Showing papers on "Codebook published in 2010"

Visual Word Ambiguity

Abstract: This paper studies automatic image classification by modeling soft assignment in the popular codebook model. The codebook model describes an image as a bag of discrete visual words selected from a vocabulary, where the frequency distributions of visual words in an image allow classification. One inherent component of the codebook model is the assignment of discrete visual words to continuous image features. Despite the clear mismatch of this hard assignment with the nature of continuous features, the approach has been successfully applied for some years. In this paper, we investigate four types of soft assignment of visual words to image features. We demonstrate that explicitly modeling visual word assignment ambiguity improves classification performance compared to the hard assignment of the traditional codebook model. The traditional codebook model is compared against our method for five well-known data sets: 15 natural scenes, Caltech-101, Caltech-256, and Pascal VOC 2007/2008. We demonstrate that large codebook vocabulary sizes completely deteriorate the performance of the traditional model, whereas the proposed model performs consistently. Moreover, we show that our method profits in high-dimensional feature spaces and reaps higher benefits when increasing the number of image categories.

Local features are not lonely – Laplacian sparse coding for image classification

Abstract: Sparse coding which encodes the original signal in a sparse signal space, has shown its state-of-the-art performance in the visual codebook generation and feature quantization process of BoW based image representation. However, in the feature quantization process of sparse coding, some similar local features may be quantized into different visual words of the codebook due to the sensitiveness of quantization. In this paper, to alleviate the impact of this problem, we propose a Laplacian sparse coding method, which will exploit the dependence among the local features. Specifically, we propose to use histogram intersection based kNN method to construct a Laplacian matrix, which can well characterize the similarity of local features. In addition, we incorporate this Laplacian matrix into the objective function of sparse coding to preserve the consistence in sparse representation of similar local features. Comprehensive experimental results show that our method achieves or outperforms existing state-of-the-art results, and exhibits excellent performance on Scene 15 data set.

Codebook: discovering and exploiting relationships in software repositories

Abstract: Large-scale software engineering requires communication and collaboration to successfully build and ship products. We conducted a survey with Microsoft engineers on inter-team coordination and found that the most impactful problems concerned finding and keeping track of other engineers. Since engineers are connected by their shared work, a tool that discovers connections in their work-related repositories can help. Here we describe the Codebook framework for mining software repositories. It is flexible enough to address all of the problems identified by our survey with a single data structure (graph of people and artifacts) and a single algorithm (regular language reachability). Codebook handles a larger variety of problems than prior work, analyzes more kinds of work artifacts, and can be customized by and for end-users. To evaluate our framework's flexibility, we built two applications, Hoozizat and Deep Intellisense. We evaluated these applications with engineers to show effectiveness in addressing multiple inter-team coordination problems.

Semantics-Preserving Bag-of-Words Models and Applications

Abstract: The Bag-of-Words (BoW) model is a promising image representation technique for image categorization and annotation tasks. One critical limitation of existing BoW models is that much semantic information is lost during the codebook generation process, an important step of BoW. This is because the codebook generated by BoW is often obtained via building the codebook simply by clustering visual features in Euclidian space. However, visual features related to the same semantics may not distribute in clusters in the Euclidian space, which is primarily due to the semantic gap between low-level features and high-level semantics. In this paper, we propose a novel scheme to learn optimized BoW models, which aims to map semantically related features to the same visual words. In particular, we consider the distance between semantically identical features as a measurement of the semantic gap, and attempt to learn an optimized codebook by minimizing this gap, aiming to achieve the minimal loss of the semantics. We refer to such kind of novel codebook as semantics-preserving codebook (SPC) and the corresponding model as the Semantics-Preserving Bag-of-Words (SPBoW) model. Extensive experiments on image annotation and object detection tasks with public testbeds from MIT's Labelme and PASCAL VOC challenge databases show that the proposed SPC learning scheme is effective for optimizing the codebook generation process, and the SPBoW model is able to greatly enhance the performance of the existing BoW model.

Spatio-temporal context for codebook-based dynamic background subtraction

Abstract: In background subtraction, it is challenging to detect foreground objects in the presence of dynamic background motions. The paper proposes two new algorithms to this problem by improving the codebook model with the incorporation of the spatial and temporal context of each pixel. The spatial context involves the local spatial dependency between neighboring pixels, and the temporal context involves the preceding detection result. Only the spatial context is incorporated into the first algorithm which makes the background representation more compact than the standard codebook. The second algorithm explicitly models the spatio-temporal context with a Markov random field model, thus achieving more accurate foreground detection. Extensive experiments on several dynamic scenes are conducted to compare the two proposed algorithms with each other and with the standard codebook algorithm. (C) 2009 Elsevier GmbH. All rights reserved.

DFT-Based Beamforming Weight-Vector Codebook Design for Spatially Correlated Channels in the Unitary Precoding Aided Multiuser Downlink

Abstract: The DFT-based beamforming weight-vector codebook is considered as an effective design for spatially correlated channels. In this paper, we demonstrate that when the antenna elements are uniformly spaced as well as linearly arranged, and the channels are spatially correlated, the codewords in a DFT-based beaforming weight-vector codebook approximately match the distribution of the optimal beamforming weight-vectors. As a result, the DFT-based codebook is indeed effective. Furthermore, we also demonstrate that if the antenna elements are uniformly spaced and circulary arranged, the statistical distribution of the optimal beamforming weight-vectors becomes different. We will demonstrate that in this scenario the DFT-based codebook will no longer outperform the Grassmannian codebook, which has not been shown in previous studies. Finally, an algorithm is proposed for constructing the DFT-based precoding matrix, which outperforms the conventional algorithm by ensuring the orthogonality of the precoding matrix.

Closed-loop transmission feedback in wireless communication systems

Abstract: A method and apparatus for providing channel feedback is provided herein. During operation a covariance matrix at time t (R) is calculated by the mobile as a function of a received downlink signal. In order to reduce overhead, R is normalized and quantized by the mobile using multiple codebook entries plus at least one constant for quantization. The mobile then transmits the normalized and quantized covariance matrix back to the base station as bit values indicating the selected entries from the codebook plus bit values corresponding to the at least one constant. The base unit then uses the covariance matrix estimate to determine appropriate channel beamforming weights, and instructs transmit beamforming circuitry to use the appropriate weights.

Building compact local pairwise codebook with joint feature space clustering

Abstract: This paper presents a simple, yet effective method of building a codebook for pairs of spatially close SIFT descriptors. Integrating such codebook into the popular bag-of-words model encodes local spatial information which otherwise cannot be represented with just individual SIFT descriptors. Many previous pairing techniques first quantize the descriptors to learn a set of visual words before they are actually paired. Our approach contrasts with theirs in that each pair of spatially close descriptors is represented as a data point in a joint feature space first and then clustering is applied to build a codebook called Local Pairwise Codebook (LPC). It is advantageous over the previous approaches in that feature selection over quadratic number of possible pairs of visual words is not required and feature aggregation is implicitly performed to achieve a compact codebook. This is all done in an unsupervised manner. Experimental results on challenging datasets, namely 15 Scenes, 67 Indoors, Caltech-101, Caltech-256 and MSRCv2 demonstrate that LPC outperforms the baselines and performs competitively against the state-of-the-art techniques in scene and object categorization tasks where a large number of categories need to be recognized.

System and method for channel information feedback in a wireless communications system

Abstract: A system and method for channel information feedback in a wireless communications system is provided. A method for communications device operation includes receiving a pilot transmitted by a controller, computing a channel estimate for a channel between the controller and a communications device, the computing based on the pilot, computing a channel correlation matrix for the channel based on the channel estimate, and transmitting a reduced rank representation of the channel correlation matrix to the controller as a first feedback information. The method also includes adapting a first codebook based on the reduced rank representation of the channel correlation matrix, computing a representation of the channel using the adapted codebook, transmitting the representation of the channel as a second feedback information, and receiving a transmission beamformed based on the first feedback information and the second feedback information.

A Survey of VQ Codebook Generation

Abstract: One of the key roles of Vector Quantization (VQ) is how to generate a good codebook such that the distortion between the original image and the reconstructed image is the minimum. In the past years, many improved algorithms of VQ codebook generation approaches have been developed. In this paper, we present a snapshot of the recent de- veloped schemes. The discussed schemes include mean-distance-ordered partial codebook search (MPS), enhance LBG (ELBG), neural network based techniques, genetic-based algorithms, principal component analysis (PCA) approaches, tabu search (TS) schemes,

Lossy communication of correlated sources over multiple access channels

Abstract: A new approach to joint source-channel coding is presented in the context of communicating correlated sources over multiple access channels. Similar to the separation architecture, the joint source-channel coding system architecture in this approach is modular, whereby the source encoding and channel decoding operations are decoupled. However, unlike the separation architecture, the same codeword is used for both source coding and channel coding, which allows the resulting coding scheme to achieve the performance of the best known schemes despite its simplicity. In particular, it recovers as special cases previous results on lossless communication of correlated sources over multiple access channels by Cover, El Gamal, and Salehi, distributed lossy source coding by Berger and Tung, and lossy communication of the bivariate Gaussian source over the Gaussian multiple access channel by Lapidoth and Tinguely. The proof of achievability involves a new technique for analyzing the probability of decoding error when the message index depends on the codebook itself. Applications of the new joint source-channel coding system architecture in other settings are also discussed.

A Limited Feedback Joint Precoding for Amplify-and-Forward Relaying

Abstract: This paper deals with the practical precoding design for a dual hop downlink with multiple-input multiple-output (MIMO) amplify-and-forward relaying. First, assuming that full channel state information (CSI) of the two hop channels is available, a suboptimal dual hop joint precoding scheme, i.e., precoding at both the base station and relay station, is investigated. Based on its structure, a scheme of limited feedback joint precoding using joint codebooks is then proposed, which uses a distributed codeword selection to concurrently choose two joint precoders such that the feedback delay is considerably decreased. Finally, the joint codebook design for the limited feedback joint precoding system is analyzed, and results reveal that independent codebook designs at the base station and relay station using the conventional Grassmannian subspace packing method is able to guarantee that the overall performance of the dual hop joint precoding scheme improves with the size of each of the two codebooks. Simulation results show that the proposed dual hop joint precoding system using distributed codeword selection scheme exhibits a rate or BER performance close to the one using the optimal centralized codeword selection scheme, while having lower computational complexity and shorter feedback delay.

Learning weights for codebook in image classification and retrieval

Abstract: This paper presents a codebook learning approach for image classification and retrieval. It corresponds to learning a weighted similarity metric to satisfy that the weighted similarity between the same labeled images is larger than that between the differently labeled images with largest margin. We formulate the learning problem as a convex quadratic programming and adopt alternating optimization to solve it efficiently. Experiments on both synthetic and real datasets validate the approach. The codebook learning improves the performance, in particular in the case where the number of training examples is not sufficient for large size codebook.

Image segmentation with topic random field

Abstract: Recently, there has been increasing interests in applying aspect models (e.g., PLSA and LDA) in image segmentation. However, these models ignore spatial relationships among local topic labels in an image and suffers from information loss by representing image feature using the index of its closest match in the codebook. In this paper, we propose Topic Random Field (TRF) to tackle these two problems. Specifically, TRF defines a Markov Random Field over hidden labels of an image, to enforce the spatial coherence between topic labels for neighboring regions. Moreover, TRF utilizes a noise channel to model the generation of local image features, and avoids the off-line process of building visual codebook. We provide details of variational inference and parameter learning for TRF. Experimental evaluations on three image data sets show that TRF achieves better segmentation performance.

A Scalable Limited Feedback Design for Network MIMO Using Per-Cell Product Codebook

Abstract: In network MIMO systems, channel state information is required at the transmitter side to multiplex users in the spatial domain. Since perfect channel knowledge is difficult to obtain in practice, limited feedback is a widely accepted solution. The dynamic number of cooperating BSs and heterogeneous path loss effects of network MIMO systems pose new challenges on limited feedback design. In this paper, we propose a scalable limited feedback design for network MIMO systems with multiple base stations, multiple users and multiple data streams for each user. We propose a limited feedback framework using per-cell product codebooks, along with a low-complexity feedback indices selection algorithm. We show that the proposed per-cell product codebook limited feedback design can asymptotically achieve the same performance as the joint-cell codebook approach. We also derive an asymptotic per-user throughput loss due to limited feedback with per-cell product codebooks. Based on that, we show that when the number of per-user feedback-bits Bk is {O}( NnTnR log2(ρgksum)), the system operates in the noise-limited regime in which the per-user throughput is {O} ( nR log2 (nRρgksum/NnT)). On the other hand, when the number of per-user feedback-bits Bk does not scale with the system SNR ρ, the system operates in the interference-limited regime where the per-user throughput is {O}(nRBk/(NnT)2). Numerical results show that the proposed design is very flexible to accommodate dynamic number of cooperating BSs and achieves much better performance compared with other baselines (such as the Givens rotation approach).

Gesture recognition by learning local motion signatures

Abstract: This paper overviews a new gesture recognition framework based on learning local motion signatures (LMSs) introduced by [5]. After the generation of these LMSs computed on one individual by tracking Histograms of Oriented Gradient (HOG) [2] descriptor, we learn a codebook of video-words (i.e. clusters of LMSs) using k-means algorithm on a learning gesture video database. Then the video-words are compacted to a codebook of code-words by the Maximization of Mutual Information (MMI) algorithm. At the final step, we compare the LMSs generated for a new gesture w.r.t. the learned codebook via the k-nearest neighbors (k-NN) algorithm and a novel voting strategy. Our main contribution is the handling of the N to N mapping between code-words and gesture labels with the proposed voting strategy. Experiments have been carried out on two public gesture databases: KTH [16] and IXMAS [19]. Results show that the proposed method outperforms recent state-of-the-art methods.

Limited feedback with joint CSI quantization for multicell cooperative generalized eigenvector beamforming

Abstract: Existing work on limited feedback for cooperative multicell beamforming quantizes the desired and interfering channel state information (CSI) using separate codebooks. In this paper, it is shown that comparatively higher sum-rates can be obtained by jointly quantizing the desired and interfering CSI using a single codebook. A selection criterion is developed for random vector quantization (RVQ) to show that joint quantization with RVQ yields higher sum-rates than those obtained using separate codebooks. The generalized Lloyd algorithm is then used to generate codebooks using the codeword design strategy proposed in this paper. Simulations are used to show that the proposed joint quantization approaches perform almost as well as the full CSI case.

An Improved Decoding Algorithm for the Davey-MacKay Construction

Abstract: The Deletion-Insertion Correcting Code construction proposed by Davey and MacKay consists of an inner code that recovers synchronization and an outer code that provides substitution error protection. The inner code uses low-weight codewords which are added (modulo two) to a pilot sequence. The receiver is able to synchronise on the pilot sequence in spite of the changes introduced by the added codeword. The original bit-level formulation of the inner decoder assumes that all bits in the sparse codebook are identically and independently distributed. Not only is this assumption inaccurate, but it also prevents the use of soft a- priori input to the decoder. We propose an alternative symbol-level inner decoding algorithm that takes the actual codebook into account. Simulation results show that the proposed algorithm has an improved performance with only a small penalty in complexity, and it allows other improvements using inner codes with larger minimum distance.

MIMO codebook generation

Abstract: Particular codebook entries used for a MIMO system of a lower dimension (i.e., having a relatively low number of antennas) can be used to generate a codebook for a MIMO system of a higher dimension (i.e., having a relatively high number of antennas). The entries in rank 1 of the codebook related to the MIMO system having four transmit antennas are used to newly construct entries for rank 1 through rank 8 of two base matrices for a MIMO base codebook related to eight transmit antennas.

Systems and methods of encoding using a reduced codebook with adaptive resetting

Abstract: To feedback MIMO channel conditions, a codeword from a codebook is selected. To reduce signalling, the codewords are organized into codeword subsets. The receiver signals an index of a codeword into a current codeword subset previously made known to the transmitter. The current codeword subset is adaptively selected based on a threshold criterion. For example, if the best codeword from the current codeword subset is not sufficiently similar to the best codeword in the full codebook, a switch in the codeword subset is made.

Method and apparatus for information feedback and pre-coding

Abstract: Methods and apparatuses for information feedback and precoding have been provided. A method for processing communication data at a user equipment in a wireless communication system may comprise: deriving a spatial correlation matrix R of multiple transmit antennas of a base station based on an obtained downlink channel transmission matrix H ; transforming a precoding codebook F according to the spatial correlation matrix R ; selecting a precoding matrix F s based on the transformed precoding codebook; and feeding back information about the spatial correlation matrix R and information about the selected precoding matrix F s to the base station. A method for data precoding at a base station in a wireless communication system may comprise: obtaining, from a user equipment, information about a spatial correlation matrix R of multiple transmit antennas of the base station and information about a precoding matrix F s selected by the user equipment; determining a desired precoding matrix F R,s based on the obtained information and a precoding codebook; and precoding downlink data to be transmitted to the user equipment with the desired precoding matrix F R,s

Codebook design and structure for multi-granular feedback

Abstract: A first precoding matrix W1 is selected from a first codebook comprising sets of rank specific precoding matrices. The first codebook is characterized by there being fewer precoding matrices associated with higher ranks than associated with lower ranks, and characterized by precoding matrices associated with ranks above a certain rank all being diagonal matrices. The selected first precoding matrix W1 is used to select a rank-specific second precoding matrix W2 from a second codebook, such that the selected first and second precoding matrices form a joint precoder specific to a desired rank. The second codebook is characterized by differently sized precoding matrices associated with each of N total ranks, in which N is an integer greater than one. Information on the joint precoder is reported to a network node over an uplink transmission channel.

Data transmission method and apparatus in multiple antenna system

Abstract: A data transmission method in a multiple antenna system is provided. The method includes: defining a codebook including at least one precoding matrix composed of a plurality of rows and columns, wherein the codebook is at least one of a first type in which all elements of the precoding matrix are non-zero elements, a second type in which any one column of the precoding matrix includes non-zero element and the remaining columns include at least one zero element, and a third type in which all columns of the precoding matrix include at least one zero element; precoding an input symbol by using the defined codebook; and transmitting the precoded symbol.

On distributed antenna systems with limited feedback precoding: Opportunities and challenges

Abstract: The distributed antenna system is a promising technique to enhance system performance in next-generation broadband wireless mobile communications. The application of linear precoding with space time codeword based on channel state information can further enhance the link quality and reduce the receiver complexity in DASs. However, designing the optimal codebook for DAS to convey the precoding information remains an open issue. This article provides a tutorial addressing various issues of DASs with limited feedback, focusing on the challenges in its codebook design. A suboptimal codebook design is then proposed, suitable for DAS. Furthermore, an effective feedback reduction strategy without performance loss is presented by exploiting the channel correlation property. Numerical results are given to demonstrate the effectiveness of the proposed scheme.

On-line adaption of class-specific codebooks for instance tracking

Abstract: In this work, we demonstrate that an off-line trained class-specific detector can be transformed into an instance-specific detector on-the-fly. To this end, we make use of a codebook-based detector [1] that is trained on an object class. Codebooks model the spatial distribution and appearance of object parts. When matching an image against a codebook, a certain set of codebook entries is activated to cast probabilistic votes for the object. For a given object hypothesis, one can collect the entries that voted for the object. In our case, these entries can be regarded as a signature for the target of interest. Since a change of pose and appearance can lead to an activation of very different codebook entries, we learn the statistics for the target and the background over time, i.e. we learn on-line the probability of each part in the codebook belonging to the target. By taking the target-specific statistics into account for voting, the target can be distinguished from other instances in the background yielding a higher detection confidence for the target, see Fig. 1. A class-specific codebook as in [1, 2, 3, 4, 5] is trained off-line to identify any instance of the class in any image. It models the probability of the patches belonging to the object class p ( c=1|L ) and the local spatial distribution of the patches with respect to the object center p ( x|c=1,L ) . For detection, patches are sampled from an image and matched against the codebook, i.e. each patch P(y) sampled from image location y ends at a leaf L(y). The probability for an instance of the class centered at the location x is then given by

Precoding codebook design for single user mimo

Abstract: A transmitter is for use with multiple transmit antennas and includes a precoder unit configured to precode data for a transmission using a precoding matrix selected from a codebook, wherein the codebook corresponds to the following three transmission properties for an uplink transmission: 1) all precoding elements from the precoding matrix have a same magnitude, 2) each precoding element from the precoding matrix is taken from a set of finite values and 3) there is only one non-zero element in any row of the precoding matrix. The transmitter also includes a transmit unit configured to transmit the precoded data.

Rank and precoding indication for mimo operation

Abstract: Certain aspects of the present disclosure relate to a technique for signaling rank and precoding indications in uplink and downlink MIMO operations using codebook and non-codebook based precoding.

New Clustering Algorithm for Vector Quantization using Rotation of Error Vector

Abstract: —The paper presents new clustering algorithm. The proposed algorithm gives less distortion as compared to well known Linde Buzo Gray (LBG) algorithm and Kekre’s Proportionate Error (KPE) Algorithm. Constant error is added every time to split the clusters in LBG, resulting in formation of cluster in one direction which is 135 0 in 2-dimensional 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 0 over 135 . The proposed algorithm takes care of this problem by introducing new orientation every time to split the clusters. The proposed method reduces PSNR by 2db to 5db for codebook size 128 to 1024 with respect to LBG. Keywords-component; Vector Quantization; Codebook; Codevector; Encoding; Compression. I. I NTRODUCTION Exhaustive Search (ES) method gives the optimal result at the World Wide Web Applications have extensively grown since last few decades and it has become requisite tool for education, communication, industry, amusement etc. All these applications are multimedia-based applications consisting of images and videos. Images/videos require enormous volume of data items, creating a serious problem as they need higher channel bandwidth for efficient transmission. Further high degree of redundancies is observed in digital images. Thus the need for image compression arises for resourceful storage and transmission. Image compression is classified into two categories, lossless image compression and lossy image compression technique. Vector quantization (VQ) is one of the lossy data compression techniques[1], [2] and has been used in number of applications, like pattern recognition [3], speech recognition and face detection [4], [5], image segmentation [6-9], speech data compression [10], Content Based Image Retrieval (CBIR) [11], [12], Face recognition[13], [14] iris recognition[15], tumor detection in mammography images [29] etc. VQ is a mapping function which maps k-dimensional vector space to a finite set CB = {C

Limited Feedback Beamforming Systems for Dual-Polarized MIMO Channels

Abstract: Dual-polarized multiple-input multiple-output (MI-MO) antenna systems, where the antennas are grouped in pairs of orthogonally polarized antennas, are a spatially-efficient alternative to single polarized MIMO antenna systems. A limited feedback beamforming technique is proposed for dual-polarized MIMO channels where the receiver has perfect channel knowledge but the transmitter only receives partial information regarding the channel instantiation. The system employs an effective signal-to-noise ratio (SNR) distortion minimizing codebook to convey channel state information (CSI) in the form of beamforming direction. By investigating the average SNR performance of this system, an upper bound on the average SNR distortion is found as a weighted sum of two beamforming distortion metrics. The distortion minimization problem is solved by designing a concatenated codebook. Finally, we propose a codebook switching scheme exploiting the cross-polar discrimination (XPD) statistics. Simulations show that the proposed codebook switching scheme with an XPD dependent concatenated codebook has the ability to adapt to dual-polarized channels.