By Topic

Information, Communications and Signal Processing (ICICS) 2013 9th International Conference on

Date 10-13 Dec. 2013

Filter Results

Displaying Results 1 - 25 of 204
  • [Front matter]

    Page(s): 1 - 4
    Save to Project icon | Request Permissions | PDF file iconPDF (3762 KB)  
    Freely Available from IEEE
  • High speed sound source detection based on ADFSD measurement

    Page(s): 1 - 4
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (279 KB) |  | HTML iconHTML  

    In this paper, a detection of semi-unknown high speed moving sound sources technique using only a single microphone in a real environment is proposed. Doppler processing techniques in cooperation with an array of high sensitive and high performance (expensive) microphones are commonly used techniques to detection, identification (classification) and various moving parameters estimation of a known moving sound source as high speed automobiles in a highway and low flying aircrafts. In this paper, it is shown that these techniques are applicable using only one microphone for detecting of a semi-unknown high speed moving sound source except the times that the source moves straightway (straightly toward) or straightaway (straightly away) from the microphone. Therefore, the Acoustic Doppler Frequency Shift Difference (ADFSD) is measured between the consecutive acoustic data frames of the moving sound source which are acquired by means of a single microphone. The simulation results on the truth data which were gathered in an airport, confirm this claim. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of abnormal sound recognition system for indoor environment

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (342 KB) |  | HTML iconHTML  

    In our living environment, there are various types of sounds. According to the uniqueness of sounds, people can further comprehend the surrounding by the sense of hearing. Nowadays, voice recognition had been widely applied in various applications. In this paper, we proposed an abnormal sound recognition system for monitoring indoor sounds. Twenty-four features were extracted from each sound frame. The sequential floating forward selection (SFFS) was then adopted to select high discriminative features. The support vector machine (SVM) was finally used to classify the sounds into six categories (screaming, infants' crying, coughing, glass breaking, laughing and doorbell ringing). From the experiment results, the proposed system can effectively recognize different kinds of abnormal sounds with a high recognition rate. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Relationship between the nonlinear distortions compensation performance and each parameter of the electro-dynamic loudspeaker system

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (131 KB) |  | HTML iconHTML  

    Loudspeaker systems generate nonlinear distortions due to their complex structures. The nonlinear distortions degrade the sound quality. A solution of this problem is to use Mirror filter. Mirror filter is used for the compensation of nonlinear distortions for electro-dynamic loudspeaker systems and is based on the nonlinear differential equations. The design of Mirror filter requires the estimated parameters of a target loudspeaker system. If you obtain the corresponding parameters of a target loudspeaker system and arrange Mirror filter designed using those parameters in front of the loudspeaker, then the nonlinear distortions can be compensated. Hence, the estimated parameters are very important to achieve high compensation performance. In this paper, we examine the relationship between the nonlinear distortions compensation performance and each parameter of the electro-dynamic loudspeaker system. Concretely, we clarify the effects by varying each nonlinear parameter in Mirror filter. Simulation and experimental results demonstrate that the compensation performance for the second order nonlinear distortions depends on a nonlinear parameter of the force factor in loudspeaker systems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Noninvasive assessment of liver viscoelasticity by acoustic radiation force with a rat model

    Page(s): 1 - 4
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (224 KB) |  | HTML iconHTML  

    Most ultrasound elastographic methods utilize a purely elastic model to describe the liver mechanical properties. However, to describe tissue that is dispersive and to obtain an accurate measure of tissue elasticity, the tissue viscoelastic property should be examined. The objective of this study was to investigate the shear viscoelastic characteristics, as measured by ultrasound elastography, of liver fibrosis in a rat model and to evaluate the diagnostic accuracy of viscoelasticity for liver fibrosis at different stages. Liver fibrosis was induced in 37 rats using carbon tetrachloride (CCI4) with 6 rats as controls. Liver viscoelasticity was measured in vitro using shear waves induced by acoustic radiation force. The measured mean values of liver elasticity and viscosity ranged from 0.84 to 3.45 kPa and from 1.12 to 2.06 Pa·s for the F0-F4 fibrosis stages, respectively. The area under receiver operating characteristic (AUROC) values were 0.97 (≥F2), 0.91 (≥F3), and 1.00 (F4) for elasticity and 0.91 (≥F2), 0.79 (≥F3), and 0.74 (F4) for viscosity, respectively. The results confirmed that the shear wave velocity was dispersive in frequency, suggesting a viscoelastic model to describe liver fibrosis. The study suggests that although viscosity is not as good as elasticity for fibrosis staging, it is important to consider viscosity for the accurate estimation of elasticity and it may provide other mechanical insights into liver tissues. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Map-Reduce based fast speaker recognition

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (205 KB) |  | HTML iconHTML  

    In text-independent speaker identification, there are a large number of likelihood computations, especial in large population. To speed up the recognition, we proposed a lightweight algorithm called CBF (Codebook Filtering). CBF provides two phase of speaker pruning to accelerate the speaker recognition. To make CBF could process large population, this paper implements CBF on Map-Reduce framework. In this approach, we encountered some problems, such as how to balance accuracy and speed-up of algorithm. This paper provides a mechanism of parameter consulting to archieve satisfactory accuracy and speed-up factor. To verify algorithm, we implement it on Phoenix, a Map-Reduce framework on multi-core. As the result of experiment, this approach has increased the speed-up factor of CBF obviously. The speed-up factor reaches 40.2 when the accuracy keeps 94.98%. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Glottal activity detection using Finite Rate of Innovation methods

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (132 KB) |  | HTML iconHTML  

    A novel algorithm which accurately determines the glottal closing instants and glottal opening instants from the Differentiated Electroglottographic signals is presented.The glottal closing instants and glottal opening instant are modeled as non-bandlimited signals and methods for sampling and reconstructing signals with Finite Rate of Innovation are then used to recover the glottal closing instants and glottal opening instants. In comparison with existing methods, the proposed scheme achieves improved performance by providing 99% reliable estimates of glottal closing and opening instants respectively. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Face recognition using Deep PCA

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (225 KB) |  | HTML iconHTML  

    In this paper, we propose a new deep learning method called Deep PCA (DPCA) for face recognition. Our method performs deep learning through hierarchically projecting face image vectors to different feature subspaces and obtaining the representations from different projections. Specifically, we perform a two-layer ZCA whitening plus PCA structure for learning hierarchical features. The whole feature representation of each face image can be extracted by concatenating the representations from the first and second layers. Our approach learns deep representations from the data, by utilizing information from the first layer to produce a new and different representation, making it more discriminative. Experimental results on the widely used FERET and AR databases are presented to show the efficiency of the proposed approach. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Spectral-spatial classification of hyperspectral image using autoencoders

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (570 KB) |  | HTML iconHTML  

    Hyperspectral image (HSI) classification is a hot topic in the remote sensing community. This paper proposes a new framework of spectral-spatial feature extraction for HSI classification, in which for the first time the concept of deep learning is introduced. Specifically, the model of autoencoder is exploited in our framework to extract various kinds of features. First we verify the eligibility of autoencoder by following classical spectral information based classification and use autoencoders with different depth to classify hyperspectral image. Further in the proposed framework, we combine PCA on spectral dimension and autoencoder on the other two spatial dimensions to extract spectral-spatial information for classification. The experimental results show that this framework achieves the highest classification accuracy among all methods, and outperforms classical classifiers such as SVM and PCA-based SVM. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A quantitative metric of visual-words separability for a more discriminative visual vocabulary in an unsupervised manner

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (216 KB) |  | HTML iconHTML  

    The task of visual vocabulary construction plays an important role in the bag-of-words based pattern analysis and robotic applications. A discriminative vocabulary generation in unsupervised case is an open issue for reducing perceptual aliasing in image matching based applications. In this paper, we present a scheme to evaluate the discriminative power of each visual word quantitatively in terms of Mahalanobis separability, and a discriminative visual vocabulary is obtained through adaptively updating the poor discriminative visual words in an unsupervised manner. The effectiveness of our metric is demonstrated in the experiment of loop-closure detection under strong perceptual aliasing condition in both indoor and outdoor image sequences. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Two dimension nonnegative partial least squares for face recognition

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (158 KB) |  | HTML iconHTML  

    For benefiting from incorporating the class information, partial least squares (PLS) and its two dimension version (2DPLS) have been widely employed in face recognition when extracting principal components. However, currently popular statistic methods, such as principal component analysis (PCA) and linear discriminant analysis (LDA), only learn holistic, not parts-based, representations which ignore available local features for face recognition. In this paper, we propose a novel approach to extract the facial features called two dimension nonnegative partial least squares (2DNPLS). Our approach can grab the local features via adding non-negativity constraint to the 2DPLS, and can also reserve the advantages of 2DPLS, which are both inherent structure and class information of images. For evaluating our approach's performance, a series of experiments were conducted on two famous face image databases include ORL and Yale face databases, which demonstrate that our proposed approach outperforms the compared state-of-art algorithms. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Skeleton based shape matching using reweighted random walks

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (278 KB) |  | HTML iconHTML  

    Shape matching is a very important issue and challenging task in computer vision. In this paper, the problem of finding a matching between two shapes is addressed by establishing correspondences between two their skeleton graphs based on random walk framework. We first propose a novel skeleton graph model in which nodes represent end-nodes of skeleton while edges describe relations between two end-nodes. Matching between two skeletons is then formulated as graph matching, which is solved by ranking on an association graph via random walks. By applying the random walks with reweighting jumps on the association skeleton graph, the proposed method can collect potential matches, eliminating the unreliable matches, which are affected by noise and distortion. Comparative experiments on several benchmark data sets show that the proposed method produces more accurate results than the previous works. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive Weighted Multi-Element Collaborative Representation for Visual Classification

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (190 KB) |  | HTML iconHTML  

    Adaptive Weighted Multi-Element Collaborative Representation for Visual Classification is proposed in this paper. To address the weak discriminative power of SRC (sparse representation classifier) method, we propose using multiple elements to represent each element and construct multiple collaborative representation for classification. To reflect the different element with different importance and discriminative power, we present an adaptive weighted residuals method to linearly combine different element representations for classification. Experimental results demonstrate the effectiveness and better classification accuracy of our proposed method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Price and renewable aware geographical load balancing technique for data centres

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (125 KB) |  | HTML iconHTML  

    We propose to address the critical issue pertaining to the energy consumption in data centers, which are the essential backbone of various IT functionalities provided over the Internet. We carry out extensive investigations on the avenues to reduce the energy cost incurred by the cloud service providers which run the requests from the customers subject to compliance with the predetermined Service Level Agreement (SLA). Secondly, we consider the social issue of increasing the renewable energy penetration into the grid. We achieve both the economical objective of energy cost reduction and social objective of increasing the penetration of renewable energy by employing a price and renewable aware load distribution algorithm. We formulate and solve a convex optimization problem, for which there exist computationally efficient algorithms. We carry out rigorous simulation experiments and demonstrate that our proposed method substantially outperforms other commonly employed algorithms. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • One-step and multi-step ahead stock prediction using backpropagation neural networks

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (186 KB) |  | HTML iconHTML  

    Forecasting stock price with traditional time series methods has proven to be difficult. An artificial neural network is probably more suitable for this task, since no assumption of a mathematical model has to be made prior to the forecasting process. Furthermore, a neural network has the ability to extract the main influential factors from large sets of data, which is often required for a successful stock prediction task. In this paper, we explore one-step ahead and multi-step ahead predictions and compare with previous work. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • One-vs-all for class imbalance learning

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (174 KB) |  | HTML iconHTML  

    The performance of support vector machines (SVMs) can deteriorate when the number of samples in one class is much greater than that in the other. Existing methods tackle this problem by modifying the learning algorithms or resampling the datasets. In this paper, we propose a new method called one-vs-all for class imbalance learning (OVACIL) which neither modifies the SVM learning algorithms nor resamples the datasets. In the OVACIL method, we re-group a given imbalanced dataset into a number of new datasets comprising of all the original samples and train standard SVM classifiers using each of the datasets. The output scores of these classifiers on a testing sample are then compared and a final decision is made without a fixed decision threshold. This comparison is not biased toward any particular class, resulting in high accuracies of both classes. The Gmean and Fmeasure values obtained by OVACIL on 18 real-world imbalanced datasets surpass the previous best values reported by other state-of-the-art CIL methods on most of these datasets. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A modified Fuzzy C-means algorithm with symmetry information for MR brain image segmentation

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (442 KB) |  | HTML iconHTML  

    In this paper, we present a novel modified Fuzzy C-means algorithm with symmetry information to reduce the effect of noise in brain tissue segmentation in magnetic resonance image (MRI). We integrate brain's bilateral symmetry into the conventional Fuzzy C-means (FCM) as an additional term. In experiments, some synthetic images, and both simulated and real brain images were used to investigate the robustness of the method against noise. Finally, the method was compared with the conventional FCM algorithm. Results show the viability of the approach and the preliminary investigation appears promising. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Stereovision based 3D hand gesture recognition for pervasive computing applications

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (241 KB) |  | HTML iconHTML  

    Hand gesture recognition, being one of the most intuitive means of Human Computer Interface has spawned many applications in the area of pervasive computing devices. This paper presents a solution for hand gesture recognition problem based on 3D stereo imaging techniques employing a low complexity algorithm on a low cost sensor. This method significantly improves upon conventional 2D techniques which are limited in their approach due to various reasons which include but are not limited to occlusion, difficulty in establishing classifiers, segmentation problems using template matching and noise. The innovation here is the use of a hybrid method that combines a simple state machine, skin tone detection and stereo cameras to provide a fast method for detecting simple hand gestures that can be implemented on most low cost System on Chips (SoC). The experimental results from the test setup are promising. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Apply Run-length encoding on pixel differences to do image hiding

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (431 KB) |  | HTML iconHTML  

    In 2011, Chen used Run Length Coding to record the repetition times of continuous data with the same value. These recorded data were then embedded in cover images by Module-based Substitution. In order to increase the quantity of repetition times of the data with the same value, in this paper, the pixel values are changed into difference values by using Median Edge Predictor (MED). The range of all possible difference values is subsequently divided into two regions. The difference values in the regions are reassigned new positive values, put through Run Length Coding, and embedded into cover images by Modulebased Substitution. According to the experimental results, the quality of the embedded image in the proposed method is better than that in Chen's method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A random increasing sequence hash chain and smart card-based remote user authentication scheme

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (114 KB) |  | HTML iconHTML  

    In 2012, Hsieh and Leu proposed a remote user authentication scheme. They use cryptographic hash functions to design their scheme. By designing a method to analyse one-way hash functions, we propose two attacks on their scheme. One is masquerade attack and another is guessing attack. In addition, we point out that their scheme provides no user anonymity and propose a remote user authentication scheme. Our scheme can provide user anonymity. Based on the one-wayness of hash functions and random increasing sequence-based hash chain, our proposed scheme can resist outsider masquerade attack and insider guessing attack. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A high payload data embedding scheme using dual stego-images with reversibility

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (162 KB) |  | HTML iconHTML  

    Lee and Huang proposed a dual-image reversible data hiding method to protect the confidential data in 2011. The image quality obtained by the method is good because each pixel in the cover image is modified by only adding or subtracting one to generate the stego image. However, in their method, each set of four cover pixels can only embed five confidential bits, thereby restricting the embedding rate. To overcome the above drawback, this work develops a reversible data hiding method based on the magic matrix. By using the proposed method, each set of two cover pixels can be used to embed at least three confidential bits. In the embedding procedure, the maximum modification level of the cover pixel is four. These features imply that the proposed method can embed a large amount of confidential data; in addition, our stego image can achieve good visual quality. Experimental results indicate that the embedding rate and the image quality of the proposed method are 1.55 bpp and 39.89 dB, which exceed those of recently presented methods. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A user authentication scheme on multi-server environments for cloud computing

    Page(s): 1 - 4
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (180 KB) |  | HTML iconHTML  

    In the cloud computing, user authentication is an important security mechanism because it provides the functions of authentication, authorization, and accounting for cloud servers. However, the previously proposed user authentication schemes have many security and efficiency problems. In addition, these schemes can only work on single server environments. To solve the above problems, we propose a new user authentication scheme on multi-server environments for cloud computing in this paper. Compared with the related works, the proposed scheme has less computation costs. In addition, the proposed scheme can be applied to multi-server environments because the ID-based concept is used. Therefore, the proposed user authentication scheme provides efficiency, security, and flexibility for the cloud computing applications. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • High efficient joint fingerprinting and decryption for multimedia distribution

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1100 KB) |  | HTML iconHTML  

    In fingerprint-based secure multimedia distribution systems, the encryption process is performed at the sender side before transmission, and the fingerprint embedding and decryption processes are performed at the receiver side simultaneously. In this paper, a low complexity joint fingerprint and decryption (JFD) scheme based on exclusive-or operations has been proposed, which can be effectively used for traitor tracing. The scheme encrypts only sub-pixels of the entire image using a tiny lookup table generated by a stream cipher. In the decryption process, some pixels can be completely recovered and the fingerprinting process depends on four different directions of sub-image blocks. The method successfully leaves the user's fingerprint during decryption and produces slightly different media copies. The experimental results show that the proposed scheme is highly effective in terms of perceptual security and fingerprinted visual quality. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Opportunities for Software-Defined Networking in Smart Grid

    Page(s): 1 - 5
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (336 KB) |  | HTML iconHTML  

    The large-scale, heterogeneous, and distributed nature of the Smart Grid poses many challenges to be overcome from communication networking to autonomous control and management. The underlying infrastructure of Smart Grid must be efficient and reliable in transmitting large amounts of real-time data, scalable and flexible in aggregating resources, and secured and convenient in providing management interfaces to upper layer application systems. Meanwhile, the recent rapidly developing technology of Software Defined Networking (SDN) is perceived to have tremendous potential for utilization by the underlying infrastructure. By abstracting control functionalities from underlying packet forwarding hardware to an external software controller, SDN offers a high degree of flexibility for implementing novel networking solutions to improve performances of distributed systems in large, complex network environments such as Smart Grid. In this paper, three potential use cases are presented to examine the opportunities for SDN technology in Smart Grid. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Enabling technologies and signal processing for NG-PON2 and future WDM-PON

    Page(s): 1 - 2
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (77 KB) |  | HTML iconHTML  

    Silicon-based optical components are now mature technologies; and they can be used in the time-and-wavelength-division-multiplexed passive-optical-network (TWDM-PON) optical networking unit (ONU) to increase the level of integration for cost reduction. We will also discuss the further migration form the NG-PON2 to the future pure WDM-PON. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.