P. Rajesh Kumar

Robust digital image watermarking based on fuzzy inference system and back propagation neural networks using DCT

This paper presents an image watermarking technique which exploits the human visual system (HVS), two artificial intelligence techniques [fuzzy inference system (FIS) and back propagation neural networks (BPNN)], and discrete cosine transform. The integration of FIS and BPNN results in a hybrid intelligence technique named neuro-fuzzy system, which combines the advantages of both the techniques. This system uses HVS model for calculating the weighing factor by giving the HVS parameters as inputs to the FIS and BPNN. The computed weighing value is used for embedding the watermark with maximum energy and imperceptibility. Compared to the existing methods, the proposed algorithm makes the watermark invisible and robust against numerous watermarking attacks including low-pass filtering, median filtering, rotation, JPEG compression, row–column copying, row–column blanking, image blurring, etc. Performance metrics used for comparison are peak signal-to-noise ratio and normalized cross correlation.

De-noising of ECG raw signal by cascaded window based digital filters configuration

An ECG is a graphic produced by an electrocardiograph that shows its significance in heart disease diagnostics, Bio-metric authentication, Human Computer Interface (HCI), etc. by recording the electrical activity of heart with respect to time. In each application noise elimination is initial step. Normally, the ECG signal is highly sensitive as it varies a lot with time. It is mainly corrupted by noises such as Power line Interference (PLI), Baseline wandering noise and noise due to muscle movements (EMG noise) during data acquisition. These noises add at different frequency range of ECG and make the signal analysis much tougher. Several researches with variety of processing methods have been taking place so as to retain the morphology of ECG signal. In this paper cascading of window based FIR filters are adopted for removing low frequency, mid frequency and high frequency noise from the noisy real ECG in a single step execution. Different ECG signals from MIT-BIH NSR, ECG ID databases are taken and the results were evaluated using MATLAB software. Finally performance measures of Signal to Noise Ratio (SNR), Mean Square Error (MSE) and Power Spectral Density (PSD) of de-noised ECG signal is calculated and compared with noisy real ECG.

Integration of HSV Color Histogram and LMEBP Joint Histogram for Multimedia Image Retrieval

In this paper, HSV color histogram and local maximum edge binary patterns (LMEBP) joint histogram are integrated for content-based image retrieval (CBIR). The local region of image is represented by LMEBP, which are evaluated by taking into consideration the magnitude of local difference between the center pixel and its neighbors. This LMEBP differs from the existing LBP in a manner that it extracts the information based on distribution of edges in an image. Further, the joint histogram is constructed between uniform two rotational invariant first three LMEBP patterns. The color feature is extracted by calculating the histogram on hue (H), saturation (S), and value (V) spaces. The experimentation has been carried out for proving the worth of our algorithm. It is further mentioned that the database considered for experiment is Corel-1K. The results after being investigated show a significant improvement in terms of their evaluation measures as compared to previously available spatial and transform domain methods on their respective databases.

An Efficient Variable Step Size Least Mean Square Adaptive Algorithm Used to Enhance the Quality of Electrocardiogram Signal

The main aim of this paper is to present an efficient method to cancel the noise in the ECG signal, due to various sources, by applying adaptive filtering techniques. The adaptive filter essentially reduces the mean-squared error between a primary input, which is the noisy ECG, and a reference input, which is either noise that is correlated in some way with the noise in the primary input or a signal that is correlated only with ECG in the primary input. The Least Mean Square (LMS) algorithm is familiar and simple to use for cancellation of noises. However, the low convergence rate and low signal to noise ratio are the limitations for this LMS algorithm. To enhance the performance of LMS algorithm, in this paper, we present an efficient variable step size LMS algorithms which will provide fast convergence at early stages and less misadjustment in later stages. Different kinds of variable step size algorithms are used to eliminate artifacts in ECG by considering the noises such as power line interference and baseline wander. The simulation results shows that the performance of the variable step size LMS algorithm is superior to the conventional LMS algorithm, while for sign based, the sign regressor variable step size LMS algorithm is equally efficient as that of variable step size LMS with additional advantage of less computational complexity.

A modified blind image watermarking scheme based on DWT, DCT and SVD domain using GA to optimize robustness

A modified blind image watermarking algorithm using discrete wavelet transform (DWT), discrete cosine transform (DCT) and singular value decomposition (SVD) is proposed in this paper. One level DWT operation is performed on original host image using Haar wavelet and approximation (LL) sub-band is selected to split it into two sub images. DCT and SVD operations are performed in both sub images using 8×8 block size. Based on pixels values of binary watermark image, singular value of first sub image is modified referring singular value of second sub image to embed watermark. Genetic algorithm (GA) technique is adopted to optimize visual quality and robustness of watermarked image. Peak Signal to Noise Ratio (PSNR) and Normalized Cross correlation (NCC) are computed to evaluate imperceptibility and robustness of the watermarked image. Experimental results obtained confirm that the proposed method is efficient and more robust against different attacks, viz. noise insertion, filtering, JPEG compression, histogram equalization, gamma correction, bit plane removal, image sharpening and cropping attacks.

Performance Evaluation of different Image Filtering Algorithms using Image Quality Assessment

The quality assessment of images is meaningful for most video and images applications. Recently a new metric called Region of Interest structure similarity index (ROISSIM) is proposed for assessing image quality with better reflection to human visual characteristics than traditional image quality assessment methods. In this paper, five different filtering algorithms are compared based on the ability to reconstruct noise affected images using ROISSIM. Experimental results give us the quality assessment of filtering algorithms based on region of interest.

Reducing the grating lobes and main lobe width for increasing range resolution using phase and frequency modulated codes

In Radar the good range resolution can be achieved by a short pulse. Short pulse systems are limited resolution because they are narrow band. Using pulse compression technique high range resolution can be achieved in radar system. The barker codes are known as bi-phase codes with minimum side lobes used in phase modulation in pulse compression technique. LFM and NLFM codes are used as frequency modulation. In this paper the wide band signals that is LFM and NLFM codes are modulates Barker codes to get good resolution while getting minimum side lobes and reduction in main lobe width. By introducing LFM in Barker codes grating lobes will be appeared in Autocorrelation function of LFM Barker sequence. To reduce the grating lobes and the main lobe width to get high delay and Doppler resolution, NLFM code is introduced in Barker codes and observed merit factor, side lobe levels, main lobe width and Doppler tolerance from Autocorrelation function as well as ambiguity function.

An efficient signal processing algorithm for accurate detection of characteristic points in Abnormal ECG signals

A lot of information on the normal and pathological physiology of heart can be obtained in the form of ECG. The irregularity of heart resembles the shape of ECG. One cardiac cycle of ECG signal consists of characteristic points P-QRS-T. The amplitudes and intervals values of P-QRS-T segment determine the functioning of heart of every human. If the electrical activity of the heart is irregular, faster, or slower than normal this is considered as abnormality in ECG signals. In this paper a Signal processing algorithm, which is the combination of both Frequency and Time analysis is used to identify P-QRS-T points. R-peak detection is the first step in characteristic points detection, for identifying R-peak, wavelet transform (sym4) decomposition method (Frequency domain) is used. After R-peak detection other characteristic points are detected by tracing to and fro from R-peak (Time domain). In this five cases of ECG signals are tested specifically for a Normal Person; Normal Sinus Rhythm (NSR), Atrial fibrillation (AF), Right Bundle Branch Block (RBBB), Left Bundle Branch Block (LBBB), Cardiac Ischemia (CI) and Sudden Cardiac Arrest (SCA). A total of 73 Abnormal ECG signals are taken from MIT-BIH arrhythmias database and Results are evaluated using MATLAB R2014a software.

Removal of Artifacts from Electrocardiogram Using Efficient Dead Zone Leaky LMS Adaptive Algorithm

The ability to extract high resolution and valid ECG signals from contaminated recordings is an important subject in the biotelemetry systems. During ECG acquisition several artefacts strongly degrades the signal quality. The dominant artefacts encountered in ECG signal such as Power Line Interference, Muscle Artefacts, Baseline Wander, Electrode Motion Artefacts; and channel noise generated during transmission. The tiny features of ECG signal are masked due to these noises. To track random variations in noisy signals, the adaptive filter is used. In this paper, we proposed Dead Zone Leaky Least Mean Square algorithm, Leaky Least Mean Froth algorithm and Median Leaky LMS algorithms to remove PLI and EM artefacts from ECG signals. Based on these algorithms, we derived some sign based algorithms for less computational complexity. We compare the proposed algorithms with LMS algorithm, which shows better performance in weight drift problem, round off error and low steady state error. The simulation results show that Dead Zone Leaky LMS algorithm gives good correlation factor and SNR ratio.

Robust Digital Image Watermarking Scheme in Discrete Wavelet Transform domain using Support Vector Machines

This paper presents a robust and blind watermarking scheme for copyright protection of images in discrete wavelet transform domain based on the support vector machines (SVMs). This scheme is based on the relation between the coefficients in various sub bands in discrete wavelet transform decomposition. The proposed scheme is very secured and robust to various attacks, viz., Low pass Filtering, Salt & Pepper noise, Gamma Correction, JPEG Compression, Row-Column Copying, Row-column blanking, Bit plane removal, Cropping, Resize and Histogram Equalization etc. Experimental results show that the proposed scheme has significant improvements in both robustness and imperceptibility and superior to an algorithm proposed by Li et al. in terms of Normalized Cross correlation (NC) and Peak Signal to Noise Ratio (PSNR).