Dr. Swati Bhisikar

@jspmrscoe.edu.in

Associate Professor
Jayawant Shikshan Prasarak Mandal's Rajarshi Shahu College of Engineering, Pune

Dr. Swati Bhisikar


            Download Compact PDF  
https://researchid.co/rscoe

EDUCATION

Engineering)

RESEARCH INTERESTS

Image and signal Processing, Artificial Intelligence, Machine Learning
12

Scopus Publications

Scopus Publications

  • Detection of Diabetic Retinopathy Using CNN Machine Learning
    Aboli Deshpande, Swati Bhisikar
    Lecture Notes in Networks and Systems, 2026
  • Advancing Energy Efficiency in 6G Networks Through Empirical Analysis of Intelligent Configuration Models
    S. P. Vibhute, S. C. Patil, S. A. Bhisikar
    Lecture Notes in Networks and Systems, 2026
  • Integrated System for Leaf Disease Prediction and Crop Recommendation Using Machine Learning and IoT
    Shradha Jadhav, Swati Bhisikar, Pranali Jadhav, Pranjali Jadhav
    Lecture Notes in Networks and Systems, 2026
  • IoT in Healthcare: Enhancing Remote Patient Monitoring
    Rajendra Bhise, Swati Bhisikar
    Icoicc 2025 3rd International Conference on Intelligent and Cloud Computing, 2025
    As technology progresses and sensors diminish in size, there is a desire to incorporate them into a wide range of sectors to improve human life. Technology integration in healthcare is an important research area. Healthcare is not always affordable, especially in poorer countries. This program aims to address a contemporary healthcare issue in society. In recent years, IoT has gained popularity in a variety of industrial industries, most notably automation and control. This project utilizes IoT technology to enhance remote patient monitoring by giving real-time data on important health signs namely blood oxygen levels (SpO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>), temperature, and heart rate. The system's primary controller is an Arduino, which is connected to an ESP8266 module for wireless communication and a SPO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> sensor for health data collecting. The data is displayed locally on an LCD screen, and any abnormal results are alerted by a buzzer. The MQTT protocol facilitates seamless data transmission to an IoT-based panel application, enabling healthcare professionals and caregivers to remotely track patient health. Design and deploy an Internet of Things (IoT) system that continuously records vital signs and delivers real-time alerts to medical experts, enhancing patient care—particularly for individuals in remote locations or home-based settings.
  • Alzheimer's Disease Diagnosis from Brain MRI Using Deep Learning Techniques
    Shubhangi Vivekanand Patil, Shailaja C. Patil, Swati A. Bhisikar
    2025 International Conference on Emerging Smart Computing and Informatics Esci 2025, 2025
    One of the major causes behind memory loss is a neurological disorder called Alzheimer's Disease. As of right now, there is no cure for this disease that can reverse or stop the disease progression. The development of an efficient model with high precision is vital for timely detection of the disease. Patients with Alzheimer's Disease are identified using different noninvasive or minimal invasive biomarkers such as Brain imaging, Bio fluid analysis and emerging biomarkers. MR imaging is to be considered as one of the noninvasive biomarker to look for evidence of memory impairment. Sometimes these indicators could be inexact when used as ground truth marking for Alzheimer’s Disease detection prior patient suffers from irreversible memory loss, which is why researchers have proposed multiple computer-aided methods. In this paper, an architecture of basic deep learning approached has been proposed. First approach uses the Convolution Neural Network design, while the second approach uses the VGG16 model, that utilized on the ADNI dataset with 3 progression stages such as Cognitive Normal, Mild Cognitive Impairment & severe stage of Alzheimer's Disease. These methods achieve impressive accuracy rates of 97% and 98% in effectively identifying and categorizing the progression levels of alzheimer’s disease. This technique can significantly contribute to improve alzheimer’s diagnosis and aid physicians in enhancing diagnostic accuracy by leveraging advanced computational techniques.
  • Enhanced Streetlight Management: Using IoT and ESP32 for Automated Fault Detection
    Swati Bhisikar, Prabodh Jagtap, Shilpa Sonawane, Hrushikesh Sawant, Prerana Pisal
    Indiscon 2024 5th IEEE India Council International Subsections Conference Science Technology and Society, 2024
    The core focus of this paper is to implement a cutting-edge system for intelligent streetlight management, leveraging an automotive monitoring system to identify and rectify any faults within the streetlight infrastructure. Rapid response of the proposed fault detection system will help save the time of the linemen, which is wasted in manually detecting the faults at streetlights, and will help to reduce the workload of the linemen. Also, most of the time, the detection of the problem mainly depends on the grievances of the local people. Thus, this causes delays in the maintenance of the streetlights, wastage of power, and problems for the local people. So, to reduce such problems, we have developed an automatic detection system for streetlight faults. During nighttime, this system will detect faults such as whether the lamp is ON or OFF, or if a wire cut is in the circuit. As the fault is detected, the information will be provided to the authorized person; the information will contain the type of fault that occurred in the streetlight system.
  • Peer-to-Peer File Sharing WebApp Enhancing Data Security and Privacy through Peer-to-Peer File Transfer in a Web Application
    Swati Bhisikar, Simran Taneja, Omkar Yadav, Swarnika Srivastava
    International Journal on Recent and Innovation Trends in Computing and Communication, 2023
    Peer-to-peer (P2P) networking has emerged as a promising technology that enables distributed systems to operate in a decentralized manner. P2P networks are based on a model where each node in the network can act as both a client and a server, thereby enabling data and resource sharing without relying on centralized servers. The P2P model has gained considerable attention in recent years due to its potential to provide a scalable, fault-tolerant, and resilient architecture for various applications such as file sharing, content distribution, and social networks.In recent years, researchers have also proposed hybrid architectures that combine the benefits of both structured and unstructured P2P networks. For example, the Distributed Hash Table (DHT) is a popular hybrid architecture that provides efficient lookup and search algorithms while maintaining the flexibility and adaptability of the unstructured network.To demonstrate the feasibility of P2P systems, several prototypes have been developed, such as the BitTorrent file-sharing protocol and the Skype voice-over-IP (VoIP) service. These prototypes have demonstrated the potential of P2P systems for large-scale applications and have paved the way for the development of new P2P-based systems.
  • Machine learning approach for prediction of lung cancer
    Hemant Kasturiwale, Swati Bhisikar, Sandhya Save
    Medical Imaging and Health Informatics, 2022
    In the current era of the introduction of artificial intelligence, there have been advances in the use of this field in image enhancement. The use of the histogram [local energy shape histogram (LESH)] approach based on local energy has previously helped diagnose breast cancer. The current support vector machine (SVM) algorithm is further advanced to AdaBoost algorithm for image extraction. The boosting algorithm of AdaBoost on the accuracy of the results will provide a much better result. For lung cancer diagnosis utilizing CT images, the LESH feature extraction algorithm is presented for lung cancer diagnoses using CT images [1]. This research builds on previous work by using the LESH with AdaBoost feature extraction methodology to detect lung cancer. The main objective of this research is to compare the LESH and HTF feature extraction approaches of SVM and AdaBoost. It is difficult to detect the specific symptoms of lung cancer since most cancer tissues are formed, and enormous tissue structures are crossed. Images will be evaluated using the LESA algorithms basic operation in this method. In this study, the GLCM technique is used to prepare snap photos and to evaluate the level of a patients condition at an early stage so that it may be established regularly or extraordinarily. The cancer stage is determined by the results. The survival rate of cancer patients can be determined using the dataset and results. The outcome is totally determined by the correct or erroneous arrangement of tissue patterning. Hence, a method must be such that it will remove the noise, extract vital information, and, at the same time, make it easy for a person to understand what is the problem with the given lung signal. In addition, the algorithm must have the ability to track the important changes and our approach should provide accurate, non-invasive assessment in clinical practice. After analyzing the signal, the method used provides vital information of linear methods, i.e., time domain and frequency domain parameters, and also provides the details of the indices of a cardiac patient and normal person which will be helpful to initiate treatment for a cardiac patient as soon as possible.
  • Classification of Rheumatoid Arthritis Based on Image Processing Technique
    S. A. Bhisikar, S. N. Kale
    Communications in Computer and Information Science, 2019
    Arthritis is a disabling and agonizing disease. The rapid growth of biomedical image processing techniques assists the doctor in diagnosis and treatment of the disease. In Rheumatoid Arthritis as the disease progresses, it results in reducing physical activity level of the patient. The method presented in this paper is a completely automated framework to detect and quantify joint space width. This system detects severe stage of RA that are contaminated by disease to the degree that the joint space is no longer noticeable in the X-ray image. In proposed work RA is classified in three stages Normal (Non-RA), Abnormal (RA) and Severe stage RA. Joint location accuracy achieved is 92%. 60 images were tested, Out of 60 Test images 20 images are Normal, 22 images are abnormal i.e. RA affected and 18 images are severe. SVM classifier with Radial basis function kernel is efficient compared to FFNN and k-NN as Non-RA i.e. Normal patient classification accuracy is 95%, RA classification accuracy is 70%, Severe stage classification accuracy of RA is 100%.
  • Automatic analysis of rheumatoid Arthritis based on statistical features
    Swati A. Bhisikar, Sujata N. Kale
    International Conference on Automatic Control and Dynamic Optimization Techniques Icacdot 2016, 2017
    Rheumatoid arthritis destroys joints of the body like erosion in bones which intern may cause deformity and ankylosis in the later stage of the disease. At the beginning of this disease mainly the joints of hand and wrist are affected making hand radiograph analysis very important. Lately manual JSW measurement in hand X-ray digital radiograph of Arthritis patients were in use but it has disadvantages like inaccuracy, inter-reader variability. Also hand radiograph analysis is difficult for radiologist since in all there are 14 number of hand joints. To avoid observer dependency, computer-aided analysis is required. We have proposed the use of image processing techniques using MATLAB to analyze joint space narrowing. In this paper bone boundaries are delineated with Active Shape Model which contains statistical model of bone shape and local texture. Joint positions are identified by local linear mapping based on texture features. We have examined five hand radiograph images affected by RA. Joint location estimate accuracy is 92%. The automated analysis helps to reduce need of skilled personnel. Also remote analysis and medication is possible.
  • EEG-based brain controlled prosthetic arm
    Dany Bright, Amrita Nair, Devashish Salvekar, Swati Bhisikar
    Conference on Advances in Signal Processing Casp 2016, 2016
  • Automatic joint detection and measurement of joint space width in arthritis
    Swati A. Bhisikar, Sujata N. Kale
    2016 IEEE International Conference on Advances in Electronics Communication and Computer Technology Icaecct 2016, 2016