Showing papers by "Budge Budge Institute of Technology published in 2021"

COVID-19 detection from lung CT-scan images using transfer learning approach

Abstract: Since the onset of 2020, the spread of coronavirus disease (COVID-19) has rapidly accelerated worldwide into a state of severe pandemic. COVID-19 has infected more than 29 million people and caused more than 900 thousand deaths at the time of writing. Since it is highly contagious, it causes explosive community transmission. Thus, health care delivery has been disrupted and compromised by the lack of testing kits. COVID-19-infected patients show severe acute respiratory syndrome. Meanwhile, the scientific community has been involved in the implementation of deep learning (DL) techniques to diagnose COVID-19 using computed tomography (CT) lung scans, since CT is a pertinent screening tool due to its higher sensitivity in recognizing early pneumonic changes. However, large datasets of CT-scan images are not publicly available due to privacy concerns and obtaining very accurate models has become difficult. Thus, to overcome this drawback, transfer-learning pre-trained models are used in the proposed methodology to classify COVID-19 (positive) and COVID-19 (negative) patients. We describe the development of a DL framework that includes pre-trained models (DenseNet201, VGG16, ResNet50V2, and MobileNet) as its backbone, known as KarNet. To extensively test and analyze the framework, each model was trained on original (i.e. unaugmented) and manipulated (i.e. augmented) datasets. Among the four pre-trained models of KarNet, the one that used DenseNet201 demonstrated excellent diagnostic ability, with AUC scores of 1.00 and 0.99 for models trained on unaugmented and augmented data sets, respectively. Even after considerable distortion of the images (i.e. the augmented dataset) DenseNet201 achieved an accuracy of 97% for the test dataset, followed by ResNet50V2, MobileNet, and VGG16 (which achieved accuracies of 96%, 95%, and 94%, respectively).

Comparison Study of DG-MOSFET with and without Gate Stack Configuration for Biosensor Applications

Abstract: In this Paper, we have studied and compared the performance of two different configurations of simulation model advanced MOSFET devices which can be used for biosensor application. The bio-molecules like protein, biotin, streptavidin, APTES, etc., undergo label free electrical detection with the help of dielectrical modulation technique in order to overcome the limitations of short channel effect in a more efficient way. The bio-molecules trapped inside the cavity region change the electrical parameters of the MOSFET. Biosensors based on MOSFETs have certain issues, like short channel effects (SCEs) and problems related to scaling and power supply. Therefore the proposed device is better withstand to SCEs and can be consider as an alternative for biosensing applications. For channel material, silicon is used for both the configurations i.e. with stack and without stack model and we have also studied the performance of the device based on the analog as well as RF parameters by considering the protein as bio-molecule in the cavity. Two different oxide materials are used to design the device structure such as HfO2 (K = 25) and SiO2 (K = 3.9) and for simulation purpose the 2D Sentrausu TCAD simulator has been used. The sensing capability of this proposed dielectric modulated device can be applicable for IOT based applications. They can also be uses in health IOT systems for medical research applications and as bio chip sensor in wearable device so as to study the protein content of the human body.

Comparative study on Analog & RF Parameter of InAlN/AlN/GaN Normally off HEMTs with and without AlGaN Back Barrier

Abstract: In this work, we have made a relative assessment of lattice-matched In 0.17 Al 0.83 N/AlN/GaN normally off HEMT device with AlGaN back-barrier (BB) and without back-barrier by using device simulator. The utility of AlGaN BB on the said E-HEMT relaxes the channel, which reduces the short channel effects. It also reduces the total gate capacitance and simultaneously improves the cut- off frequency. The numerical modelings are done by the 2Dimenssional TCAD by means of HD mobility and matched with the previously accepted experimental result. Different device parameters are analyzed and compared with BB and without BB with the help of the numerical modeling. AlGaN back-barrier has further benefits in device parameters with comparison to without back-barrier i.e. less total gate capacitance and higher cut-off frequency. These outcomes prove the utility of proposed BB in such E-Mode GaN HEMTs can be a substitute way out in support of high power along with high-frequency purposes.

Experimental and statistical model analysis on Bambusa Tulda (bio-material) for removal of dye contamination from wastewater

Abstract: The present study applies modified Bambusa Tulda (bio-material) to replace Granular activated carbon to remove Malachite Green dye contaminant from wastewater. The optimum operational condition was found to be at pH 7200 RPM (rotation per minute), at 298 K temperature with an equilibrium contact time of 60 min and dosage of 10 g/L—the study of regression analysis conducted to predict the percentage removal of malachite green dye by Na2CO3 modified Bambusa Tulda (NC-BT) bio-adsorbent. Bamboo bio-waste can reuse for energy generation. Various approaches for dye removal require physical force and limited removal effectiveness, so adsorption techniques with bio-material adsorbent are prevalent because of their high efficiency, less energy consumption, recyclability, cost-effectiveness, ease of handling, and the absence of secondary pollution. There is an increasing emphasis on identifying low-cost biomaterials that can be used to alternative Granular activated carbon. From all the results, modified Bambusa Tulda can be a suitable and efficient replacement of commercially available Granular activated carbon to remove Malachite Green contaminant from wastewater.

Hybrid Genetic Algorithm-Gravitational Search Algorithm to Optimize Multi-Scale Load Dispatch

Abstract: Genetic algorithm (GA) and gravitational search algorithm (GSA) both have successfully been applied in solving ELD problems of electrical power generation systems. Each of these algorithms has their limitations and advantage. GA's global search and GSA's local search capability are their strong points while long execution period of GA and premature of convergence of GSA hinders the possibility of optimum result when applied separately in ELD problems. To mitigate these limitations, experiment is done for the first time by combining GA and GSA suitably and applying the hybrid in non-linear ELD problems of 6, 15, and 40 unit test systems. The paper reports the details of this study including comparative analysis considering similar hybrid algorithms. The result strongly attests the quality, consistency, and overall effectiveness of the GA-GSA hybrid in ELD problems.

Role of Communication in Ethical HR Practice

Abstract: At present, we see HR Professionals are tasked with various ethical as well as professional challenges such as inequality in workplace, sexual harassments, setting standards for workplace conducts amongst few. When certain types of scandals threaten the status quo of a company, it becomes the prerogative of the HR to take tough decisions and handle the matter swiftly so that the daily work culture is not affected. Thus, an HR professional can be termed as the “moral police” of a company. But this task is no easy feat. It befalls on a them to master the art of ethical decision making and become a valuable resource by communicating them not only to the employees but for the sake of company as well. In this study, I aim to point out the major problems that an HR face while taking tough calls on moral decisions and how he/she can overcome their dilemma and overcome the communication barrier that is common in most organizational sectors. My study will mainly focus on the four Ws : What kind of situation may arise that can threaten the moral grounds of decision making, Why such threatening situations may arise , What can be ethically done to overcome such situation and Ways to communicate them to the employees of the organization.

An ANN Approach of Twisted Fiestel Block Ciphering

Abstract: The twenty-first century is a digital era when bulk volume of data is engendering per second over the cyber world which is somehow an indispensability of digital life. So, for the purport of its auspice from unauthorized betokens, sundry encryption and decryption techniques were engendered which are predicated on different aspects of bulwarking the data. One of them is the Feistel cipher algorithm which is predicated on symmetric key cryptography where a single key is utilized for both encryption and decryption of the text. In this method, there is immensely colossal drawback that the Feistel cipher only transmutes a component of the internal state in each round, i.e., after user feeds the data, only half part of the main text is manipulated in each round which as a result is bit insecure as the left half is the pristine text for the very first round. So, to surmount this issue we have proposed one algorithm predicated on Feistel cipher which transmutes the entire internal state in each round and makes the data entered by user more secure. Not only that, our method is more time-efficient than that of the traditional Feistel cipher. In this paper, we introduced an incipient convoluted structure where our main motive is to make an algorithm which is more expeditious, lighter and time preserver than the traditional Feistel cipher.

Mechanics of Damage at Steel-Concrete Interface in RC Structures

Abstract: One of the major problems affecting present infrastructure is shortening of service life due to durability issues. A large number of steel-concrete composite structures fail not from overloading but because of durability issues. One of the main causes of short service life is corrosion of steel in steel-concrete composite structures. The current state of practice in design codes deals with this issue by assuming a perfect bond between steel and concrete for design considerations and provision of minimum cover depth for prevention against corrosion. The degradation at the interface that can have long-term deteriorative effects on strength and durability cannot be addressed with the present framework. Thus a need for more exact design and analysis recommendations is needed to address and deal with the issue of durability caused by corrosion. The objective of this study is to investigate the effect of corrosion at the steel-concrete interface on the structural behaviour of steel-concrete composite structures through numerical methods. The mechanics of degradation of the interface have been studied using finite element method. Effects of corrosion at the steel-concrete interface such as weakening of the interface, reduction of structural steel area and de-lamination at the interface have been studied. These phenomenon can be modelled using finite element approaches. Modelling of corrosion using solid elements presented with problems of mesh refinement, coupling damage with a constitutive response. To address the above, a UEL has been implemented in ABAQUS using a traction separation law based on the Mohr-Coulomb failure criteria, which will be used to model corrosion at the interface. Experimental data on corrosion samples were used to obtain the material properties. The developed model shall provide a way to study the effect of the presence of varying amounts of corrosion on the performance of a structure. The present study aims to provide a robust numerical scheme to analyse composite structures with varying degrees of corrosion. On the basis of obtained results, recommendations for addressing deterioration due to corrosion would be recommended.

Assessment of Electromagnetic Energy Absorption Rate in Hibiscus Flower Model at 947.50 MHz and 1842.50 MHz

Abstract: Living biological tissues in plants, fruits and flowers possess considerably high permittivity and electrical conductivity over wide microwave frequency spectrum. Plants, being immobile, are continuously exposed to electromagnetic radiation emitted from cell tower antennas. As a consequence, plants are expected to absorb quite a large amount of incident electromagnetic energy over multiple frequency bands. Different global and national regulatory bodies have put limits on maximum permissible electromagnetic exposure to restrict absorption in humans and minimize related health risks. However, electromagnetic energy absorption in plants along with associated physiological and molecular effects has not been considered in these guidelines. Plants, in general, possess higher surface area while compared to humans or other living objects. In particular, flowers own reasonably high surface to volume ratio and consequently dissipate quite high amount of electromagnetic energy in a relatively less tissue mass. Hence, this paper aims at estimating specific absorption rate data for a hibiscus flower model at 947.50 MHz and 1842.50 MHz. Linearly polarized plane waves at those two frequencies impinge on the hibiscus flower model in separate simulations as per the International Commission on Non-Ionizing Radiation Protection guidelines. Maximum local point, averaged over 1g contiguous tissue and whole body averaged specific absorption rate data at those two frequencies are significantly different than the data reported earlier at 2450 MHz - indicating, the nature of dependence of specific absorption rate data on frequency of irradiation, incident field strength and dielectric properties of constituent flower tissue.