Codebook Size Research Articles

A Novel Efficient Initial Access Method for 5G Millimeter Wave Communications Using Genetic Algorithm

Initial Access (IA) in millimeter wave (mmWave) communications, i.e., the mechanism by which the Base Station (BS) and the User Equipment (UE) establish a directional communication link, is a challenging problem. A significant delay can be incurred when the BS and the UE try to find the appropriate beam alignment for obtaining a directional link, and solving this problem efficiently has become an important research topic. In this work, we propose a new beam refinement method based on Genetic Algorithms (GAs) which is particularly advantageous when the number of antennas is large. In particular, we consider a scenario where both the BS and the UE are equipped with a single RF chain and multiple antennas, i.e., a single beam analog beamforming is considered at the BS and UE. The efficiency of the proposed method and the effect of several parameters, such as the number of transmit and receive antennas, codebook size, and transmission power are investigated. In the case of small parameters, it is shown that the proposed method achieves the same capacity of Exhaustive Search (ES) but with a much lower complexity. Moreover, in comparison with existing methods, our method has a superior performance in terms of capacity, outage probability, and power consumption.

The impact of the image processing in the indexation system

This paper presents an efficient word spotting system applied to handwritten Arabic documents, where images are represented with bag-of-visual-SIFT descriptors and a sliding window approach is used to locate the regions that are most similar to the query by following the query-by-example paragon. First, a pre-processing step is used to produce a better representation of the most informative features. Secondly, a region-based framework is deployed to represent each local region by a bag-of-visual-SIFT descriptors. Afterward, some experiments are in order to demonstrate the codebook size influence on the efficiency of the system, by analyzing the curse of dimensionality curve. In the end, to measure the similarity score, a floating distance based on the descriptor’s number for each query is adopted. The experimental results prove the efficiency of the proposed processing steps in the word spotting system.

Grassmannian Product Codebooks for Limited Feedback Massive MIMO With Two-Tier Precoding

In frequency-division duplex full-dimension massive MIMO systems, one of the main challenges is to obtain sufficiently accurate channel state information (CSI) at the transmitter to enable efficient multi-user MIMO transmission. In this paper, we propose a novel dual stage Grassmannian product quantization approach that is suitable for high-dimensional CSI quantization and feedback. We apply the proposed method for outer-tier CSI feedback in two-tier precoding architectures, which employ channel-subspace based outer-tier precoding strategies, such as maximum eigenmode transmission. The proposed method is especially effective if the channel can be decomposed in the angular domain, such that DFT codebooks enable an efficient CSI compression. Our dual stage product codebook quantization approach mitigates the well-known inefficiency of oversampled DFT codebooks for growing codebook sizes, by varying the intermediate quantization dimension of the proposed quantizer.

Video compression based on Multiwavelet and Multi Stage Vector Quantization using Adaptive Diamond Refinement Search Algorithm

Due to the advances in the digital technology, multimedia processing has become the essential requirement in many applications. These applications find wide use in mobile, personal computer(PC), TV, surveillance and satellite broadcast. Also it is necessary that the video coding algorithms to be updated in order to meet the requirements of latest hardware devices. The processing speed and bandwidth are essential parameters in these applications. A good video compression standard can achieve these parameters adequately. In the proposed system, the video coding standard is implemented using the three important stages. In which the first sage uses multiwavelets to achieve good compression rate. Also it reduces the memory and bandwidth requirement. Second stage is the Multi Stage Vector Quantization(MVSQ) which reduces the complexity of searching process and the size of codebook. Third stage uses Adaptive Diamond Refinement Search(ADRS) algorithm for the motion estimation which has better performance than the Adaptive Diamond Orthogonal Search(ADOS) and Diamond Refinement Search(DRS) algorithms. The combination of multiwavelet, Multi Stage Vector Quantization(MVSQ) and Adaptive Diamond Refinement Search(ADRS) algorithm gives the high compression ratios. Preliminary results indicate that the proposed method has good performance in terms of average number of search points, PSNR values and compression rates.

A Codeword Search Scheme for Massive MIMO Systems

In limited-feedback systems, conventional codeword search scheme targets the entire codebook to find a codeword, which maximizes an inner product norm with a channel direction. Since the computational complexity of the conventional codeword search scheme is proportional to the multiplication of the number of antennas and the codebook size, it spends too much time in the massive MIMO systems. From a geometric observation, we propose a fast codeword search scheme that stops search when an inner product norm between a codeword and a channel direction is larger than a threshold to reduce the search time. We derive an optimal threshold that can reduce the average number of searches without search accuracy loss and a suboptimal threshold that can reduce the average number of searches largely at the cost of search accuracy. We analyze the performances of the proposed scheme as functions of the thresholds, and show that the optimal threshold loses its benefit in the massive MIMO systems and the suboptimal threshold can reduce the number of searches by $e^{-1}\approx \text{37}\%$ without performance loss when the number of antennas is infinite. The simulation results show that our theoretical analysis is accurate.

A centroid based vector quantization reversible data hiding technique

In this paper, we propose a reversible data hiding scheme that exploits the centroid formula. Specifically, we use it to define a centroid boundary vector and a centroid state codebook CSCB. Initially, our centroid boundary vectors and CSCBs are the same as the side match vector quantization (SMVQ) algorithm’s boundary vectors and state codebooks SCBs. For each VQ index, the proposed scheme exploits the centroid formula to update its centroid boundary vector and the corresponding CSCB. The updating is coupled with a heuristic to select the best state codebook (i.e., either SCB or CSCB) for each VQ index, which generates a highly compressible distribution of index values. Our experimental results show that the proposed scheme can embed n = 1, 2, 3, and 4 bit per index (bpi) at bit rates of 0.332, 0.394, 0.457, and 0.519 bit per pixel (bpp), respectively, for the main codebook size N = 256. These results confirm that the proposed scheme improves recent VQ and SMVQ based reversible data hiding schemes.

An Optimal Codebook for Content-Based Image Retrieval in JPEG Compressed Domain

Images in JPEG compressed domain have a widespread use and computationally efficient retrieval of those images is of prime concern. In this paper, a content-based image retrieval system in JPEG compressed domain is presented, which generates an optimal codebook and extract features that only require partial decoding of images. The codebook is generated by selecting an optimum number of training images and feature vector length using an optimization cost based on precision and recall. The cost gives the difference in average precision and average recall, which the retrieval system can tolerate, while reducing the codebook size and feature vector length. The efficiency in retrieval performance is achieved by generating an optimal codebook. Experimental results have shown better performance for the proposed retrieval system in terms of precision, recall and number of operations, when compared with state-of-the-art image retrieval systems in JPEG compressed domain. The proposed retrieval system also shows robustness against compressed query images by using different quantization parameters.

Automatic Scale Severity Assessment Method in Psoriasis Skin Images Using Local Descriptors.

Psoriasis is a chronic skin condition. Its clinical assessment involves four measures: erythema, scales, induration, and area. In this paper, we introduce a scale severity scoring framework for two-dimensional psoriasis skin images. Specifically, we leverage the bag-of-visual words (BoVWs) model for lesion feature extraction using superpixels as key points. BoVWs model is based on building a vocabulary with specific number of words (i.e., codebook size) by using a clustering algorithm with some local features extracted from a constructed set of key points. This is followed by three-class machine learning classifiers for scale scoring using support vector machine (SVM) and random forest. Besides, we examine eight different local color and texture descriptors, namely color histogram, local binary patterns, edge histogram descriptor, color layout descriptor, scalable color descriptor, color and edge directivity descriptor (CEDD), fuzzy color and texture histogram, and brightness and texture directionality histogram. Further, the selection of codebook and superpixel sizes are studied intensively. A psoriasis image set, consisting of 96 images, is used in this study. The conducted experiments show that color descriptors have the highest performance measures for scale severity scoring. This is followed by the combined color and texture descriptors, whereas texture-based descriptors come last. Moreover, K-means algorithm shows better results in vocabulary building than Gaussian mixed model, in terms of accuracy and computations time. Finally, the proposed method yields a scale severity scoring accuracy of 80.81% using the following setup: a superpixel of size [Formula: see text], a combined color and texture descriptor (i.e., CEDD), a constructed codebook of size 128 using K-means, and SVM for scale scoring.

Real-Time Image Compression System Using an Embedded Board

Digital image processing is the use of computer algorithms to improve image quality, to extract or add information. Image compression is a part of image processing and is used to reduce the quantity of data to store. This paper presents the implementation of image compression operations on low cost embedded systems (Raspberry Pi 3, Arduino Uno R3). Motivated by the work of Tchagna et al. (DOI: 10.5815/ijigsp.2018.11.05), we proposed within this paper a real-time implementation of the compression algorithm on embedded boards. Our investigation in this paper is to make a real-time compression system able to capture an image, apply compression algorithm and save compression image on an SD card (for Arduino or Raspberry) or sent directly compressed image to cloud. Compression system with a Raspberry Pi basically used a webcam USB camera to capture the images, the compression function based on python language, and a function to store compressed image to an SD card or to Cloud. The compression system with Arduino used an SD card where the image to be compressed are stored, an external SRAM chip, and an Ethernet shield. The proposed hardware system can decompress the image. In opposition to the approach adopted in the literature, all the results presented within this work use the vector quantization. Eight images have been used to evaluate and compared the compression time for each board according to codebook size used during vector quantization step. Based on our results, we remark that compression and decompression time using Raspberry Pi is lower than compression and decompression time using Arduino. Raspberry Pi offers many possibilities and its processor is bigger than Arduino processor. This justifies the obtained compression and decompression time using Raspberry Pi compared to those with Arduino.

Deep Patch Representations with Shared Codebook for Scene Classification

Scene classification is a challenging problem. Compared with object images, scene images are more abstract, as they are composed of objects. Object and scene images have different characteristics with different scales and composition structures. How to effectively integrate the local mid-level semantic representations including both object and scene concepts needs to be investigated, which is an important aspect for scene classification. In this article, the idea of a sharing codebook is introduced by organically integrating deep learning, concept feature, and local feature encoding techniques. More specifically, the shared local feature codebook is generated from the combined ImageNet1K and Places365 concepts (Mixed1365) using convolutional neural networks. As the Mixed1365 features cover all the semantic information including both object and scene concepts, we can extract a shared codebook from the Mixed1365 features, which only contain a subset of the whole 1,365 concepts with the same codebook size. The shared codebook can not only provide complementary representations without additional codebook training but also be adaptively extracted toward different scene classification tasks. A method of fusing the encoded features with both the original codebook and the shared codebook is proposed for scene classification. In this way, more comprehensive and representative image features can be generated for classification. Extensive experimentations conducted on two public datasets validate the effectiveness of the proposed method. Besides, some useful observations are also revealed to show the advantage of shared codebook.

A proposal of prior probability-oriented clustering in feature encoding strategies.

Codebook-based feature encodings are a standard framework for image recognition issues. A codebook is usually constructed by clusterings, such as the k-means and the Gaussian Mixture Model (GMM). A codebook size is an important factor to decide the trade-off between recognition performance and computational complexity and a traditional framework has the disadvantage to image recognition issues when a large codebook; the number of unique clusters becomes smaller than a designated codebook size because some clusters converge to close positions. This paper focusses on the disadvantage from a perspective of the distribution of prior probabilities and presents a clustering framework including two objectives that are alternated to the k-means and the GMM. Our approach is first evaluated with synthetic clustering datasets to analyze a difference to traditional clustering. In the experiment section, although our approach alternated to the k-means generates similar results to the k-means results, our approach is able to finely tune clusters for our objective. Our approach alternated to the GMM significantly improves our objective and constructs intuitively appropriate clusters, especially for huge and complicatedly distributed samples. In the experiment on image recognition issues, two state-of-the-art encodings, the Fisher Vector (FV) using the GMM and the Vector of Locally Aggregated Descriptors (VLAD) using the k-means, are evaluated with two publicly available image datasets, the Birds and the Butterflies. For the results of the VLAD with our approach, the recognition performances tend to be worse compared to the original VLAD results. On the other hand, the FV using our approach is able to improve the performance, especially in a larger codebook size.

Beam Codebook Design for 5G mmWave Terminals

A beam codebook of 5G millimeter wave (mmWave) for data communication consists of multiple high-peak-gain beams to compensate the high pathloss at the mmWave bands. These beams also have to point to different angular directions, such that by performing beam searching over the codebook, a good mmWave signal coverage over the full sphere around the terminal (spherical coverage) can be achieved. A model-based beam codebook design that assumes ideal omni-directional antenna pattern, and neglects the impact of terminal housing around the antenna, does not work well because the radiation pattern of a practical mmWave antenna combined with the impact of terminal housing is highly irregular. In this paper, we propose a novel and efficient data-driven method to generate a beam codebook to boost the spherical coverage of mmWave terminals. The method takes as inputs the measured or simulated electric field response data of each antenna and provides the codebook according to the requirements on the codebook size, spherical coverage, etc. The method can be applied in a straightforward manner to different antenna type, antenna array configuration, placement and terminal housing design. Our simulation results show that the proposed method generates a codebook better than the benchmark and 802.15.3c codebooks in terms of the spherical coverage.

An Efficient SCMA Codebook Design Based on Lattice Theory for Information-Centric IoT

Different from traditional communication systems, information-centric Internet of things (IC-IoT) as a novel smart paradigm needs to guarantee end-to-end connectivity for rapidly growing smart devices. How to meet the demands of massive connection has become a key problem for IC-IoT. Sparse code multiple access (SCMA) as a code-domain non-orthogonal multiple access (NOMA) technique has been intensively investigated. With SCMA, each user employs different sub-carrier frequencies for transmission, and different users can share the same sub-carrier frequency via superposition coding. The spectrum efficiency is thus improved. However, designing large-scale codebook sets is still an open problem for SCMA. In this paper, a new lattice-based codebook design method is proposed via mother constellation optimization. First, the optimization problem of the mother constellation is formulated. Through analysis, the problem can be converted into two sub-problems. Accordingly, we first use the lattice theory to find a constellation containing infinite points with large coding gain. After that, we search for a boundary that contains a set of points via spherical packing. Such an approach enables us to obtain a real constellation satisfying a power saving criterion. Secondly, the mother constellation with large power variance is obtained from the real multi-dimensional constellation. Finally, lattice-based codebooks are generated by combining the mother constellation and the mapping matrices with constellation rotation. Simulations demonstrate that the designed codebooks have improved bit error rate (BER) performance with large codebook size, especially at high signal to noise ratio (SNR).

Multi‐beam receive scheme for millimetre wave wireless communication system

Different from the conventional receive scheme for millimetre wave (mmWave) communication, this study proposes a multi-beam receive scheme to improve the quality of the received signal and enhance the robustness of the system. The authors proposal is firstly formulated as an optimisation problem, where each signal received by different beams is combined through phase compensation to harvest more transmission energy. Then the original problem is divided into a series of sub-problems and an analytical solution is further obtained for each sub-problem. Furthermore, considering the spatial sparsity of the mmWave channel, they propose further a low-computational complexity algorithm by choosing a few candidate codewords with non-negligible receive signal power. Numerical results show that their proposal achieves remarkable performance improvements even with a small size codebook and is more robust for different scenarios, especially for a non-line of sight scenario. Compared with doubling receive antennas to obtain diversity gain, their proposal obtains larger signal-to-noise ratio gain while using fewer hardware resources.

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.

Sparsity-aided codebook for limited feedback in FDD massive MIMO system

Channel state information at transmitter (CSIT) is vital for massive multiple-input-multiple-output (MIMO) system to achieve unprecedented improvement in capacity and spectral efficiency. Unfortunately, CSIT is acquired via channel feedback in frequency-division duplexing (FDD) system, which suffers from large volume of feedback overhead and size of codebook, since they increase with the number of antennas at transmitter. In this paper, by exploiting the channel sparsity in virtual angular domain, a sparsity-aided codebook is proposed for FDD massive MIMO system to alleviate the channel feedback overhead and codebook size. Specifically, a sparsity aided limited channel feedback model is proposed, where the high dimensional channel is decomposed to the limited linear combination with channel coefficients in virtual angular domain. Then, the high dimensional channel can be quantized by the proposed codebook in a more precise way with significantly small codebook size and fed back in low overhead within virtual coherence time. Furthermore, the upper bound of inter-user interference and the per user rate loss incurred by the proposed sparsity-aided codebook are theoretically analyzed, in which the required quantization bits only increase linearly with the channel sparsity level rather than number of BS antennas when the rate loss is limited to a constant level.

Fingerprint Classification using KMCG Algorithm under Varying Window and Codebook Sizes

Fingerprint Classification using KMCG Algorithm under Varying Window and Codebook Sizes

Joint image coding and lossless data hiding in VQ indices using adaptive coding techniques

In this paper, we propose a novel lossless data hiding method for vector quantization (VQ) compressed images. The existing methods sort the codebook based on the intensity of codewords to obtain rearranged indices (RI). The indices in RI are predicted, and the secret bits, indicators, and prediction errors are then encoded to construct an output code stream (CS). However, the neighboring indices of RI might not have similar intensity, which could impede the coding efficiency. Moreover, the equal-length indicators used in the existing methods might not efficiently represent the cases categorized in the encoding stage. As a result, the size of CS in these methods is likely to increase unnecessarily. The proposed method rearranges the indices to make their spatial correlations stronger, so that the obtained RI is more advantageous for predictive coding. A weight-controlled least square estimator is employed for prediction, and the indicators are better assigned according to the occurrence frequencies of prediction errors. The run-length coding technique is adopted to further reduce the code length. Experimental results show that the proposed method effectively reduces the bitrate for various codebook sizes.

Transmit Precoding Based Low Complexity Codebooks and Finite Feedback Rate for Space Shift Keying MIMO Systems

Transmit Precoding (TPC) considerably improves the performance of Space Shift Keying (SSK) systems. Codebook-based TPC is able to tackle several difficulties encountered by non-codebook TPC techniques. Channel State Information (CSI) is required at the receiver only, and the index of the best codeword is easily identified and fed-back from receiver to transmitter with low rate message. Motivated by these advantages, this paper contributes to the codebook-based TPC in the following aspects. First, the paper presents a Factorized form of the Full-Combination (FC) codebook with phase rotation only, and shows that only four phases are needed, which significantly simplifies practical implementation. As a second contribution, the paper introduces two statistically filtered codebooks: Index-Filtration FC and Average-Filtration FC. These codebooks considerably reduce system complexity due to their small codebook size, while maintaining almost the same BER of the FC codebook. As a third contribution, this paper proposes a new codeword selection criteria based on Long term statistics of the CSI (LCSI), instead of the conventional criteria using instantaneous CSI. This reduces the feedback rate of the index of the selected codeword while providing performance improvement over non-precoded SSK. Simulation results show the effectiveness of the proposed codebooks and the performance improvement with LCSI selection criteria.

SIFT Meets CNN: A Decade Survey of Instance Retrieval.

In the early days, content-based image retrieval (CBIR) was studied with global features. Since 2003, image retrieval based on local descriptors (de facto SIFT) has been extensively studied for over a decade due to the advantage of SIFT in dealing with image transformations. Recently, image representations based on the convolutional neural network (CNN) have attracted increasing interest in the community and demonstrated impressive performance. Given this time of rapid evolution, this article provides a comprehensive survey of instance retrieval over the last decade. Two broad categories, SIFT-based and CNN-based methods, are presented. For the former, according to the codebook size, we organize the literature into using large/medium-sized/small codebooks. For the latter, we discuss three lines of methods, i.e., using pre-trained or fine-tuned CNN models, and hybrid methods. The first two perform a single-pass of an image to the network, while the last category employs a patch-based feature extraction scheme. This survey presents milestones in modern instance retrieval, reviews a broad selection of previous works in different categories, and provides insights on the connection between SIFT and CNN-based methods. After analyzing and comparing retrieval performance of different categories on several datasets, we discuss promising directions towards generic and specialized instance retrieval.