MIT-World Peace University (MIT-WPU) is a top educational institute located in Pune, India. It is recognized within&nbsp;UGC under Government of Maharashtra Act No. XXXV 2017.&nbsp;

It is part of the MIT group of institutions (established in 1983) which covers 10+ campuses across India and has more than 50,000+ students enrolled across various disciplines at a given point of time. The institute houses 15 schools ranging from Engineering &amp; Technology, Design, Pharmacy, Commerce, Governance among others. It offers undergraduate, postgraduate and PhD programs among different disciplines.&nbsp;

&nbsp;

MIT World Peace University

School of Computer Engineering & Technology

The Internet of Things (IoT) has emerged as one of the most vital technologies of the twenty-first century in recent years. IoT can be described as the common physical objects that are digitally connected. Things in IoT refers to all the physical objects like sensors that can be used to be embedded with software. A drone is a type of unmanned aerial vehicle. It does not have a human pilot on board, as the name implies. Because of its compact size and lightweight mechanism, it has a lot of potential to carry out tasks that will make our lives easier. Drones may be controlled via different mediums like wifi transmissions and radio frequency signals. In the case of radio frequency drones, the drones are typically controlled via a remote controller, although wifi equipped drones may also be controlled by devices with wifi access, such as smart phones. Due to the qualities of the drone, they have become popular for use in both legal and unlawful activities, such as delivering illegal materials across country borders. Anti-drone systems are thus being developed in anticipation of their potential use in future smart cities. However, the cost of each anti-drone technology is currently too exorbitant to be employed in civilian settings. As a result, we've offered a notion for creating an anti-drone system utilizing IoT technologies in this article. Making an anti-drone system with Raspberry Pi and Wifi-Pineapple is the subject of this study. Wireless assaults can knock down or even take control of wifi-enabled drones, forcing them to make a safe or emergency landing. In this paper we have tried to bring down the drone in the vicinity of the system by means of a wireless attack using the above components. The wireless attack consists of deauthentication attack and wireless password cracking. In this study, the findings and the attack success rate have been described. The DJI Tello, a WIFI-enabled drone, was the one used in this experiment, and the system was automated so that it could decide for itself how to carry out the attacks. © 2022 IEEE.

World University root node

School of Management PG

Management

Engineering and Technology

2022 6th International Conference on Computing, Communication, Control and Automation, ICCUBEA 2022

System to capture WiFi based Drones using IoT

Over the years, many researchers have sought to use Deep Learning techniques to detect the malaria-infected cells in blood sample images. Although extremely dangerous, the spread of malaria can be restricted when treated in the early stages. This serves as an impetus for implementing an accurate solution for the detection of malaria that can replace the traditional manual process. The manual process consists of visually examining the blood samples and counting the parasitized and non-parasitized red blood cells. This process is extremely time-consuming, requires the presence of trained medical personnel, and is susceptible to human errors. With these aspects in mind, we aimed to develop a solution that could be used by medical staff with minimal training, thereby saving on time and labour. Having studied various research papers related to the use of Deep Learning techniques for the detection of malaria, we have proposed a model that addresses the gaps we identified in these systems while not compromising on the accuracy of the results. Our proposed model comprises a pre-processing module, the frozen Encoder of an Autoencoder model, a few dense CNN layers, and the classifier (Softmax). © 2022 IEEE.

2022 2nd Asian Conference on Innovation in Technology, ASIANCON 2022

Malaria Cell Image Classification using Autoencoder

Proceedings - 2022 IEEE World Conference on Applied Intelligence and Computing, AIC 2022

An IoT based Smart Medicine Dispenser Model for Healthcare

ATM(Automated Teller Machine) is among the most important tools for financial transactions. Nevertheless, for years, the visually impaired have been denied various banking facilities, especially ATM cards. The most common arguments presented by banks are 'someone might misuse your card or cheat you, you might be stalked by miscreants and robbed'. Therefore, an ATM interface should be made which allows the visually impaired to do transactions safely, independently, and efficiently. This work aims to implement an ATM interface that has all these qualities. In this interface, the user will hear audio instructions from the interface about various facilities in the ATM, implemented using text-to-speech conversion. Moreover, the ATM will also notify the breakup of dispensed cash and directions for the cash dispenser via audio instructions. The user will also be able to give voice instructions to the interface for speeding up the transactions, implemented using NLP(Natural Language Processing). For avoiding the snooping of ATM PIN(Personal Identification Number), a distance detection algorithm will be used, to ensure that there is no one else present in the ATM. © Grenze Scientific Society, 2022.

13th International Conference on Advances in Computing, Control, and Telecommunication Technologies, ACT 2022

Designing a Secure Interactive ATM for Visually Impaired People

BACKGROUND: The nutrition screening tool determines whether or not a patient is at risk of malnutrition. Many organizations contributed to the form's design by researching various nutritional screening tools. OBJECTIVE: The purpose of this study is to analyse the different screening forms for the hospitalised geriatric patients to understand the nutritional status as well as the comorbidities. In addition, the study intends to identify the essential parameters from three nutrition screening forms-MUST,NRS-2002,MNA. The primary outcome is to identify the essential parameters from the screening form and calculating the nutrition risk score is considered a secondary outcome. METHODS: The research was done on geriatric patients who were admitted to the hospital from November 2019 to March 2020, suffering from comorbidities. The study comprises 192 hospitalized patients with an average age of 74 years including both genders. The patients' data was collected by a certified dietician to understand the essential parameters of these nutrition screening forms. The Pearson correlation test is used to find the correlation value between the screening form parameters. Based on the analysis, features with a correlation value greater than 0.5 with the nutrition score are considered a significant parameter for the research. RESULTS: The analysis of the nutrition screening strategy demonstrates the different significant parameters of the above three screening forms. The experimentation shows, the significant parameters are the Body Mass Index, impaired nutrition status, the severity of disease, mobility, mode of feeding, nutritional status, and weight loss score. CONCLUSIONS: The research is performed to understand the correlation between the screening form parameters MNA, MUST, NRS2002 using a non-invasive method and analyze the score in patients to improve health rates by using nutritional risk therapy and nutritional care for such patients. Nutrition therapy can be applied to patients suffering from malnutrition and at risk of malnutrition to improve the health of the patient. © 2022-IOS Press. All rights reserved.

Mediterranean Journal of Nutrition and Metabolism

Analysis of Nutritional Risk Tools to Identify Essential Parameters in Assessing Nutritional Risk in Elderly Patients

The continuous battle against the variants of Corona Virus demands speedy treatment and quick diagnostic reporting on priority basis. With millions of people contracting the infection every day and a mortality rate of 2%, our goal is to solve this growing problem by developing an important and substantive method for diagnosing COVID19 patients. Due to a proportionally reduced number of medical practitioners, testing kits, and other resources in densely populated nations, the exponential development of COVID19 cases is having a significant impact on the health care system, making it increasingly important to identify infected patients. The goal of this work is to develop an exact, productive and time-saving algorithm to identify positive corona patients that addresses the aforementioned issues. In this paper, a Deep Convolution Neural Network model called "EfficientNet"is implemented and explored that can reveal significant diagnostic characteristics to enable radiologists and medical specialists locate COVID-19 infected patients using X-ray pictures of the chest and aid in the fight against the pandemic. The experimental findings conclusively indicate that an accuracy rate of 99.71 percent was obtained for binary classification of Non-COVID and COVID Chest X-ray pictures. Our pretrained Deep Learning classification model can be a significant contribution to recognizing COVID-19 inflicted individuals due to its high diagnostic accuracy. © 2022 IEEE.

2022 IEEE 7th International conference for Convergence in Technology, I2CT 2022

Deep Pre-Trained Convolutional Neural System for High-Accuracy Covid19 Forecasting from Chest X-Rays

Segmentation of retinal structures, namely optic disc, vessel, demarcation line, and ridge, is essential for describing the characteristics of Retinopathy of Prematurity (ROP). Computerized systems are being developed for automatic segmentation in fundus images to assist the medical experts and bring consistency in the diagnosis. There are multiple challenges in the segmentation task of premature infants’ fundus images. The annotation and ground truth preparation required for the segmentation is complex, challenging, and expensive. Further, ROP datasets are not available publicly, and hence carrying out such a task needs a primary dataset and significant assistance from the domain expert. To address this gap, two primary datasets named HVDROPDB-BV and HVDROPDB-RIDGE were developed. The datasets consist of images captured by two different imaging systems, having different sizes, resolutions, and illumination. This made the trained models generic and robust to data variability and heterogeneity. We propose the modified U-Net architectures by incorporating squeeze and excitation (SE) blocks and attention gates (AG) to segment the demarcation line/ridge and vessel from these datasets. These modifications were tested and validated by ROP experts. The performance of all the three networks (U-Net, AG U-Net, and SE U-Net) was promising, with a variation of 1 to 6% in the dice coefficient for the HVDROPDB datasets. The area under the curve (AUC) obtained for all three networks was above 0.94, indicating them as excellent models. AG U-Net outperformed the other two networks, with a sensitivity of 96% and specificity of 89% for stage detection via new test images. © 2022, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.

Multimedia Tools and Applications

Deep dive in retinal fundus image segmentation using deep learning for retinopathy of prematurity

Chatbots, also called chatterbots are applications that simulate human conversation either taking voice or text commands or both as input. They are automated programs that utilize Natural language processing techniques to understand process and responds accordingly to the users. Chatbots are cost effective and require a very little investment cost in the beginning stage of development. The software program can be built using AIML that use Extensible Mark-up Language as a base. In python, libraries like NLTK and JSON package can be used for the efficient designing of a Chatbot model. Chatbots use Corpus which is the vocabulary or the huge amount of data from different human dataset to the model. This paper elaborates on a Chatbot model that uses the functions of nltk library in python to pre-process data and responds accurately to the user using the cosine similarity approach. This chatbot can be implemented for any specific domain by changing the corpus. The corpus used in this design is about chatbots itself. The chatbot is capable of accurately responding to the queries related to chatbots. Accuracy of the proposed chatbot is 93% that is out of 150 queries it answered 140 queries correctly.

2022 4th International Conference on Smart Systems and Inventive Technology (ICSSIT)

Design and Implementation of Multipurpose Chatbot

2022 IEEE Students Conference on Engineering and Systems, SCES 2022

Hand Gesture Recognition using YOLO Models for Hearing and Speech Impaired People

Local feature-based approaches are proven to be most successful for the application of human action recognition. This work aims to study the effect of the number of cuboids chosen on the performance of human action recognition. A 3D region, called cuboid, is extracted around every spatio-temporal interest point and an entropy-based cuboid selection method is implemented for choosing the cuboids having maximum information. The proposed method is evaluated using UT interaction and SBU kinect interaction datasets. Results show that maximum classification accuracy and F1 Score are obtained when the top 40 cuboids are selected for the feature extraction. The classification accuracy and F1 Score increase when the number of selected cuboids is increased from 10 to 40 and remain constant when the number is increased beyond 40. The selected 40 cuboids are then used for feature extraction. To carry out this experiment, wavelet coefficients extracted from selected cuboids are used as features. Copyright © 2022 Inderscience Enterprises Ltd.