Codebook Size Research Articles

Efficient Resource Allocation in a Rateless-Coded MU-MIMO Cognitive Radio Network With QoS Provisioning and Limited Feedback

In this paper, we design an efficient resource-allocation strategy for a multiuser multiple-input-multiple-output (MU-MIMO) rateless-coded cognitive radio network (CRN) with quality-of-service (QoS) provisioning. We consider a limited feedback MU-MIMO CRN, where zero-forcing beamforming (ZFBF) is performed under imperfect channel state information (CSI) at a cognitive base station to mitigate both inter- and intranetwork interferences. To minimize the total feedback amount while satisfying the interference constraint and QoS requirements simultaneously, we propose to adaptively adjust the transmit power, select the transmission mode, and choose the feedback codebook size according to the interference constraint, CSI, and QoS requirements. The optimization problem is shown to be an integer programming problem, and we propose a heuristic algorithm that can provide an optimal solution for most practical scenarios. Results show that our resource-allocation strategy can decide the feedback amount and transmission mode adaptively based on the delay requirements.

Phase ambiguity mitigation for per-cell codebook-based limited feedback coordinated multi-point transmission systems

Limited feedback techniques, that employ codebook-based quantisation, are used in coordinated multi-point (CoMP) transmission to convey the channel direction information between the mobile user and each base station in the cooperative set and also among the cooperating base stations. However, unlike single-cell multiple-user multiple input multiple output (MU-MIMO) systems, CoMP systems face additional challenges in this codebook design. Among these challenges are the huge feedback overhead and the need for dynamic codebook size. These problems are solved when CoMP uses the so-called per-cell codebook, in which the channel to each cell is quantised separately. However, this results in a further problem, that is having random phase values between the quantised per-cell channels of the same user. In this study, the phase ambiguity problem is explained and its effect on performance degradation is quantified. A new quantisation technique is proposed where the quantisation of each of the real and the imaginary parts is performed independently using the same real codebook. This novel solution prevents the phase ambiguity problem instead of trying to mitigate its effects, in addition to simplifying the codebook design. The proposed quantisation technique outperforms two recently introduced solutions in the literature. The first solution is to quantise phase ambiguity as part of channel state information (CSI), whereas the other avoids phase ambiguity by an iterative selection procedure. Mathematical validations and simulations are provided to verify these findings.

Characteristics of the use of coupled hidden Markov models for audio-visual polish speech recognition

Abstract. This paper focuses on combining audio-visual signals for Polish speech recognition in conditions of the highly disturbed audio speech signal. Recognition of audio-visual speech was based on combined hidden Markov models (CHMM). The described methods were developed for a single isolated command, nevertheless their effectiveness indicated that they would also work similarly in continuous audiovisual speech recognition. The problem of a visual speech analysis is very difficult and computationally demanding, mostly because of an extreme amount of data that needs to be processed. Therefore, the method of audio-video speech recognition is used only while the audiospeech signal is exposed to a considerable level of distortion. There are proposed the authors’ own methods of the lip edges detection and a visual characteristic extraction in this paper. Moreover, the method of fusing speech characteristics for an audio-video signal was proposed and tested. A significant increase of recognition effectiveness and processing speed were noted during tests - for properly selected CHMM parameters and an adequate codebook size, besides the use of the appropriate fusion of audio-visual characteristics. The experimental results were very promising and close to those achieved by leading scientists in the field of audio-visual speech recognition.

A novel VQ-based reversible data hiding scheme by using hybrid encoding strategies

Reversible data hiding is a special algorithm that not only guarantees the confidential data will be extracted accurately but also allows the original cover image to be reconstructed without distortion after the confidential data are completely extracted. This paper proposes a new index compression and reversible data hiding scheme based on side-match vector quantization (SMVQ) and search order coding (SOC). In this proposed scheme, the confidential data are embedded into the transformed index table of a cover image. During the extracting phase, simple steps are employed to extract the confidential data and reconstruct the original cover image. The experimental results show that with a codebook size of 256, the average compression rate of the proposed scheme is 0.325bpp, which is superior to that of the methods proposed by Chen and Huang (0.426bpp) and Chang et al. (0.429bpp). Additionally, the embedding and extracting times of our scheme are 5.491s and 0.352s, respectively, demonstrating that the execution time of the proposed scheme is much faster than that of the methods of Chen and Huang and Chang et al. Moreover, our scheme achieves better performance than other selected reversible data hiding schemes with respect to the data embedding and data compression rates.

A Study on the Effect of Codebook and CodeVector Size on Image Retrieval Using Vector Quantization

In this paper, we study the effect of codebook size and codevector size using vector quantization (VQ) for retrieval of images, not restricted to the compressed domain. We use the image index model, to study the precision and recall values for different similarity measures. The study presents the following findings. The codebook size and codevec- tor size are directly proportional to the precision value for a locally global codebook, but are dependent on the size of the source image for a local codebook. The Encoding Distortion similarity measure calculated from the local codebook pro- duces the highest precision for the same recall over all other similarity measures. The Histogram Intersection using locally global codebook gives higher precision for higher codebook sizes. It is established that VQ can be used to create a single valued feature to represent the image in the image index model. This feature based on distortion measure can be effectively used for image retrieval based on the experimental results.

IMAGE COMPRESSION USING SELF-ORGANIZING FEATURE MAP AND WAVELET TRANSFORMATION

In this paper, a new method of vector quantizer design for image compression using Generic codebook and wavelet transformation is proposed. In the proposed method, Self Organizing Feature Map (SOFM) is used for initial codebook generation. A new scheme of wavelet transformation based Vector Quantization (VQ) technique is proposed to replace the SOFM code vectors by VQ code vectors. The proposed wavelet transform is used to generate wavelet coefficients which are then converted into VQ code vectors. Discrete Cosine Transformation based vector quantization technique is proposed in the existing image compression algorithms with low quality images with greater amount of information loss. Hence to increase the psycho visual quality of the reconstructed image wavelet transformation based vector quantization technique is proposed in this paper. Performance of the proposed work is tested with varying codebook size and various training images. Experimental results show that the reconstructed images obtained by the proposed method are of good quality with better compression ratio and higher Peak Signal–to– Noise Ratio.

Discovering Image Semantics in Codebook Derivative Space

The sparse coding based approaches for image recognition have recently shown improved performance than traditional bag-of-features technique. Due to high dimensionality of the image descriptor space, existing systems usually require very large codebook size to minimize coding error in order to get satisfactory accuracy. While most research efforts try to address the problem by constructing a relatively smaller codebook with stronger discriminative power, in this paper, we introduce an alternative solution by enhancing the quality of coding. Particularly, we apply the idea similar to Fisher kernel to the coding framework, where we use the image-dependent codebook derivative to represent the image. The proposed idea is generic across multiple coding criteria, and in this paper, it is applied to enhance the locality-constraint linear coding (LLC). Experiments show that, the extracted new feature, called “LLC+,” achieved significantly improved accuracy on several challenging datasets even with a small codebook of 1/20 the reported size used by LLC. This obviously adds to LLC+ the modeling accuracy, processing speed and codebook training advantages.

Effective Codebooks for Human Action Representation and Classification in Unconstrained Videos

Recognition and classification of human actions for annotation of unconstrained video sequences has proven to be challenging because of the variations in the environment, appearance of actors, modalities in which the same action is performed by different persons, speed and duration, and points of view from which the event is observed. This variability reflects in the difficulty of defining effective descriptors and deriving appropriate and effective codebooks for action categorization. In this paper, we propose a novel and effective solution to classify human actions in unconstrained videos. It improves on previous contributions through the definition of a novel local descriptor that uses image gradient and optic flow to respectively model the appearance and motion of human actions at interest point regions. In the formation of the codebook, we employ radius-based clustering with soft assignment in order to create a rich vocabulary that may account for the high variability of human actions. We show that our solution scores very good performance with no need of parameter tuning. We also show that a strong reduction of computation time can be obtained by applying codebook size reduction with Deep Belief Networks with little loss of accuracy.

Co.Vi.Wo.: Color Visual Words Based on Non-Predefined Size Codebooks

Due to the rapid development of information technology and the continuously increasing number of available multimedia data, the task of retrieving information based on visual content has become a popular subject of scientific interest. Recent approaches adopt the bag-of-visual-words (BOVW) model to retrieve images in a semantic way. BOVW has shown remarkable performance in content-based image retrieval tasks, exhibiting better retrieval effectiveness over global and local feature (LF) representations. The performance of the BOVW approach depends strongly, however, on predicting the ideal codebook size, a difficult and database-dependent task. The contribution of this paper is threefold. First, it presents a new technique that uses a self-growing and self-organized neural gas network to calculate the most appropriate size of a codebook for a given database. Second, it proposes a new soft-weighting technique, whereby each LF is classified into only one visual word (VW) with a degree of participation. Third, by combining the information derived from the method that automatically detects the number of VWs, the soft-weighting method, and a color information extraction method from the literature, it shapes a new descriptor, called color VWs. Experimental results on two well-known benchmarking databases demonstrate that the proposed descriptor outperforms 15 contemporary descriptors and methods from the literature, in terms of both precision at K and its ability to retrieve the entire ground truth.

Supervised learning probabilistic Latent Semantic Analysis for human motion analysis

Latent topic models such as Latent Dirichlet Allocation (LDA) and probabilistic Latent Semantic Analysis (pLSA) have demonstrated success in computer vision tasks. Most existing approaches train LDA and pLSA in an unsupervised manner, where the training data does not include any class label information. However, the class labels in training data are very important for the task of classification. In this paper, we propose to train a pLSA model in a supervised manner for the task of human motion analysis using the bag-of-words representation. Each frame in a video is treated as a word, and all the frames in the training videos are clustered to construct a codebook. The class label information is used to learn the pLSA model in a supervised manner, which not only makes the training more efficient, but also improves the overall recognition accuracy significantly. In addition, we employ the pyramid Histogram of orientation Gradient (HoG) to encode a human figure in each frame. The pyramid HoG descriptor does not require extraction of silhouettes, and is invariant to translations and rotations to some extent. The method is validated using two standard datasets. The experimental results show that our method can accurately recognize human motion in video sequences. Moreover, the overall recognition accuracy is rather stable with respect to the codebook size.

Supervised Learning and Codebook Optimization for Bag-of-Words Models

In this paper, we present a novel approach for supervised codebook learning and optimization for bag-of-words models. This type of models is frequently used in visual recognition tasks like object class recognition or human action recognition. An entity is represented as a histogram of codewords, which are traditionally clustered with unsupervised methods like k-means or random forests and then classified in a supervised way. We propose a new supervised method for joint codebook creation and class learning, which learns the cluster centers of the codebook in a goal-directed way using the class labels of the training set. As a result, the codebook is highly correlated to the recognition problem, leading to a more discriminative codebook. We propose two different learning algorithms, one based on error backpropagation and the other based on cluster label reassignment. We apply the proposed method to human action recognition from video sequences and evaluate it on the KTH data set, reporting very promising results. The proposed technique allows us to improve the discriminative power of an unsupervised learned codebook or to keep the discriminative power while decreasing the size of the learned codebook, thus decreasing the computational complexity due to the nearest neighbor search.

New Clustering Algorithm for Vector Quantization using Walsh Sequence

In this paper we present an effective clustering algorithm to generate codebook for vector quantization (VQ). Constant error is added every time to split the clusters in LBG, resulting in formation of cluster in one direction which is 135 in 2dimensional case. Because of this reason clustering is inefficient resulting in high MSE in LBG. To overcome this drawback of LBG proportionate error is added to change the cluster orientation in KPE. Though the cluster orientation in KPE is changed, its variation is limited to ± 45 over 135. KEVR introduces new orientation every time to split the clusters. But in KEVR the error vector sequence is the binary representation of numbers, so the cluster orientation change slowly in every iteration. To overcome this drawback we propose the technique which uses Walsh sequence to rotate the error vector. The proposed technique (Kekre’s error vector rotation using Walsh – KEVRW) is based on KEVR algorithm. The proposed methodology is tested on different training images for code books of sizes 128, 256, 512, 1024. Our result shows that KEVRW gives less MSE and high PSNR compared to LBG, KPE and KEVR. General Terms Image Processing, Vector Quantization, Data Compression.

Codebook-Based Lattice-Reduction-Aided Precoding for Limited-Feedback Coded MIMO Systems

Lattice-reduction-aided precoding (LRP) provides near-capacity rates with the use of low-complexity linear receivers for coded multiple-input multiple-output (MIMO) systems. However, a large amount of feedback in the feedback for an integer or binary precoding matrix has been a bottleneck in its implementation. In this paper, we propose a codebook-based LRP scheme for limited-feedback coded MIMO systems. The proposed LRP scheme follows the fundamentals of the previous LRP scheme that employed multilevel binary coset coding so that the precoding matrix is binary. In the proposed LRP scheme, the conventional precoding matrix obtained from the Lenstra-Lenstra-Lovasz algorithm is modified to specific forms that are predefined in a codebook set. The new precoding matrix is selected such that the lower bound on the capacity is maximized for a given codebook set, while the codebook set is designed offline such that the upper bound of the average capacity loss induced by the limitation on the codebook size is minimized. The simulation results show that the proposed LRP scheme nearly achieves the achievable rate of the conventional LRP scheme with a greatly reduced amount of feedback.

A Reduced Feedback Precoder for MIMO-OFDM Cooperative Diversity Systems

A new precoder that reduces the feedback overhead in an orthogonal frequency-division multiplexing (OFDM)-based relay network is proposed. A joint singular value decomposition (SVD) and QR decomposition is employed on the source-destination link and the source-relay-destination link. Using the proposed precoder, only one precoding matrix needs to be fed back, independent of the total number of subcarriers. Kernel density estimation is applied to determine approximate forms of unknown distributions of the end-to-end effective channel powers obtained by this precoder. Using approximated probability density functions (pdfs) for these effective channel powers, closed-form expressions for the link achievable average sum-rate and the outage probability can be derived and then their accuracies verified via Monte Carlo simulations. The achievable sum-rates are also compared with that of SVD-based precoders and for varying codebook size.

Performance Evaluation of Closed-Loop Spatial Multiplexing Codebook Based on Indoor MIMO Channel Measurement

Closed-loop MIMO technique standardized in LTE can support different layer transmissions through precoding operation to match the channel multiplexing capability. However, the performance of the limited size codebook still needs to be evaluated in real channel environment for further insights. Based on the wideband MIMO channel measurement in a typical indoor scenario, capacity loss (CL) of the limited size codebook relative to perfect precoding is studied first in two extreme channel conditions. The results show that current codebook design for single layer transmission is nearly capacity lossless, and the CL will increase with the number of transmitted layers. Furthermore, the capacity improvement of better codebook selection criterions is very limited compared to CL. Then we define the maximum capacity boost achieved by frequency domain layer adaption (FDLA) and investigate its sensitivity to SNR and channel condition. To survey the effect of frequency domain channel variation on MIMO-OFDM system, we define a function to measure the fluctuation levels of the key channel metrics within a subband and reveal the inherent relationship between them. Finally, a capacity floor resulted as the feedback interval increases in frequency domain.

Dynamic selection of generative–discriminative ensembles for off-line signature verification

In practice, each writer provides only a limited number of signature samples to design a signature verification (SV) system. Hybrid generative–discriminative ensembles of classifiers (EoCs) are proposed in this paper to design an off-line SV system from few samples, where the classifier selection process is performed dynamically. To design the generative stage, multiple discrete left-to-right Hidden Markov Models (HMMs) are trained using a different number of states and codebook sizes, allowing the system to learn signatures at different levels of perception. To design the discriminative stage, HMM likelihoods are measured for each training signature, and assembled into feature vectors that are used to train a diversified pool of two-class classifiers through a specialized Random Subspace Method. During verification, a new dynamic selection strategy based on the K-nearest-oracles (KNORA) algorithm and on Output Profiles selects the most accurate EoCs to classify a given input signature. This SV system is suitable for incremental learning of new signature samples. Experiments performed with real-world signature data (composed of genuine samples, and random, simple and skilled forgeries) indicate that the proposed dynamic selection strategy can significantly reduce the overall error rates, with respect to other EoCs formed using well-known dynamic and static selection strategies. Moreover, the performance of the SV system proposed in this paper is significantly greater than or comparable to that of related systems found in the literature.

Joint power control and beamforming codebook design for MISO channels under the outage criterion

This paper investigates the joint design of power control and beamforming codebooks for limited-feedback multiple-input single-output (MISO) wireless systems. The problem is formulated as the minimization of the outage probability subject to the transmit power constraint and cardinality constraints on the beamforming and power codebooks. We show that the two codebooks need to be designed jointly in this setup, and provide a numerical method for the joint optimization. For independent and identically distributed (i.i.d.) Rayleigh channel, we also propose a low-complexity approach of fixing a uniform beamforming codebook and optimizing the power codebook for that particular beamformer, and show that it performs very close to the optimum. Further, this paper investigates the optimal tradeoffs between beamforming and power codebook sizes. We show that as the outage probability decreases, optimal joint design should use more feedback bits for power control and fewer feedback bits for beamforming. The jointly optimized beamforming and power control modules combine the power gain of beamforming and diversity gain of power control, which enable it to approach the performance of the system with perfect channel state information as the feedback link capacity increases—something that is not possible with either beamforming or power control alone.

Improvements in image categorization using codebook ensembles

The problem of object category classification by committees or ensembles of classifiers, each of which is based on one diverse codebook, is addressed in this paper. Two methods of constructing visual codebook ensembles are proposed in this study. The first technique introduces diverse individual visual codebooks using different clustering algorithms. The second uses various visual codebooks of different sizes for constructing an ensemble with high diversity. Codebook ensembles are trained to capture and convey image properties from different aspects. Based on these codebook ensembles, different types of image representations can be acquired. A classifier ensemble can be trained based on different expression datasets from the same training image set. The use of a classifier ensemble to categorize new images can lead to improved performance. Detailed experimental analysis on a Pascal VOC challenge dataset reveals that the present ensemble approach performs well, consistently improves the performance of visual object classifiers, and results in state-of-the-art performance in categorization.

An HMM Based Pitch-Contour Generation Method for Mandarin Speech Synthesis

In this paper, a method is proposed to generate pitch-contours for Mandarin speech synthesis. In this method, an HMM (hidden Markov model) is used to model the pro- sodic states implicitly stayed and a syllable's pitch-contour is treated as an observation generated from a prosodic state. Such an HMM is called a syllable pitch-contour HMM (SPC-HMM). For training the SPC-HMM, we developed a feasible method to normalize a pitch-contour's height. After normalization, each training syllable's pitch-contour is vector quantized and represented with a VQ (vector quantization) code. Then, the VQ code and its adjacent syllables' lexical tones are combined to define an observation symbol for training the SPC-HMM. In the synthesis phase, a sentence-wide most prob- able observation symbol sequence is searched on the SPC-HMM using a dynamic pro- gramming algorithm proposed here. Then, the observation symbol found for a syllable is decoded to obtain its pitch-contour VQ code. We conducted testing experiments to de- termine the size of a pitch-contour codebook and the number of states for an SPC-HMM. The results indicate that setting the codebook size to eight and using six states are the best choices. Also, we conducted perception tests to compare the naturalness levels of synthetic speech files. The results show that the two generation modes for operating an SPC-HMM studied here are comparable to each other in naturalness level.

정규화 코드북을 이용한 분할 벡터 구조의 ISF 적응적 양자화 기법

In most of the ISF (or LSF) based real time speech codec, SVQ (split vector quantization) method is used to decrease the quantizer complexity and memory size of codebook. However, it produces drawback that the level of correlation between code vectors can not be used during vector splits. This paper presents a new method of adaptive ISF vector quantization, which compensates the drawbacks of SVQ structured quantizer for wideband speech codec. In each different frame, the proposed method makes use of the correlation between splitted vectors by adaptively changing codebook distribution according to ordering property of ISF. The algorithm is evaluated in AMR-WB, and shows about 1.5 bit per frame improvement.