Ahlad Kumar

Moment domain representation of nonblind image deblurring

In this paper, we propose the use of geometric moments to the field of nonblind image deblurring. Using the developed relationship of geometric moments for original and blurred images, a mathematical formulation based on the Euler-Lagrange identity and variational techniques is proposed. It uses an iterative procedure to deblur the image in moment domain. The theoretical framework is validated by a set of experiments. A comparative analysis of the results obtained using the spatial and moment domains are evaluated using a quality assessment method known as the Blind/Reference-less Image Spatial Quality Evaluator (BRISQUE). The results show that the proposed method yields a higher quality score when compared with the spatial domain method for the same number of iterations.

Geometric moment extraction equipment for image processing applications

An equipment for calculating 2nd, 3rd, and higher order geometric moments by using accumulators, adders, subtractors, and multiplier blocks has been presented. The performance analysis of the proposed equipment with the existing systems in terms of speed and power dissipation has been carried out and has been shown that the computational time to calculate the geometric moments is reduced to half and the power dissipation is reduced by a factor of about 3 at a clock frequency of 10 MHz. The hardware has been implemented in BSIM4.3.0 50 nm technology operating at 1 V and its functionality has been verified using P-Spice simulator.

Tchebichef moment based restoration of Gaussian blurred images

With the knowledge of how edges vary in the presence of a Gaussian blur, a method that uses low-order Tchebichef moments is proposed to estimate the blur parameters: sigma (σ) and size (w). The difference between the Tchebichef moments of the original and the reblurred images is used as feature vectors to train an extreme learning machine for estimating the blur parameters (σ,w). The effectiveness of the proposed method to estimate the blur parameters is examined using cross-database validation. The estimated blur parameters from the proposed method are used in the split Bregman-based image restoration algorithm. A comparative analysis of the proposed method with three existing methods using all the images from the LIVE database is carried out. The results show that the proposed method in most of the cases performs better than the three existing methods in terms of the visual quality evaluated using the structural similarity index.

Nonlocal means image denoising using orthogonal moments.

An image denoising method in moment domain has been proposed. The denoising involves the development and evaluation based on the modified nonlocal means (NLM) algorithm. It uses the similarity of the neighborhood, evaluated using Krawtchouk moments. The results of the proposed denoising method have been validated using peak signal-to-noise ratio (PSNR), a well-known quality measure such as structural similarity (SSIM) index and blind/referenceless image spatial quality evaluator (BRISQUE). The denoising algorithm has been evaluated for synthetic and real clinical images contaminated by Gaussian, Poisson, and Rician noise. The algorithm performs well compared to the Zernike based denoising as indicated by the PSNR, SSIM, and BRISQUE scores of the denoised images with an improvement of 3.1 dB, 0.1285, and 4.23, respectively. Further, comparative analysis of the proposed work with the existing techniques has also been performed. It has been observed that the results are competitive in terms of PSNR, SSIM, and BRISQUE scores when evaluated for varying levels of noise.

A quad tree based method for blurred and non-blurred video text frames classification through quality metrics

Blur is a common artifact in video, which adds more complexity to text detection and recognition. To achieve good accuracies for text detection and recognition, this paper suggests a new method for classifying blurred and non-blurred frames in video. We explore quality metrics, namely, BRISQUE, NRIQA, GPC and SI, in a new way for classification. We estimate the values of these metrics with the help of predefined samples called reference values. To widen the difference between metric values for better classification, we introduce scaling factors as a non-linear sigmoidal function, which considers the metric of each current frame and its reference and results in templates. Based on the characteristics of metrics, the proposed method finds a relationship between the metrics to derive rules for classification. To classify the frame containing local blur, we explore quad tree division with classification rules which divide non-blurred blocks to identify local blur. We use standard databases, namely, ICDAR 2013, ICDAR 2015 and YVT videos for experimentation, and evaluate the proposed method in terms of text detection and recognition rates given by text detection and binarization methods before and after classification.

Complementary metal-oxide semiconductor implementation of digital filters for signal processing applications

Realisations of filters in signal processing using interconnects as delay elements have been presented. Normally, these filters are implemented using switched capacitor technique. However, in digital realisations of these filters, flip flops are used for obtaining the delays. Here, the implementation of these filters using interconnects has been presented. Moreover, filter architectures which acts as basic building blocks for other complex filter structures have been explored and discussed. Spice simulations of these basic building blocks are carried out using BSIM 4.3 50 nm technology with a supply voltage of 1 V.

Stability of a Two-Stage Opamp: A New Perspective

ABSTRACT A new insight into the cause of ringing is presented, which is attributed to the inductance created at the input of the second stage of the operational amplifier. This approach has been verified through the simulation results carried out using BSIM 4.3 50 nm technology.

Integrated Analysis of Well Logs and Seismic Data for Reservoir Characterization to Estimate Hydrocarbon

The Main objective of oil industry worldwide is determination of accurate reservoir model. These models make an increased percentage of the world’s hydrocarbon reserves. The model requires complete information of subsurface properties such as porosity, permeability, etc. But the fundamental challenges for geologists and geophysicists to predict these properties are reservoir specificity and heterogeneity which affects reservoir performance and their well productivity. Moreover, nonlinear multivariable regression technique like Probabilistic Neural Network has been utilizes to correlate statistically the seismic attribute to achieve high correlation coefficients when cross-plotted with reservoir properties. It results in better (r2 = 0.82) correlation coefficient than linear regression model showed (r2= 0.74). The Issue is better seismic-well tie to generate synthetic seismic traces and their correlation between predicted and the true seismic trace. Therefore, we can propose to generate pseudo porosity log from the 3-D seismic volume using polynomial neural network, helps in better integration between seismic attribute and well logs to improve the reservoir characterization by providing petrophysical properties away from well controls. The proposed model tries to achieve high attribute correlation which improves the reservoir characterization lead in estimating hydrocarbon reserves. This model also assists oil and gas companies to obtain higher drilling success.