Gangadhar Ch, Nama Ajay Nagendra, Syed Mutahar Aaqib, C.M. Sulaikha, Shaheena Kv and Karanam Santoshachandra Rao
COVID-19 would have a far-reaching impact on the international health-care industry and the patients. For COVID-19, there is a need for unique screening tests to reliably and…
Abstract
Purpose
COVID-19 would have a far-reaching impact on the international health-care industry and the patients. For COVID-19, there is a need for unique screening tests to reliably and rapidly determine who is infected. Medical COVID images protection is critical when data pertaining to computer images are being transmitted through public networks in health information systems.
Design/methodology/approach
Medical images such as computed tomography (CT) play key role in the diagnosis of COVID-19 patients. Neural networks-based methods are designed to detect COVID patients using chest CT scan images. And CT images are transmitted securely in health information systems.
Findings
The authors hereby examine neural networks-based COVID diagnosis methods using chest CT scan images and secure transmission of CT images for health information systems. For screening patients infected with COVID-19, a new approach using convolutional neural networks is proposed, and its output is simulated.
Originality/value
The required patient’s chest CT scan images have been taken from online databases such as GitHub. The experiments show that neural networks-based methods are effective in the diagnosis of COVID-19 patients using chest CT scan images.
Details
Keywords
Nama Ajay Nagendra and Lakshman Pappula
The issues of radiating sources in the existence of smooth convex matters by such objects are of huge significance in the modeling of antennas on structures. Conformal antenna…
Abstract
Purpose
The issues of radiating sources in the existence of smooth convex matters by such objects are of huge significance in the modeling of antennas on structures. Conformal antenna arrays are necessary when an antenna has to match to certain platforms. A fundamental problem in the design is that the possible surfaces for a conformal antenna are infinite in number. Furthermore, if there is no symmetry, each element will see a different environment, and this complicates the mathematics. As a consequence, the element factor cannot be factored out from the array factor.
Design/methodology/approach
This paper intends to enhance the design of the conformal antenna. Here, the main objective of this task is to maximize the antenna gain and directivity from the first-side lobe and other side-lobes in the two way radiation pattern. Thus the adopted model is designed as a multiobjective concern. In order to attain this multiobjective function, both the element spacing and the radius of each antenna element should be optimized based on the probability of the Crow Search Algorithm (CSA). Thus the proposed method is named Probability Improved CSA (PI-CSA). Here, the First Null Beam Width (FNBW) and Side-Lobe Level (SLL) are minimized. Moreover, the adopted scheme is compared with conventional algorithms, and the results are attained.
Findings
From the analysis, the gain of the presented PI-CSA scheme in terms of best performance was 52.68% superior to ABC, 25.11% superior to PSO, 13.38% superior to FF and 3.21% superior to CS algorithms. Moreover, the mean performance of the adopted model was 62.94% better than ABC, 13.06% better than PSO, 24.34% better than FF and 10.05% better than CS algorithms. By maximizing the gain and directivity, FNBW and SLL were decreased. Thus, the optimal design of the conformal antenna has been attained by the proposed PI-CSA algorithm in an effective way.
Originality/value
This paper presents a technique for enhancing the design of the conformal antenna using the PI-CSA algorithm. This is the first work that utilizes PI-CSA-based optimization for improving the design of the conformal antenna.