LBG Algorithm Research Articles

Fast Linde-Buzo-Gray (FLBG) Algorithm for Image Compression through Rescaling Using Bilinear Interpolation.

Vector quantization (VQ) is a block coding method that is famous for its high compression ratio and simple encoder and decoder implementation. Linde-Buzo-Gray (LBG) is a renowned technique for VQ that uses a clustering-based approach for finding the optimum codebook. Numerous algorithms, such as Particle Swarm Optimization (PSO), the Cuckoo search algorithm (CS), bat algorithm, and firefly algorithm (FA), are used for codebook design. These algorithms are primarily focused on improving the image quality in terms of the PSNR and SSIM but use exhaustive searching to find the optimum codebook, which causes the computational time to be very high. In our study, our algorithm enhances LBG by minimizing the computational complexity by reducing the total number of comparisons among the codebook and training vectors using a match function. The input image is taken as a training vector at the encoder side, which is initialized with the random selection of the vectors from the input image. Rescaling using bilinear interpolation through the nearest neighborhood method is performed to reduce the comparison of the codebook with the training vector. The compressed image is first downsized by the encoder, which is then upscaled at the decoder side during decompression. Based on the results, it is demonstrated that the proposed method reduces the computational complexity by 50.2% compared to LBG and above 97% compared to the other LBG-based algorithms. Moreover, a 20% reduction in the memory size is also obtained, with no significant loss in the image quality compared to the LBG algorithm.

On the Initialization of Swarm Intelligence Algorithms for Vector Quantization Codebook Design.

Vector Quantization (VQ) is a technique with a wide range of applications. For example, it can be used for image compression. The codebook design for VQ has great significance in the quality of the quantized signals and can benefit from the use of swarm intelligence. Initialization of the Linde-Buzo-Gray (LBG) algorithm, which is the most popular VQ codebook design algorithm, is a step that directly influences VQ performance, as the convergence speed and codebook quality depend on the initial codebook. A widely used initialization alternative is random initialization, in which the initial set of codevectors is drawn randomly from the training set. Other initialization methods can lead to a better quality of the designed codebooks. The present work evaluates the impacts of initialization strategies on swarm intelligence algorithms for codebook design in terms of the quality of the designed codebooks, assessed by the quality of the reconstructed images, and in terms of the convergence speed, evaluated by the number of iterations. Initialization strategies consist of a combination of codebooks obtained by initialization algorithms from the literature with codebooks composed of vectors randomly selected from the training set. The possibility of combining different initialization techniques provides new perspectives in the search for the quality of the VQ codebooks. Nine initialization strategies are presented, which are compared with random initialization. Initialization strategies are evaluated on the following algorithms for codebook design based on swarm clustering: modified firefly algorithm-Linde-Buzo-Gray (M-FA-LBG), modified particle swarm optimization-Linde-Buzo-Gray (M-PSO-LBG), modified fish school search-Linde-Buzo-Gray (M-FSS-LBG) and their accelerated versions (M-FA-LBGa, M-PSO-LBGa and M-FSS-LBGa) which are obtained by replacing the LBG with the accelerated LBG algorithm. The simulation results point out to the benefits of the proposed initialization strategies. The results show gains up to 4.43 dB in terms of PSNR for image Clock with M-PSO-LBG codebooks of size 512 and codebook design time savings up to 67.05% for image Clock, with M-FF-LBGa codebooks with size N=512, by using initialization strategies in substitution to Random initialization.

Stochastic Bat Optimization Model for Secured WSN with Energy‐Aware Quantized Indexive Clustering

The wireless sensor networks (WSNs) with dynamic topology communication among the sensor nodes is vulnerable to numerous attacks. As they have limited power, there arises a conflict between the complex security scheme and the consumption of energy which are inversely proportional to each other. Hence, a trade‐off should be accomplished between the implemented scheme and energy. A novel secure and energy‐aware routing technique quantized indexive energy‐aware clustering‐based combinatorial stochastic sampled bat optimization (QIEAC‐CSSBO) is proposed which consists of clustering, optimal route path identification, and route maintenance. The clustering process and selection of cluster head (CH) with high residual energy is done using the quantized Schutz indexive Linde–Buzo–Gray algorithm (QIEAC). Optimal route identification is done using CSSBO (combinatorial stochastic sampled Prevosti’s bat optimization), and fitness of every bat is measured on combinatorial functions, namely, distance, energy, trust, and link stability among nodes. Stochastic universal sampling selection procedure is applied to select the global best optimal path for secure data transmission. Lastly, route maintenance process is performed to identify alternative route while link failure occurs among nodes. Experimental assessment is performed using various performance metrics, namely, energy consumption, packet delivery ratio, packet drop rate, throughput, and delay. The proposed method QIEAC‐CSSBO enhances the performance of packet delivery ratio by 4%, throughput by 26%, and packet drop rate by 27% and reduces energy consumption by 11%, as well as delay by 16% as compared to existing techniques.

A Weight Based Hierarchical Vector Quantization Algorithm for Space Environment Volume Data

Visualization has been widely applied in space environment domain. However, compressed volume rendering algorithms based on VQ are concerned on fidelity and compression rate, not combined with specific application. To fulfill the specific visualization requirements for space environment volume data, an application-driven compression and rendering algorithm is proposed, which is Weight Based Hierarchical Vector Quantization (WHVQ). The volume data is initially partitioned into disjoint 4³ blocks. Weights are assigned to the blocks according to their importance. The blocks are then decomposed into a three level hierarchical representation and each block is represented by a mean value and two detail vectors. To the top two levels, a splitting based on principal component analysis and weight is adopted to form their initial codebooks. Then, LBG algorithm based on weight is conducted for codebook refinement and quantization. The experimental results show that WHVQ is able to improve the quality of reconstruction in interested area on the premise of the good overall fidelity.

Lossless Image Compression Using Wavelet and Vector Quantization.(Dept.E)

In this paper, we propose a new low bit rate scheme that combines the Discrete Wavelet transform (IWT) and Vector Quantization (VQ) for image compression with high Compression Ratio (CR) in order to achieve acceptable quality of output images, Image compression using the combination of DWT willy VQ has been considered in many recent works; here in our paper the DWT was performed on several gray scale standard images, each of (256X256) byte. We used 4 images as the training images of the train VO using LBG algorithm and got the final optimal codebook. Then we used another two images to compare our output image with JPEG 2000 standard winch based on Wavelet and Scalar quantization. Our proposed system used a VO codebook with Variable cod hook size and dimensions were used to compress wavelet Transform) coefficients resulting using Danny wavelet types. Our pro system is slowed so that the best Wavelet types that you can use for compression is COIF 2. Our proposed system gave better PSNR than G2000 standard at high (CR), above than (5L), and this achieved a VQ codebook size (1024 -512) with dimension equal to 256. In the paper a typical system for the application of image compression system and new scientific results are present.

A novel image compression model by adaptive vector quantization: modified rider optimization algorithm

In recent days over the internet, the uploading of enormous new images is being made every day, and they necessitate large storage to accumulate the image data. For the earlier few decades, more analysts have evolved skillful image compression schemes to enhance the compression rates and the image quality. In this work, Vector Quantization is used, which uses the Linde–Buzo–Gray algorithm. As a novel intention, the codebooks are optimized by an improved optimization algorithm. In this approach, the database image is firstly separated into a set of blocks, i.e., pixels, and these sets of blocks are referred to as vectors. Then a suitable codeword is selected for each vector such that is the closest representation of that input vector. The encoder generates a codebook by mapping the vectors on the basis of these code words, and the compression of the vectors takes place. The encoder then sends a compressed stream of these vectors by pointing out their indices from the codebook to the decoder through a channel. The decoder then decodes the index to find out the compressed vector and places it on the image. For attaining a better image compression effect, the codebook is optimized using the Best Fitness Updated Rider Optimization Algorithm. The optimization of codebooks is done so that the summation of the compression ratio and the error difference between the original and decompressed images has to be minimized. Moreover, the proposed model is scruntized with other existing algorithms, and the experimental outcomes are validated.

Comparitive Analysis and Findings on Dct & Lbg Compression Techniques

The main objective of this project is to develop an application to find the best compression technique to store Muscat College students' photographs in less storage. MATLAB software will be used to develop a Graphical User Interface GUI application and implement two image compression techniques which are lossless compression using the DCT algorithm and lossy compression using the LBG algorithm. The application shall allow the user to select and test a sample image by applying both these techniques for any student image he\she selects in order to compare the results by display the image after compression and the histogram to find which the most suitable compression technique is. Also, the application shall show the size of images before and after applying the compression process and show the compression ratio and relative data redundancy of compressed image/images. The main functionality is that the application shall allow the user to do bulk processing to apply image enhancement and image compression technique to enhance and compress all the photographs of students and store them in less space.

Lung diseases identification method based on capsule neural network

In the automatic identification of lesions, aiming at the misdiagnosis of the lung nodule size, shape, blood vessels and other lung tissues on CT images of pneumonia, this paper proposes a method for identifying lung diseases based on capsule neural networks. Considering that the texture features of lung CT images contain important medical information, this article first combines gray average, entropy, fractal dimension and fractal intercept to form feature vectors as texture features, and introduces context models to obtain context information. The LBG algorithm was used to achieve context quantization, and the CT images of lungs were preprocessed. Then, for the problem of less CT image data of lungs and fewer features of lung diseases in this paper, the existing images are enhanced and the data set is expanded. Finally, in the recognition process of pneumonia, the characteristic pose information will affect the recognition. In order to solve this problem, this paper uses a capsule neural network to identify and detect lung CT images. The application results show that the accuracy rate of identifying lung diseases by capsule neural network is 98.7%. Compared with traditional convolutional networks, capsule neural networks can better identify lung diseases.

Using Vector Quantization Technique in Recognition of Gurbani Hymns (Japji Sahib): LBG Algorithm (VQ)

An improved and different variation of Automatic Speech Recognition (ASR) is presented which is based on Vector Quantization (VQ). ASR for different languages and different applications has been introduced so far. In this paper, we have presented a Speech Recognition system to recognize the hymns (paath) of Gurbani (sentences of Japji Sahib) as continuous mode of speech. For this, speech corpus has been generated in which the entire path has been recited by different speakers. The speech mode here can be taken as continuous speech encapsulated with background music and different kinds of additional noises and have been eliminated. The work has been done by using VQ approach of speech recognition and LBG algorithm which design optimal codebooks for the process of recognition. Experimental results are included which show that recognition accuracy for such system was found to be 92.6% and 95.8% for different and same speakers with different and same sentences.

Clustering Approaches for Management Zone Delineation in Precision Agriculture for Small Farms

For enhancing the quality and the productivity of the crop, usage of modern tools and techniques has become inevitable. One of such technique is Precision Agriculture (PA). PA collects and controls agronomic information to furnish actual nutrient needs to parts of fields rather than average needs to complete fields. These parts of fields are management zones which can be used to treat the within field variation. Management zone delineation (MZD) has become an integral part and pillar of Precision agriculture by dividing the field according to soil physical and chemical characteristics. Concept of clustering from data mining domain is suitable to create such management zones within the field. This paper experiments and compares K mean, FCM, PFCM and LBG clustering algorithm for delineating the management zones in precision agriculture. The objective of determination of zone delineation is for the application of fertilization process. Sugarcane (Saccharum officinarum) has been selected as case study for the experimentation of MZD. It considers 14 important nutrients of the crop for the delineation. Real time data set is generated for the experimentation. Result shows that PFCM algorithm works better over k-mean, FCM and LBG algorithm with spatial data set.

Recurrent Neural Network-Based Dictionary Learning for Compressive Speech Sensing

We propose a novel dictionary learning technique for compressive sensing of speech signals based on the recurrent neural network. First, we exploit the recurrent neural network to solve an $$\ell _{0}$$ -norm optimization problem based on a sequential linear prediction model for estimating the linear prediction coefficients for voiced and unvoiced speech, respectively. Then, the extracted linear prediction coefficient vectors are clustered through an improved Linde–Buzo–Gray algorithm to generate codebooks for voiced and unvoiced speech, respectively. A dictionary is then constructed for each type of speech by concatenating a union of structured matrices derived from the column vectors in the corresponding codebook. Next, a decision module is designed to determine the appropriate dictionary for the recovery algorithm in the compressive sensing system. Finally, based on the sequential linear prediction model and the proposed dictionary, a sequential recovery algorithm is proposed to further improve the quality of the reconstructed speech. Experimental results show that when compared to the selected state-of-the-art approaches, our proposed method can achieve superior performance in terms of several objective measures including segmental signal-to-noise ratio, perceptual evaluation of speech quality and short-time objective intelligibility under both noise-free and noise-aware conditions.

Vector quantization codebook design based on Fish School Search algorithm

Vector Quantization (VQ) has been used in image coding systems since it allows high compression rates. Codebook design can be seen as a high dimensional optimization problem and, in this scenario, swarm intelligence techniques have been used. This paper presents a new VQ codebook design algorithm based on swarm clustering. The method, based on Fish School Search (FSS) algorithm, is introduced. The FSS is embedded in Linde–Buzo–Gray (LBG) algorithm as a swarm clustering method, here called FSS-LBG. Also, a modification in the original FSS breeding operation is proposed in order to favor the exploration ability and, therefore, achieve better results in terms of PSNR of the reconstructed images. Simulation results show gains up to 1.57 dB in terms of PSNR when compared to LBG algorithm for the image Lena at 0.5625 bpp using a codebook of size 512.

Adaptive Radiographic Image Compression Technique using Hierarchical Vector Quantization and Huffman Encoding

The objective of this paper is to provide a proposed method for radiographic image compression with maximum Compression Ratio (CR) as possible and keeping all details, especially in the Region Of Interest (ROI) that contains the important information in the image. In the proposed method; firstly, the ROI is separated from the image background using an automatic threshold based on the occurrence histogram of the variance image, then the image background is compressed with the maximum possible compression ratio using image pyramid compression followed by lossy Vector Quantization (VQ) compression technique based on Generalized Lloyd Algorithm (GLA) method for generating the codebook. After that, ROI is compressed using the Huffman Code (HC) with low compression ratio and with minimum loss in details. Finally, the compressed image is obtained by combing both the compressed background and the compressed ROI. The results are evaluated by calculating the Normalized Cross Correlation (NCC) and the Structural Similarity Index (SSIM) between the original image and the recovered image after decompression at different compression ratios. The obtained results are compared with the results obtained from compressing the whole image without separation using lossy VQ, HC, Discrete Cosine Transform (DCT), and JPEG2000 compression methods. The results show that, the proposed method owns more reliable performance than completely compressing the radiographic image without separation using VQ compression or Huffman coding compression.

SAR IMAGE COMPRESSION USING ADAPTIVE DIFFERENTIAL EVOLUTION AND PATTERN SEARCH BASED K-MEANS VECTOR QUANTIZATION

A novel Vector Quantization (VQ) technique for encoding the Bi-orthogonal wavelet decomposed image using hybrid Adaptive Differential Evolution (ADE) and a Pattern Search optimization algorithm (hADEPS) is proposed. ADE is a modified version of Differential Evolution (DE) in which mutation operation is made adaptive based on the ascending/descending objective function or fitness value and tested on twelve numerical benchmark functions and the results are compared and proved better than Genetic Algorithm (GA), ordinary DE and FA. ADE is a global optimizer which explore the global search space and PS is local optimizer which exploit a local search space, so ADE is hybridized with PS. In the proposed VQ, in a codebook of codewords, 62.5% of codewords are assigned and optimized for the approximation coefficients and the remaining 37.5% are equally assigned to horizontal, vertical and diagonal coefficients. The superiority of proposed hybrid Adaptive Differential Evolution and Pattern Search (hADE-PS) optimized vector quantization over DE is demonstrated. The proposed technique is compared with DE based VQ and ADE based quantization and with standard LBG algorithm. Results show higher Peak Signal-to-Noise Ratio (PSNR) and Structural Similiraty Index Measure (SSIM) indicating better reconstruction.

Power Allocation for Distributed Detection Systems in Wireless Sensor Networks with Limited Fusion Center Feedback

We consider a distributed detection system for a wireless sensor network over slow-fading channels. Each sensor only has knowledge of quantized channel state information (CSI) which is received from the fusion center via a limited feedback channel. We then consider transmit power allocation at each sensor in order to maximize a J-divergence based detection metric subject to a total and individual transmit power constraints. Our aim is to jointly design the quantization regions of all sensors CSI and the corresponding power allocations. A locally optimum solution is obtained by applying the generalized Lloyd algorithm (GLA). To overcome the high computational complexity of the GLA, we then propose a low-complexity near-optimal scheme which performs very close to its GLA based counterpart. This enables us to explicitly formulate the problem and to find the unique solution despite the non-convexity of the optimization problem. An asymptotic analysis is also provided when the number of feedback bits becomes large. Numerical results illustrate that only a small amount of feedback is needed to achieve a detection performance close to the full CSI case.

A highly efficient method for improving the performance of GLA-based algorithms

Motivated by the observation that most methods for accelerating the generalized Lloyd algorithm (GLA) normally lack the capability to improve the quality of its end result and that most methods for improving the quality of the end result of GLA usually lack the capability to speed it up, an efficient and effective method is presented in this paper to enhance the performance of GLA and its variants, in terms of both the computation time and the quality of the end result, by leveraging the strengths of several reduction methods and the multiple stage mechanism. Simulation results show that the proposed method can efficiently and effectively reduce the computation time of GLA by up to about 93% while at the same time improving its quality by up to about 1 dB in terms of the peak-signal-to-noise-ratio in most cases.

Vibration-based damage detection using online learning algorithm for output-only structural health monitoring

Damage-sensitive features such as natural frequencies are widely used for structural health monitoring; however, they are also influenced by the environmental condition. To address the environmental effect, principal component analysis is widely used. Before performing principal component analysis, the training data should be defined for the normal condition (baseline model) under environmental variability. It is worth noting that the natural change of the normal condition may exist due to an intrinsic behavior of the structural system. Without accounting for the natural change of the normal condition, numerous false alarms occur. However, the natural change of the normal condition cannot be known in advance. Although the description of the normal condition has a significant influence on the monitoring performance, it has received much less attention. To capture the natural change of the normal condition and detect the damage simultaneously, an adaptive statistical process monitoring using online learning algorithm is proposed for output-only structural health monitoring. The novelty aspect of the proposed method is the adaptive learning capability by moving the window of the recent samples (from normal condition) to update the baseline model. In this way, the baseline model can reflect the natural change of the normal condition in environmental variability. To handle both change rate of the normal condition and non-linear dependency of the damage-sensitive features, a variable moving window strategy is also proposed. The variable moving window strategy is the block-wise linearization method using k-means clustering based on Linde–Buzo–Gray algorithm and Bayesian information criterion. The proposed method and two existing methods (static linear principal component analysis and incremental linear principal component analysis) were applied to a full-scale bridge structure, which was artificially damaged at the end of the long-term monitoring. Among the three methods, the proposed method is the only successful method to deal with the non-linear dependency among features and detect the structural damage timely.

A New Spliting LBG Algorithm for Codebook Generation in Image Compression

A New Spliting LBG Algorithm for Codebook Generation in Image Compression

Genetic Simulated Annealing-Based Kernel Vector Quantization Algorithm

Genetic Algorithm (GA) has been successfully applied to codebook design for vector quantization and its candidate solutions are normally turned by LBG algorithm. In this paper, to solve premature phenomenon and falling into local optimum of GA, a new Genetic Simulated Annealing-based Kernel Vector Quantization (GSAKVQ) is proposed from a different point of view. The simulated annealing (SA) method proposed in this paper can approach the optimal solution faster than the other candidate approaches. In the frame of GA, firstly, a new special crossover operator and a mutation operator are designed for the partition-based code scheme, and then a SA operation is introduced to enlarge the exploration of the proposed algorithm, finally, the Kernel function-based fitness is introduced into GA in order to cluster those datasets with complex distribution. The proposed method has been extensively compared with other algorithms on 17 datasets clustering and four image compression problems. The experimental results show that the algorithm can achieve its superiority in terms of clustering correct rate and peak signal-to-noise ratio (PSNR), and the robustness of algorithm is also very good. In addition, we took “Lena” as an example and added Gaussian noise into the original image then adopted the proposed algorithm to compress the image with noise. Compared to the original image with noise, the reconstructed image is more distinct, and with the parameter value increasing, the value of PSNR decreases.

Research on Techniques of Multifeatures Extraction for Tongue Image and Its Application in Retrieval.

Tongue diagnosis is one of the important methods in the Chinese traditional medicine. Doctors can judge the disease's situation by observing patient's tongue color and texture. This paper presents a novel approach to extract color and texture features of tongue images. First, we use improved GLA (Generalized Lloyd Algorithm) to extract the main color of tongue image. Considering that the color feature cannot fully express tongue image information, the paper analyzes tongue edge's texture features and proposes an algorithm to extract them. Then, we integrate the two features in retrieval by different weight. Experimental results show that the proposed method can improve the detection rate of lesion in tongue image relative to single feature retrieval.