Pavithra Guru R

@srmist.edu.in

Assistant Professor and Department of Computing Technology
SRM Institute of Science and Technology Kattankulathur

Pavithra Guru R


                                    Download Compact PDF  
https://researchid.co/pavithragurur
I am Dr. R. Pavithra Guru, Assistant Professor in the Department of CTech, SRMIST, with research expertise in VLSI Physical Design, intelligent optimization algorithms, ML-assisted EDA, and digital system design.

My scholarly work includes multiple peer-reviewed publications, national patents, and ongoing research projects in circuit partitioning, CA-based optimization, and ML-driven semiconductor workflows. I regularly serve as a technical reviewer for conferences and journals in AI, VLSI, microelectronics, and embedded systems.

With a strong academic foundation (B.E, M.E, PhD, PDF–UK), I am committed to advancing semiconductor research while fostering high-quality learning through project-based teaching, research mentorship, and industry-oriented curriculum development.

EDUCATION

Ph.D. in Information & Communication Engineering
Anna University, Chennai
2017 – 2022
Research focus on VLSI design optimization and machine learning applications.

M.E. in VLSI Design
Anna University, Chennai
2014 – 2016
Graduated with first class honors.

B.E. in Electronics and Communication Engineering
Anna University, Chennai
2010 – 2014
Graduated with first class honors.

RESEARCH, TEACHING, or OTHER INTERESTS

Electrical and Electronic Engineering, Computer Science, Artificial Intelligence, Engineering
17

Scopus Publications

Scopus Publications

  • Optimized Auxiliary Classifier Wasserstein Generative Adversarial Network espoused Hypergraph Partitioning for VLSI Circuit Design
    R. Pavithra Guru
    International Journal of Information Technology and Decision Making, 2026
    Partitioning is a fundamental problem in computer science and plays a significant role in very large-scale integration (VLSI) circuit design. Effective partitioning significantly impacts design quality and efficiency, particularly in chip design workflows. Existing partitioning methods, such as machine-learning (ML) methods, have shown limitations in performance, necessitating the development of more effective solutions. This study aims to improve partitioning and floor planning in VLSI circuit design by proposing an optimized method that enhances partitioning accuracy, minimizes area usage and reduces delay. The Optimized Auxiliary Classifier Wasserstein Generative Adversarial Network espoused Hypergraph Partitioning for VLSI Circuit Design (HGP-ACWGAN-BEPOA) framework is proposed to generate hypergraph partitions for VLSI circuit design, ensuring improved floor planning. To enhance partitioning efficiency, the binary emperor penguin optimization algorithm (BEPOA) is applied, optimizing hypergraph partitioning for better design quality. The model is trained and evaluated using benchmark circuit datasets to assess its effectiveness. Experimental results demonstrate that HGP-ACWGAN-BEPOA achieves superior performance compared to the existing models. It reduces area usage by 20.19%, 15.61%, and 18.27%, decreases delay by 22.78%, 20.67%, and 15.98%, and lowers power consumption by 25.79%, 20.19%, and 30.12% when compared to the existing models, such as OPFP-WO-ABSO-VLSI, SSGNN-GPP-VLSI, and AP-VDUG-VLSI, respectively. The proposed HGP-ACWGAN-BEPOA framework effectively enhances hypergraph partitioning for VLSI circuit design, improving efficiency in floor planning, reducing resource consumption and outperforming existing methodologies. This approach provides a promising direction for advancing partitioning techniques in VLSI circuits.
  • ABO optimized hybrid Trans-CNN-Bi-GRU approach for intrusion detection in IoT networks: a privacy-preserving solution
    R. Pavithra Guru, Thomas M. Chen, Mithileysh Sathiyanarayanan
    Cluster Computing, 2026
  • Cellular automata-based framework for yield optimization in VLSI physical design of large-scale benchmark circuits
    R. Pavithra Guru, V. Vaithianathan
    Journal of Ambient Intelligence and Humanized Computing, 2026
  • Real-Time Audio Pattern Detection for Enhancing Situational Awareness in Headphone Users Using Deep Learning
    Dheeraj Konakalla, Shiv Akash S M, Pavithra Guru R
    7th International Conference on Innovative Trends in Information Technology Icitiit 2026, 2026
    The enjoyment of personal audio experiences via headphones is rising rapidly, while users' abilities to hear important sounds around them have decreased, resulting in a major safety concern in day-to-day living. To help people become more aware of the sounds that occur around them, including both general-purpose environmental sounds and sounds of speech made by individuals (e.g., someone nearby talking to them), we have developed a real-time audio pattern detection framework based on deep learning technology. Developed using continuous real-time monitoring and analysis of environmental sounds using Mel spectrogram feature extraction, the framework consists of a lightweight convolutional neural network (CNN) which uses an on-device CPU to classify a variety of types of environmental sounds and speech made by specific individuals. An extensive, aggregate dataset containing 3,500 records of both general-purpose environmental sounds and records of speech sounds made by specific individuals was used to test the framework's performance. The average classification metrics of the framework were found to be 82.4 In summary, the results of the proposed framework demonstrate that headphone users can improve their situational awareness while having their privacy maintained with this low-latency method of detecting real-time audio patterns.
  • Grey wolf optimization (GWO) based efficient partitioning algorithm VLSI circuits for reducing the interconnections
    R. Pavithra Guru, V. Vaithianathan
    Analog Integrated Circuits and Signal Processing, 2025
  • A secure authentication and optimization based device-to-device communication for 5G enabled IoT applications of smart cities
    Therkkumthala Prakash Rani, Ayyadurai Maruthu, Pavithra Guru Ramakrishnan, Cypto Jayakumar
    Wireless Networks, 2025
  • Medvision: A Secure Web-Based Platform for Remote Medical Consultations and Case Management
    Pavithra Guru R, Shezil Ahammed C, Tarun M
    2025 International Conference on Computing Technologies and Data Communication Icctdc 2025, 2025
    The COVID-19 pandemic highlighted significant gaps in healthcare accessibility, particularly for patients requiring specialist consultations. This paper presents MedVision, a webbased telemedicine platform that facilitates secure, asynchronous consultations between patients and doctors. The system enables patients to submit detailed medical cases, receive expert diagnoses, and communicate with healthcare professionals through a secure messaging interface. Built on React, TypeScript, and Supabase technologies, MedVision implements role-based access control, comprehensive medical record management, and secure patient-doctor communication channels. Evaluation results demonstrate the platform's effectiveness in reducing consultation wait times by 68 % while maintaining diagnostic accuracy comparable to in-person consultations. MedVision addresses critical challenges in modern healthcare delivery by improving accessibility, reducing costs, and enabling efficient remote medical consultations without compromising quality of care.
  • Design and Simulation of Bi-Layer Optimized High K- Dielectric Medium for N-Mosfet with Wild Horse Optimization to Improve Electrical Characteristics
    R. Pavithra Guru
    Ecs Journal of Solid State Science and Technology, 2024
    Electronic devices for advanced modern semiconductor based technology, mainly focus on the design regarding lighter, faster and more affordable solutions to meet the specifications of modern digital electronics. Some of the drawbacks for minimizing device size in MOSFET include gate insulator scaling, Short-Channel Effects (SCEs), shallow junction technology and off-state leakage current in MOSFET devices. In addition, the traditional SiO2 as a dielectric material contains restricted maximum capacitance as well as increased tunnel current leakage due to the thickness. Hence, a High-k dielectric is required to replace SiO2 to overcome the mentioned issues. In this model, the N-type MOSFET is designed based on the bi-layer high K-dielectric medium with optimized thickness according to the maximum capacitance and minimum threshold voltage, which are implemented on VLSI based applications such as 6 T SRAM for evaluating the performance. The drain current of HfO2, Al2O3 and HfO2+Si3N4 for 2.5 v drain voltage are 1.87 mA, 1.51 mA and 3.54 mA. Then, the read and write delay of the single and bi-layer MOSFET are 70.84 ps, 82.64 ps, 95.21 ps and 10.24 ps, 15.47 ps, 21.74 ps. Thus, the designed and simulated bi-layer optimized high k- dielectric medium for N-MOSFET with wild horse optimization performs better electrical characteristics than the single layer dielectric medium MOSFET.
  • An Exploration Performance-Efficiency and Knowledge Enhanced Text Classification using Gated Graph Neural Networks
    Pavithra Guru R, Tanuj Ravi Rao, Disshad K P, Bhaskar Jyoti Pathak
    2nd IEEE International Conference on Advances in Information Technology Icait 2024 Proceedings, 2024
    Dealing with the extensive length, feature sparsity, and substantial ambiguity inherent in text poses significant challenges for classification tasks. Prompt-learning offers a promising approach by inserting templates into inputs, effectively converting text classification tasks into predictions. However, existing methods often manually expand label words or solely rely on class names for information incorporation in prediction, leading to biases in classification. In this paper, we present a straightforward approach leveraging prompt based learning with knowledgeable expansion, aiming to address these limitations.
  • Enhanced Cardiovascular Disease Detection from Phonocardiogram Signals Using Deep Learning & Wavelet-Based Denoising
    R. Pavithra Guru, C Varun Kumar, M Chandra Sekhar, Maloth Vamshi
    2024 4th Asian Conference on Innovation in Technology Asiancon 2024, 2024
    Heart beat reveals crucial information about the heart condition at an early stage. The AI based diagnosis systems plays a crucial role for classifying heart sounds in CVD diagnosis. To accurately extract information from heart auscultation regarding a heart health condition, it’s essential that the heart sound signals remain uncontaminated by noise. This ensures that misclassification between normal and abnormal situations can be avoided. PCG signal contains crucial details regarding the functioning of heart and this information is used in diagnosis to detect various kinds of heart related diseases. Also, this PCG signals are very much distorted with noise which makes the denoising of the signal important for accurate detection of cardiovascular diseases (CVD). In this study, an automated CVD detection technique is introduced which detects various heart conditions from the PCG signal and also, we integrate a denoising technique to the diagnosis system which denoises the PCG signal if found to have any kind of internal or ambient noises. The findings presented in this paper suggest that our proposed architecture exhibited very good accuracy in detecting five classes of heart diseases with considerable denoising capabilities across various datasets featuring different noise levels and characteristics. This method is rooted on VGG based deep neural network model for detection of cardiac disease and the PCG denoising is done through wavelet denoising technique, represents a groundbreaking advancement aiming to increase the precision of automated CVD detection systems in identifying cardiac diseases within noisy environments.
  • Demodulating an acoustic signal stimulated by photo-thermal elastic energy conversion using quartz tuning forks
    M. Tamilselvi, T. M. Amirthalakshmi, R. Pavithra Guru, R. Neelaveni, G. Ramya, Yusuf Siraj Usmani, Mohd Zahid Ansari
    Optical and Quantum Electronics, 2024
  • Calorie Recognition from Food Images Using CNN for Nutritional Analysis
    R. Pavithra Guru, M. Ayyadurai, Tm Amirthalakshmi, R. Neelaven, K. Sujatha
    2023 International Conference on Data Science Agents and Artificial Intelligence Icdsaai 2023, 2023
  • An Estimation of the Performance of Deep Learning Based Hard Link Boot Caffe Neural Network for Network Anomaly Detection
    R. Deeptha, M. Ayyadurai, K. Sujatha, R.Pavithra Guru, D. Sasireka
    2023 International Conference on Data Science Agents and Artificial Intelligence Icdsaai 2023, 2023
  • Cloud Computing Based Workload Optimization using Long Short Term Memory Algorithm
    Ayyadurai M, Suresh S, Pavithra Guru. R, Ganesh B. S
    Proceedings of the 3rd International Conference on Smart Technologies in Computing Electrical and Electronics Icstcee 2022, 2022
  • An efficient VLSI circuit partitioning algorithm based on satin bowerbird optimization (SBO)
    R. Pavithra Guru, V. Vaithianathan
    Journal of Computational Electronics, 2020
  • Ant colony optimization based partition model for VLSI physical design
    Pavithra Guru R., V. Vaithianathan
    2020 International Conference on Computer Communication and Informatics Iccci 2020, 2020
  • Verilog module for on the Go implementation
    R. Pavithra Guru, C. Kalyana Sundram
    2016 International Conference on Energy Efficient Technologies for Sustainability Iceets 2016, 2016

RESEARCH OUTPUTS (PATENTS, SOFTWARE, PUBLICATIONS, PRODUCTS)

Method for Improving the Performance of Very Large Scale Integration
Patent Application No.: 202341049442A
Date of Publication: 01 September 2023

4G and 5G Technology-Based Several-Antennas for Smartphone App Using Multimode
Patent Application No.: 202241042837
Date of Publication: 29 July 2022

Method for Designing a Very Large Scale Integration (VLSI) Circuit
Patent Application No.: 202241060276
Journal No.: 44/2022
Date of Publication: 04 November 2022

AI-Based Smart Walking Stick for Visually Impaired with Alerting Unit (Granted)
Patent Application No.: 369641-001A
Date of Publication: 04 February 2022

Machine Learning for Price Discovery in Stock Market and Improving Reliability of Centralized Data Feeds
Patent Application No.: 201911042946A
Date of Publication: 04 September 2023

IoT-Based Remote Patient Monitoring Solution with AI-Driven Analytics
Patent Application No.: 202441012223A
Date of Publication: 15 March 2024

Automated Call Management System for Emergency Services Based on Artificial Intelligence
Patent Application No.: 202441049808
Date of Publication: 28 June 2024

🌍 International Patents

Device for Production of Cancer Vaccines
Patent Type: UK Design Patent
Patent No.: 6297329
Date Granted: 16 July 2023
Authority: UK Intellectual Property Office

Intrusion Detection System for Communication Networks
Patent Type: UK Design Patent
Patent No.: 6472447
Date Granted: 24 September 2025
Authority: UK Intellectual Property Office