimental results show how the scheme can be applied to detect forgery attacks, and JPEG2000 compression of the images can be applied without removing the authentication mark. The scheme is also compared to other works in the literature.

Vector quantization: a review

Vector quantization (VQ) is a very effective way to save bandwidth and storage for speech coding and image coding. Traditional vector quantization methods can be divided into mainly seven types, tree-structured VQ, direct sum VQ, Cartesian product VQ, lattice VQ, classified VQ, feedback VQ, and fuzzy VQ, according to their codebook generation procedures. Over the past decade, quantization-based approximate nearest neighbor (ANN) search has been developing very fast and many methods have emerged for searching images with binary codes in the memory for large-scale datasets. Their most impressive characteristics are the use of multiple codebooks. This leads to the appearance of two kinds of codebook: the linear combination codebook and the joint codebook. This may be a trend for the future. However, these methods are just finding a balance among speed, accuracy, and memory consumption for ANN search, and sometimes one of these three suffers. So, finding a vector quantization method that can strike a balance between speed and accuracy and consume moderately sized memory, is still a problem requiring study.

Prior-Based Quantization Bin Matching for Cloud Storage of JPEG Images.

Millions of user-generated images are uploaded to social media sites like Facebook daily, which translate to a large storage cost. However, there exists an asymmetry in upload and download data: only a fraction of the uploaded images are subsequently retrieved for viewing. In this paper, we propose a cloud storage system that reduces the storage cost of all uploaded JPEG photos, at the expense of a controlled increase in computation mainly during download of requested image subset. Specifically, the system first selectively re-encodes code blocks of uploaded JPEG images using coarser quantization parameters for smaller storage sizes. Then during download, the system exploits known signal priors-sparsity prior and graph-signal smoothness prior-for reverse mapping to recover original fine quantization bin indices, with either deterministic guarantee (lossless mode) or statistical guarantee (near-lossless mode). For fast reverse mapping, we use small dictionaries and sparse graphs that are tailored for specific clusters of similar blocks, which are classified via tree-structured vector quantizer. During image upload, cluster indices identifying the appropriate dictionaries and graphs for the re-quantized blocks are encoded as side information using a differential distributed source coding scheme to facilitate reverse mapping during image download. Experimental results show that our system can reap significant storage savings (up to 12.05%) at roughly the same image PSNR (within 0.18 dB).

Accelerating simulation for the multiple-point statistics algorithm using vector quantization.

Multiple-point statistics (MPS) is a prominent algorithm to simulate categorical variables based on a sequential simulation procedure. Assuming training images (TIs) as prior conceptual models, MPS extracts patterns from TIs using a template and records their occurrences in a database. However, complex patterns increase the size of the database and require considerable time to retrieve the desired elements. In order to speed up simulation and improve simulation quality over state-of-the-art MPS methods, we propose an accelerating simulation for MPS using vector quantization (VQ), called VQ-MPS. First, a variable representation is presented to make categorical variables applicable for vector quantization. Second, we adopt a tree-structured VQ to compress the database so that stationary simulations are realized. Finally, a transformed template and classified VQ are used to address nonstationarity. A two-dimensional (2D) stationary channelized reservoir image is used to validate the proposed VQ-MPS. In comparison with several existing MPS programs, our method exhibits significantly better performance in terms of computational time, pattern reproductions, and spatial uncertainty. Further demonstrations consist of a 2D four facies simulation, two 2D nonstationary channel simulations, and a three-dimensional (3D) rock simulation. The results reveal that our proposed method is also capable of solving multifacies, nonstationarity, and 3D simulations based on 2D TIs.

An efficient low bit-rate compression scheme of acoustic features for distributed speech recognition

Due to the limited network bandwidth, a noise robust low bit-rate compression scheme of Mel frequency cepstral coefficients (MFCCs) is desired for distributed speech recognition (DSR) services. In this paper, we present an efficient MFCCs compression method based on weighted least squares (W-LS) polynomial approximation through the exploitation of the high correlation across consecutive MFCC frames. Polynomial coefficients are quantized by designing a tree structured vector quantization (TSVQ) based scheme. Recognition experiments are conducted on the noisy Aurora-2 database, under both clean and multi-condition training modes. The results show that the proposed W-LS encoder slightly exceeds the ETSI advanced front-end (ETSI-AFE) baseline system for bit-rates ranging from 1400 bps to 1925 bps under clean training mode. However, a negligible degradation is observed in case of multi-condition training mode (around 0.6% and 0.2% at 1400 bps and 1925 bps, respectively). Furthermore, the obtained performance generally outperforms the ETSI-AFE source encoder at 4400 bps under clean training and provides similar performance, at 1925 bps, under multi-condition training.

Spectrum Map and Its Application in Resource Management in Cognitive Radio Networks

Measurements on radio spectrum usage have revealed an abundance of under-utilized bands of spectrum that belong to primary (licensed) networks. Prior knowledge about the occupancy of such bands and the expected achievable performance on those bands can help secondary (unlicensed) networks to devise effective strategies to improve utilization. Such prior spatio-temporal spectrum usage statistics can either be obtained from a database that is maintained by the primary networks or could be measured by customized sensors deployed by the secondary networks. In this paper, we use Shepard's interpolation technique to estimate a spatial distribution of spectrum usage over a region of interest, which we call the spectrum map. The interpolation is achieved by intelligently fusing the data shared by the the secondary nodes considering their mutual distances and spatial orientation with each other. The obtained map is a two-dimensional (2-D) interpolation function that is continuously differentiable and passes through all the spectrum usage values recorded at arbitrary locations; thus providing a reference for primary occupancy in that region. For determining the optimal locations for sensing primary activity, we use an iterative clustering technique that uses tree structured vector quantization. We use the spectrum map to estimate different radio and network performance metrics like channel capacity, network throughput, and spectral efficiency. As a comprehensive case study, we demonstrate how the spectrum map can be used for efficient resource allocation in TV white space. In particular, we consider an IEEE 802.22-based WRAN and show how the rendezvous probability can be improved for better radio resource allocation, thereby increasing the secondary spectrum utilization.

Automatic Human Mocap Data Classification

Automatic classification of human motion capture (mocap) data has many commercial, biomechanical, and medical applications and is the principal focus of this paper. First, we propose a multi-resolution string representation scheme based on the tree-structured vector quantization (TSVQ) to transform the time-series of human poses into codeword sequences. Then, we take the temporal variations of human poses into account via codeword sequence matching. Furthermore, we develop a family of pose-histogram-based classifiers to examine the spatial distribution of human poses. We analyze the performance of the temporal and spatial classifiers separately. To achieve a higher classification rate, we merge their decisions and soft scores using novel fusion methods. The proposed fusion solutions are tested on a wide variety of sequences from the CMU mocap database using five-fold cross validation, and a correct classification rate of 99.6% is achieved.

An efficient tree‐structured vector quantization using dynamic triangular inequality elimination

As applied to a vector quantization (VQ) codebook search, a combined version of a dynamic triangular inequality elimination (DTIE) and a tree‐structured VQ (TSVQ) algorithm, designated as the DTIE–TSVQ approach, is presented in this letter as an efficient way to reach the aim of search performance improvement by successive updating of the search scope and reduced search load through the DTIE algorithm. In this manner, this proposal features the combined advantages of a TIE and a TSVQ algorithm such that 100% search accuracy is rendered together with a remarkable reduction in computational load. At the end of this work, this proposal is validated as a superior algorithm over TIE and conventional TSVQ algorithms by a high computational load saving up to 97.45% when dealing with line spectral frequency (LSF) quantization in a G.729 speech codec and image VQ encodings. © 2014 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

An Efficient Information Hiding Scheme Based on Closest Paired Tree Structure Vector Quantization

Information hiding schemes based on vector quantization (VQ) usually require lengthy VQ encoding and decoding processes. In this paper, we propose an efficient information hiding method based on closest paired tree structure vector quantization (CPTSVQ). The simulation result shows that the execution time of the proposed scheme is much shorter than that attained by previous approaches.

An Improvement to Tree-Structured Vector Quantization

This study proposes an improvement to tree-structured vector quantization (TSVQ). The selected codeword of TSVQ is employed as the reference codeword. The dynamic triangular inequality elimination (DTIE) is employed to improve the performance of TSVQ. In the DTIE-TSVQ approach, the selected codeword and SNR are identical to that of full-search vector quantization (FSVQ), and the computational saving is better than that of TIE-VQ. The experimental results confirm that DTIE-TSVQ outperforms TIE-VQ, TSVQ.

Identification of multifunctional cytotoxic T-cell subsets as immune correlates with clinical outcomes in a phase II study of AGS-003, an autologous dendritic cell-based therapy administered to patients with newly diagnosed, metastatic RCC.

80 Background: AGS-003 is an autologous dendritic cell (DC) immunotherapy prepared from matured monocyte-derived DC co-electroporated with the subject’s own amplified tumor RNA and synthetic CD40L RNA. The mechanism of action (MOA) of AGS-003 was evaluated in combination with sunitinib for treatment of advanced renal cell carcinoma (RCC) in AGS-003-006, an open label phase II trial including subjects with newly diagnosed, unfavorable-risk, metastatic clear cell RCC. The goal of the immune monitoring platform is to identify unique cytotoxic T-cell (CTL) signatures that correlate with clinical outcome in subjects receiving AGS-003 in combination with sunitinib. Methods: Multiparametric flow cytometry was used to identify tumor-reactive CTL subsets induced by AGS-003 based on combinatorial expression patterns of surface markers CD28, CD45RA, CD27 and CCR7. Moreover, further partitioning of each CTL subset identified combinatorial expression patterns of Markers of Immune Function (MIFs) defined as cytokines (IFN-γ TNF-α, IL-2), cytolytic markers (Granzyme b, CD107) and proliferation. Correlates of CTL signatures with clinical outcome were analyzed using an adaptation of a binary tree-structured vector quantization (BTSVQ) approach, originally developed to cluster and visualize large microarray data sets. The BTSVQ approach implements a two-way unsupervised clustering that allows a subject’s CTL signature to be mapped based on both surface marker and MIFs expression patterns to identify unique clustering patterns linked to clinical outcome. Results: Data analysis identified a unique CTL signature (CD28+/CCR7+/CD45RA- phenotype) displaying a broad MIFs profile as a statistically significant correlate to PFS and OS in patients treated with AGS-003. Conclusions: These results support the intended MOA of AGS-003 in vivo, as the induction of anti-tumor central and effector memory CTL responses. These data warrant further immunological evaluation of AGS-003 in the randomized phase III ADAPT study using AGS-003 in combination with standard treatment in RCC subjects.

A novel semi-fragile forensic watermarking scheme for remote sensing images

A semi-fragile watermarking scheme for multiple band images is presented. We propose to embed a mark into remote-sensing images applying a tree-structured vector quantization approach to the pixel signatures, instead of processing each band separately. The signature of the multi- or hyperspectral image is used to embed the mark into it in order to detect any significant modification of the original image. The image is segmented into three-dimensional blocks and a tree-structured vector quantizer (TSVQ) is built for each block. These trees are manipulated using an iterative algorithm until the resulting block satisfies a required criterion which establishes the embedded mark. The method is shown to be able to preserve the mark under lossy compression (above a given threshold) but, at the same time, it detects possibly forged blocks and their position in the whole image.

Fine granularity scalable speech coding using embedded tree-structured vector quantization

This paper proposes an efficient codebook design for tree-structured vector quantization (TSVQ) that is embedded in nature. We modify two speech coding standards by replacing their original quantizers for line spectral frequencies (LSF’s) and/or Fourier magnitudes quantization with TSVQ-based quantizers. The modified coders are fine-granular bit-rate scalable with gradual change in quality for the synthetic speech. A fast search encoding algorithm using multistage tree-structured vector quantization (MTVQ) is proposed for quantization of LSF’s. The proposed method is compared to the multipath sequential tree-assisted search (MSTS) and to the well known multipath sequential search (MSS) or M-L search algorithms.

Sex-specific differences in placental global gene expression in pregnancies complicated by asthma

BackgroundChronic maternal asthma is associated with reduced growth of the female fetus and normal growth of the male fetus. The mechanisms that control the differential effects of maternal asthma on the fetus have not been fully elucidated but alterations in placental function may play a role. In the current study we have used microarray platform to examine fetal sex-specific global changes in placental gene expression in pregnancies complicated by asthma as compared to non-asthmatic subjects. MethodsPlacental RNA was extracted from 11 control subjects and 38 asthmatic subjects. Labeled cDNA was hybridized to an oligonucleotide chip with 1700 double spotted well-characterized human genes. Global gene expression data analysis and visualization were performed using the Binary Tree-Structured Vector Quantization (BTSVQ) software. Functional relationships of differentially expressed genes were assessed using protein–protein interaction database I2D, network analysis and visualization software NAViGaTOR and Ingenuity Pathway Analysis software. ResultsOverall, 65 genes were found to be altered in placentae of pregnancies complicated by asthma. Of these, only 6 genes were altered in male placentae. There were 59 gene changes in female placentae of asthmatic mothers relative to control placentae. Some of the sex-specific genes were associated with growth, inflammation and immune pathways. ConclusionThere are sex-specific alterations in placental gene expression in the presence of maternal asthma. Given that many of the identified genes in the female placentae were associated with or involved in cellular growth and tissue development, these may contribute to the sexually dimorphic difference in fetal growth in response to maternal asthma.

Vector quantisation based on a quasi-binary search algorithm

This study presents an efficient quasi-binary search algorithm for vector quantisation (VQ). The proposed algorithm adopts a tree-structured VQ (TSVQ) with overlapped codewords (TSOC) to reduce computational complexity and enhance quantisation quality. This algorithm uses overlapped codewords to expand the scope of the search path to traverse more appropriate codewords. In the authors' speech experiment, compared with the full search VQ (FSVQ), the average computational savings for triangle inequality elimination (TIE), TSVQ and TSOC are 23.65, 88.63 and 59.43%, respectively. In this experiment, the quantisation accuracy of TIE, TSVQ and TSOC are 100, 46.61 and 99.16%, respectively. To further evaluate computations at each stage of the proposed algorithm, both speech and images are considered. With codebook sizes of 256, 512 and 1024, the corresponding optimal computational savings for images are 84.59, 91.08 and 93.51%, respectively, compared with the FSVQ. For speech, the optimal computational savings reached 59.43% for a codebook size of 128. The results indicate that the proposed algorithm can save a significant number of computations, depending on the size of codebook. The TSOC algorithm is a trade-off between TSVQ and TIE, which provides a satisfactory computation quality. Moreover, unlike the TIE method, our algorithm does not depend on the high correlation characteristics of signals to reduce the amount of computation, but the TIE method can be incorporated into our algorithm to dramatically reduce the amount of computation.

Nonproduct Data-Dependent Partitions for Mutual Information Estimation: Strong Consistency and Applications

A new framework for histogram-based mutual information estimation of probability distributions equipped with density functions in (Rd,B(Rd)) is presented in this work. A general histogram-based estimate is proposed, considering nonproduct data-dependent partitions, and sufficient conditions are stipulated to guarantee a strongly consistent estimate for mutual information. Two emblematic families of density-free strongly consistent estimates are derived from this result, one based on statistically equivalent blocks (the Gessaman's partition) and the other, on a tree-structured vector quantization scheme.

Information divergence estimation based on data-dependent partitions

This work studies the problem of information divergence estimation based on data-dependent partitions. A histogram-based data-dependent estimate is proposed adopting a version of Barron-type histogram-based estimate. The main result is the stipulation of sufficient conditions on the partition scheme to make the estimate strongly consistent. Furthermore, when the distributions are equipped with density functions in ( R d , B ( R d ) ) , we obtain sufficient conditions that guarantee a density-free strongly consistent information divergence estimate. In this context, the result is presented for two emblematic partition schemes: the statistically equivalent blocks ( Gessaman's data-driven partition) and data-dependent tree-structured vector quantization (TSVQ).

Random Projection Trees for Vector Quantization

A simple and computationally efficient scheme for tree-structured vector quantization is presented. Unlike previous methods, its quantization error depends only on the intrinsic dimension of the data distribution, rather than the apparent dimension of the space in which the data happen to lie.

Performance evaluation of a novel sampling-based texture synthesis technique using different sized patches

We propose a novel sampling-based texture synthesis algorithm called Multipatch, which improves on the results of previous sampling-based algorithms by using patches of different size, and by minimizing global pasting errors. A key feature of the proposed algorithm is that it always converges to a local minimum. Multipatch, the patchwork algorithm, and Wei and Levoy’s multi-resolution texture synthesis algorithm, which is based on a tree-structured vector quantization method, are statistically analyzed and subjectively evaluated. The results of simulations show that the patchwork algorithm yields a perceptually acceptable texture in a shorter expected running time than the other two algorithms; however, Multipatch is the most efficient in terms of obtaining a good quality texture image.