Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Data Mining - Concepts and Techniques.

Compressive Sensing (CS) is a new sensing modality, which compresses the signal being acquired at the time of sensing. Signals can have sparse or compressible representation either in original domain or in some transform domain. Relying on the sparsity of the signals, CS allows us to sample the signal at a rate much below the Nyquist sampling rate. Also, the varied reconstruction algorithms of CS can faithfully reconstruct the original signal back from fewer compressive measurements. This fact has stimulated research interest toward the use of CS in several fields, such as magnetic resonance imaging, high-speed video acquisition, and ultrawideband communication. This paper reviews the basic theoretical concepts underlying CS. To bridge the gap between theory and practicality of CS, different CS acquisition strategies and reconstruction approaches are elaborated systematically in this paper. The major application areas where CS is currently being used are reviewed here. This paper also highlights some of the challenges and research directions in this field.

A Systematic Review of Compressive Sensing: Concepts, Implementations and Applications

Most image encryption algorithms based on low-dimensional chaos systems bear security risks and suffer encryption data expansion when adopting nonlinear transformation directly. To overcome these weaknesses and reduce the possible transmission burden, an efficient image compression–encryption scheme based on hyper-chaotic system and 2D compressive sensing is proposed. The original image is measured by the measurement matrices in two directions to achieve compression and encryption simultaneously, and then the resulting image is re-encrypted by the cycle shift operation controlled by a hyper-chaotic system. Cycle shift operation can change the values of the pixels efficiently. The proposed cryptosystem decreases the volume of data to be transmitted and simplifies the keys distribution simultaneously as a nonlinear encryption system. Simulation results verify the validity and the reliability of the proposed algorithm with acceptable compression and security performance.

Image compression–encryption scheme based on hyper-chaotic system and 2D compressive sensing

The existing ways to encrypt images based on compressive sensing usually treat the whole measurement matrix as the key, which renders the key too large to distribute and memorize or store. To solve this problem, a new image compression–encryption hybrid algorithm is proposed to realize compression and encryption simultaneously, where the key is easily distributed, stored or memorized. The input image is divided into 4 blocks to compress and encrypt, then the pixels of the two adjacent blocks are exchanged randomly by random matrices. The measurement matrices in compressive sensing are constructed by utilizing the circulant matrices and controlling the original row vectors of the circulant matrices with logistic map. And the random matrices used in random pixel exchanging are bound with the measurement matrices. Simulation results verify the effectiveness, security of the proposed algorithm and the acceptable compression performance.

Novel image compression–encryption hybrid algorithm based on key-controlled measurement matrix in compressive sensing

In this paper, a new method for the efficient clustering of human activity data is proposed. Unlike the traditional human activity clustering approaches, our method relies on the skeletal data recorded with the help of motion capture (mocap) systems to achieve the goal. The proposed method is structured around the kernel-based robust covariance descriptor. By introducing a data re-framing technique that efficiently utilizes the temporal properties of the human activity data, we have alleviated the data redundancy and insufficiency issues associated with action sequences. The optimization model developed encompasses the combined benefits of low-rank representation and least square regression. The formulation is strengthened by incorporating the temporal dependency of the human activity sequences with the help of a temporal Laplacian regularizer. With the proposed algorithm, a representation matrix is learned from the raw data, which is then used to perform subspace clustering. Experiments conducted on multiple human activity datasets reveal the ability of the proposed method to achieve better clustering results compared to state-of-the-art counterparts.

An Efficient Framework for the Clustering of Human Activity Datausing Kernelized Robust Covariance Descriptors

Hyperspectral imaging technology has great role in performing computer vision tasks efficiently. How-ever the acquired hyperspectral images (HSIs) are contaminated by different types of noises and other unwanted signals. This paper proposes a new tensor svd based low rank decomposition together with spatial spectral total variation (SSTV) regularization for removing the noise artefacts in HSIs. The proposed optimization model uses tensor decomposition to express the correlation among the different frequency bands. The sparse noise is detected using a 𝑙 1 norm, and in addition, a Frobenius norm is added to remove heavy Gaussian noise from the images. A SSTV norm is added to preserve the piecewise smoothness structure in the spatial and spectral domains. An efficient solution for the optimization problem is developed based on the alternating direction method of multipliers (ADMM). From the experiments conducted on noisy HSIs, it can be observed that our method achieves better results compared to the already existing ones.

A New Optimization Model for the Restoration of the Deteriorated Hyperspectral Images

Automatic recognition of speech is an active area of research which provides a smooth platform for human-machine interaction. Complexity and recognition accuracy mainly depend on the selection of signal features and classifier. Commonly used features in the field of speech recognition are mel-frequency cepstral coefficients (MFCCs), line spectral frequencies (LSF), short time energy (STE) and linear prediction coefficients (LPC). In this paper, instead of using these well-established features, sparse feature derived from the dictionary of signal atoms using sparse coding is used for feature extraction. To improve the performance, artificial neural network (ANN) is used for the classification of isolated speech. Experimental results show that the proposed method works better in noisy environment upto 20dB SNR without using any speech enhancement method. To remove heavy background noise, a sparsity based speech enhancement algorithm is also proposed in the preprocessing stage of speech recognition. The proposed algorithm is compared with other popular speech recognition methods and it is observed that the proposed method can achieve better performance than the others.

Dictionary learning based sparse coefficients for speech recognition in noisy environment

Conventional band selection methods, which vectorize the spectral bands of Hyperspectral Imaging (HSI) data, may fail to consider its intrinsic two dimensional structure. Hence, an effective tensor based band selection scheme, using submodule clustering is proposed, where this framework is capable of preserving the spatial structure of individual spectral bands. In the proposed optimization problem, a representation tensor is derived from the self-expressive representation of spectral bands in the HSI paper data. A similarity matrix is constructed from the representation tensor and then, appropriate number of non-redundant bands are determined by Silhouette clustering evaluation criterion. To capture better low rankness and self-expressiveness property, $l_{\frac{1}{2}}$ -induced Tensor Nuclear Norm (TNN) and $l_{\frac{1}{2}}$ norm regularization are incorporated in the formulated optimization model. Further, the reduced band subset is used in the clustering of hyperspectral paper data. Experimental results prove the efficacy of newly formed band subset on the clustering performance.

A Tensor Non-Convex Low Rank And Sparse Constrained Band Selection Scheme For Clustering Of Hyperspectral Paper Data

Particle Swarm Optimization(PSO) technique is proposed to find the optimum firing angle of thyristorized static switch-fed three-phase induction motors which gives better starting performance. Various performance indices considered for obtaining optimum firing angle are peak positive and negative torque pulsation, peak starting current, and acceleration time. Optimum firing angle is taken as the angle at which, negative torque pulsations is zero with minimum acceleration time and other performance indices are having satisfactory values. Results show that optimum firing angle obtained through PSO algorithm is matching with the graphical method, and by setting this firing angle, the machine can be started like direct on line starting with better starting performance.

Sudhish N. George

Papers

An Efficient Framework for the Clustering of Human Activity Datausing Kernelized Robust Covariance Descriptors

A New Optimization Model for the Restoration of the Deteriorated Hyperspectral Images

Dictionary learning based sparse coefficients for speech recognition in noisy environment

A Tensor Non-Convex Low Rank And Sparse Constrained Band Selection Scheme For Clustering Of Hyperspectral Paper Data

Performance optimization of static switch fed-three phase induction motors using PSO