Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
37 results
Search Results
Item Convolutional neural networks for medical image analysis(CRC Press, 2022) Gogineni, R.; Chaturvedi, A.The enormous advancement of Deep Learning (DL) algorithms in various fields has been adopted by many perspectives of medical image analysis. In the context of deep learning, convolutional neural networks (CNNs) are widely used models for medical imaging tasks and lead to a huge progress in computer aided diagnosis. CNNs are being profoundly absorbed in the research of medical imaging because of their pertinent features in preserving local image relations and dimensionality reduction. CNNs demonstrate excellent performance in lesion detection and other vision tasks. This chapter presents the modules of convolutional neural networks and their prominent architectures employed for medical image analysis. The commonly used variants of CNNs in medical image analysis including ResNet, GoogleNet, and fully convolutional neural networks are analyzed. The principal research areas and applications of medical image registration, segmentation, detection and classification are discussed. The overview of the above mentioned tasks in a few of the usual diagnosis areas of the human body such as brain, lungs and eye is presented in this chapter. The crucial challenge in medical image analysis using CNNs is the availability of training data sets. This problem is addressed in this chapter, by providing the existing medical image datasets for different application areas like chest, diabetic retinopathy, lungs and heart. Finally, the chapter is concluded by discussing the challenges associated with CNN architectures and promising future research directions © 2023 Şaban öztürk. All rights reserved.Item Design and analysis of microstrip elliptical low pass filter(IEEE Computer Society, 2009) Kumar, P.; Chaturvedi, A.Design and analysis of a stepped impedance elliptical microstrip low pass filter has been described in this paper. To improve the frequency response, fractals have been proposed in the conventional design. Conventional geometry and fractalized geometry have been simulated and thus obtained results are reported and compared.Item Truncated tip triangular microstrip patch antenna(2010) Chaturvedi, A.; Bhomia, Y.; Yadav, D.This paper presents a design of triangular microstrip antenna with truncated tip and experimentally studied on Ansoft Designer v-2.2.0 software. This design technology is achieved by cutting all three tips of the triangular microstrip antenna and placing a single coaxial feed. Triangular patch antenna is designed on a FR4 substrate of thickness 1.6 mm and relative permittivity of 4.4 and mounted above the ground plane at a height of 6 mm. Bandwidth as high as 11.07% are achieved with stable pattern characteristics, such as gain and cross polarization, within its bandwidth. Impedance bandwidth, antenna gain and return loss are observed for the proposed antenna. Details of the measured and simulated results are presented and discussed. ©2010 IEEE.Item Edge detection of femur bone - A comparative study(2010) Harish Kumar, H.; Chaturvedi, A.Edge detection of femur in X - ray images is an important pre processing step in segmentation and 3 - D reconstruction of femur. A typical femur image is generally very noisy. A lot of edges caused by the muscles and other bones can easily mislead the edge detection algorithm. Particularly, the femoral head overlapping the pelvic bone makes it very difficult to get a clear edge of the femur head. The edge caused by the abdominal muscles around the femur shaft can also mislead the edge detection algorithm. These extraneous edges and noise make edge detection very difficult and challenging, which is not well solved. Classical edge detectors fail miserably due to the high inhomogeneous nature of the femur X - ray images. This paper compares a new approach to edge detection of femur X - ray images using Wavelet transforms with classical edge detectors. The Wavelet based edge detection algorithm combines the coefficients of wavelet transforms on a series of scales and significantly improves the result. It is found that Wavelet based technique works much better than classical edge detectors. ©2010 IEEE.Item Estimation of degradation of performance of optical network due to crosstalk(2011) Gupta, M.L.; Chaturvedi, A.; Bhomia, Y.The aim of this paper is to carry out an analysis of optical network by calculating bit error rate (BER) as a function of signal to noise ratio by using Hermitian polynomial for different homodyne crosstalk levels. The results have been plotted for different crosstalk level for BER as a function of single to noise ratio (SNR) by using exact analysis and Gaussian approximation and the results are compared. Simulations have been carried out in MAT Lab Version 7.4. We present an exact analytical probability density function and a closed form bit error rate formula for optical network which performance is degraded by homodyne crosstalk. In homodyne (intrachannel) crosstalk, the interfering signal is at same wavelength as the desired signal. This effect is more severe than interchannel crosstalk (in which interfering signal operates at different wavelength), since the interference falls completely within the receiver bandwidth. © 2011 IEEE.Item An improved mechanism of leaf node identification for radial distribution networks(2011) Sharma, D.P.; Chaturvedi, A.; Purohit, G.; Prasad, G.Optimal operational & control aspects of distribution networks have been a thrust research area in academics as well as in industries since last two-three decades. In day to day practice, every one of us uses services offered by public utility distribution networks namely, water distribution network, electrical power distribution network etc. Operational topology of power distribution networks are radial in nature and hence are termed as radial distribution networks (RDNs). Network reconfiguration has been exercised as one of the prime and widely adopted approach for operational, maintenance and control activities of an RDN. Since last two decades, researchers have been using evolutionary computation based techniques (Genetic Algorithm [1], Simulating Annealing etc) for optimal network reconfiguration. The choices of network topology for the specific purpose/application requires a careful analysis of its merits and demerits. In RDNs, the ultimate performance of a specific network topology is usually assessed by an iterative algorithm known as Load flow analysis (LFA) and its execution results in estimation of voltages, currents and losses profiles which in turn decides, whether the obtained network topology is good or bad. Thus, an obvious need is in favor of developing a conceptual frame work for faster load flow algorithm especially to meet near real-time operation requirements. In this paper, a simple and computationally efficient method for terminal (leaf) nodes identification is presented and thus, by integrating this subroutine in LFA definitely leads to an efficient and faster LFA algorithm. © 2011 IEEE.Item Geographical location based hierarchical routing strategy for wireless sensor networks(2012) Kumar, P.; Chaturvedi, A.; Kulkarni, M.The overall performance and thus the achievable efficiency of wireless sensor networks (WSNs) rely on the type of protocols deployed to support particular application in hand. In WSN, network's lifetime depends upon the residual energy of individual nodes. In this paper; a brief survey comprises of various reported methodologies that make uses of geographical location attributes for routing in WSN is presented. To assist the routing algorithm on location aspect of the source and destination as well as other intermediate nodes; a location index is formulated based on binary encoded spatial frames. The merit of proposed spatial encoding scheme is its supportiveness on scalability aspects, so it can be easily tailored to accommodate hierarchy based network architectures. Further, the scheme incorporates an effective mechanism to select cluster heads for each cluster in such a way so as to avoid the localization of hot spot effects right from cluster level to network level. © 2012 IEEE.Item Microstrip patch antenna combining crown and Sierpinski fractal-shapes(2012) Bhomia, Y.; Chaturvedi, A.; Sharma, Y.K.This paper presents a design of Microstrip Patch Antenna Combining Crown and Sierpinski Fractal-Shapes and experimentally studied on IE3D software [10]. This design is achieved by cutting multi shapes in square pattern combined with crown & sierpinski fractal shapes and placing a single coaxial feed. The Sierpinski pre-fractal can be defined by an iterated function system (IFS). As a consequence, the geometry has a multilevel structure with many equal sub domains. It is experimentally found that the resonant frequency of the patch can be greatly lowered, and the higher the iteration order of the fractal shapes, the lower the resonant frequency becomes. And this property can be utilized to reduce the size of the microstrip patch antennas. It is also found that, the radiation patterns of the proposed fractal-shaped antennas maintained because of the self-similarity and Centro- symmetry of the fractal shapes. Crown & sierpinski fractal shapes patch antenna is designed on a FR4 substrate of thickness 1.524 mm and relative permittivity of 4.4 and mounted above the ground plane at a height of 6 mm. Bandwidth as high as 31.14% are achieved with stable pattern characteristics, such as gain and cross polarization, within its bandwidth. Impedance bandwidth, antenna gain and return loss are observed for the proposed antenna. Details of the measured and simulated results are presented and discussed. © 2012 ACM.Item An energy efficient algorithm to avoid hot spot effects in Wireless Sensor Networks(IEEE Computer Society help@computer.org, 2013) Kumar, P.; Chaturvedi, A.In this paper, a novel approach is proposed to minimize the hot spot effect to improve the life time of Wireless Sensor Networks (WSNs) based on energy of each sensor node. To implement the proposed approach, the spatial locations of geographical area under surveillance are motioned using binary location index. The simulation work is carried out for two different case studies; in first case the sink/base station is remains stationary during entire observation, whereas in other case the sink is reallocated to appropriate locations at suitable time instants. Timely varying pattern of residual energy of all network nodes and total number of queries supported by entire network till it attains targeted life time is presented and discussed. © 2013 IEEE.Item Life time enhancement of wireless Sensor Network using fuzzy c-means clustering algorithm(Institute of Electrical and Electronics Engineers Inc., 2014) Kumar, P.; Chaturvedi, A.The major issues in wireless Sensor Networks (WSNs) are efficient uses of limited resources and appropriate routing of network paths under severely constrained energy scenarios. To overcome these issues; k-means and fuzzy c-means algorithms are investigated to form clusters and for subsequent selection of cluster heads. For all these clusters; selection of cluster head is done based on member sensor nodes residual energy status (RES) and estimation of Euclidean distances. Depending upon the Euclidean distance measure between the sink node and center of gravity of clusters; these clusters are classified into five types. Further, RES estimations are presented for cluster heads as well simple sensor network nodes. © 2014 IEEE.