Binary classification of lung cancer using vision transformer models on CT images Neha Thakur, Pradeep Chouksey, Ashok Sharma, Mayank Chopra, Parveen Sadotra, Sunil Kumar Discover Computing, 2026 Lung cancer remains a leading cause of cancer-related deaths, primarily due to late-stage detection. Although medical imaging and biopsy-based evaluations have improved, early identification of lung cancer continues to be challenging. To address this, we propose a Vision Transformer (ViT)-based model for binary lung nodule classification using computed tomography (CT) images. This study uses a Kaggle-hosted subset of the LIDC–IDRI dataset containing 315 CT nodule patches, where the original malignancy scores were converted into benign and malignant binary classes. Given the small dataset size, an extensive augmentation pipeline was designed to enhance model generalization. The lightweight ViT-Small/16 architecture demonstrated strong performance, achieving 92.3% accuracy, 90.5% precision, 93.8% recall, and a 92.1% F1-score. These results highlight the potential of compact transformer models for early lung cancer identification. This work is among the first to evaluate ViT-Small/16 on a small-scale CT nodule dataset using a tailored augmentation strategy for limited-data medical imaging.
A Novel Hybrid Threat Modeling Framework for IoT Security using STRIDE-DREAD and Machine Learning Gaurav Thakur, Pradeep Chouksey, Mayank Chopra, Parveen Sadotra, Neha Thakur, Diksha Sharma, Arpit Koundal, Shaina Mahajan Journal of Communications Software and Systems, 2026 The rapid growth of Internet of Things (IoT) deployments has increased security risks due to diverse device vulnerabilities, large scale interconnected environments, and the heterogeneity of communication protocols. Traditional threat assessment methods such as STRIDE and DREAD provide a structured foundation for identifying and categorizing security risks, yet they lack automated, real-time detection capabilities required for modern IoT systems that operate in dynamic and resource-constrained environments. To address these limitations, this study presents a hybrid threat modeling framework that integrates machine learning with STRIDE–DREAD to enhance threat identification, prioritization, and quantitative risk analysis. An ML-based classifier is trained on the CIC-BCCC-NRC TabularIoTAttack-2024 dataset to detect and categorize various IoT attack types, with particular emphasis on DDoS variants due to their high prevalence. Ensemble learning techniques are applied to pre-processed network traffic, enabling accurate, scalable, and computationally efficient classification suitable for deployment on lightweight IoT hardware. The proposed system achieves 92.5% detection accuracy, surpassing conventional STRIDE–DREAD assessments by 10–15% while providing enriched decision support for security analysts. Overall, the results demonstrate that integrating ML with established threat modeling methods significantly improves automation, reduces manual evaluation time, and strengthens the precision, adaptability, and operational reliability of IoT security assessment frameworks.
A comprehensive review on the hybrid BB84 E91 QKD protocol for enhanced security efficiency and practical hardware implementation in quantum cryptography Gaurav Thakur, Pradeep Chouksey, Mayank Chopra, Parveen Sadotra, Sunil Kumar Discover Computing, 2025 Cryptography is a popular term in this modern world of digital era which has achieved a great focus and interest in the modern inter-connected digital world. There is a lot of productive research underway in such domain of communication so as to make it more secure, because it is the most fundamental technique used in day-to-day life of the common people as well as cyber professionals. Over time, the field of cryptography has experienced various breakthroughs that are being driven by the progress of quantum cryptography. Unsurprisingly, in this digital age of modern times, quantum cryptography is one of the new and efficient means of secure communication. Quantum cryptography syndicates the concepts of quantum physics to project safe channels to achieve secure communication. The calibre of research being done in the realm of quantum cryptography is set to rise amidst an increasingly volatile digital world. This paper discusses the latest trends of quantum cryptography specifically focussing on the quantum cryptography protocols with a significant contribution of recently published essential research papers with quite good productivity. The primary focus in this paper is concerned with the investigation of quantum cryptography protocols on various grounds such as quantum key distribution strategies, quantum distribution protocols and a comprehensive comparison between the past techniques and recent advancements. In light of these cutting-edge developments, a hybrid framework utilizing quantum cryptography protocols is also proposed. This study synthesizes existing findings to offer insights into quantum cryptography protocols, future challenges, and potential applications.
Fortifying E-Voting Systems: Integrating Visual Cryptography with ECC and ChaCha20-Poly1305 for Enhanced Security Gaurav Thakur, Pradeep Chouksey, Mayank Chopra, Parveen Sadotra Journal of Communications Software and Systems, 2025 The growing reliance on digital technologies demands an urgent advancement of a secure framework for remote voting systems. This paper proposes a novel e-voting framework which is reinforced by a combination of visual cryptography, elliptic curve cryptography (ECC) and ChaCha20-Poly1305 encryption methods. Visual cryptography ensures the anonymity of voters, ECC provides a robust public key infrastructure and ChaCha20-Poly1305 provides authenticated encryption to ensure data integrity. The proposed approach eliminates some of the most vulnerable weaknesses in electronic voting systems, such as unauthorized access and manipulation, while ensuring transparency and verifiability. The complete proposed framework is thus feasible for practical application, as the results prove its effectiveness and efficiency in protecting remote voting procedures.
Multi-Class Identification of Lung Cancer Subtypes Using Swin Transformer Neha Thakur, Pradeep Chouksey, Parveen Sadotra, Mayank Chopra 2025 International Conference on Engineering Innovations and Technologies Icoeit 2025, 2025 Lung cancer is still among the deadliest in the world and requires early detection and accurate determination for better outcomes for patients. In this research, we provide a deep learning-based model for classifying different forms of lung cancer, including adenocarcinoma, large-cell carcinoma, squamous-cell carcinoma, and normal tissues, using the Swin Transformer model. A systematic data preprocessing process was used that involved resizing, normalization, and standardization to ensure proper feature extraction and model convergence. Swin Transformer's hierarchical structure and shifted window attention mechanism promote efficient multi-scale feature learning for better classification performance. The model obtains an overall accuracy of 95.5%, precision of 96.1%, recall of 95.5%, and specificity of 95.3%, which proves the efficiency of the model in classifying squamous-cell carcinoma, adeno-carcinoma, large-cell carcinoma, and normal lung tissue. The outcome reflects the ability of transformer-based models in medical imaging for accurate lung cancer detection. More work will be directed to-wards advancing feature representation further and investigating hybrid deep learning methods to improve classification accuracy.
Generative AI for the Creation of Images Sahil, Pratham, Neha, Parveen Sadotra, Pradeep Chouksey, Mayank Chopra, Aditya Thakur Lecture Notes in Electrical Engineering, 2025
Research review on task scheduling algorithm for green cloud computing Parveen Sadotra, Pradeep Chouksey, Mayank Chopra, Rabia Koser, Rishikesh Rawat Scalable Modeling and Efficient Management of Iot Applications, 2024 In green cloud computing, task scheduling entails assigning tasks to virtual machines in an approach that minimizes energy use whilst still reaching the performance targets. Green cloud computing is an evolving field that lowers the energy and carbon footprint of systems for using the cloud. In green cloud computing, task planning is a crucial problem because it determines how computational resources are allotted to workloads in order to decrease energy consumption and increase efficiency. Different task-scheduling techniques have been put forth in recent years for green cloud computing. The authors look at some current studies on task scheduling methods for green cloud computing in this overview of the literature.
Enhancing E-Voting Security with Multi-Factor Authentication Using Fingerprint and Cryptography Protocols in India Gaurav Thakur, Pradeep Chouksey, Mayank Chopra, Parveen Sadotra Proceedings 2024 International Conference on Progressive Innovations in Intelligent Systems and Data Science Icpids 2024, 2024 The integrity and security of e-voting systems are the core to ensure free and fair elections in any country. As the process involves an enormous, highly diversified population in India, e-voting systems are open to a hugely diversified variety of imminent threats, therefore, extraordinary security measures are incumbent on safeguarding them. This paper deals with the implementation of multi-factor authentication in a manner intended to improve the security of e-voting systems by integrating fingerprint biometrics with visual cryptography and also the state-of-the-art cryptographic protocols. It is proposed that some of these vulnerabilities associated with conventional methods of authentication should be tackled by the use of biometric verification, which can provide each voter with a unique and immutable identity. A detailed survey is presented on the current challenges to e-voting security in India and the battle-hardened limitations of existing systems, providing a compelling case for robust MFA solutions. The proposed system uses fingerprint authentication as its primary factor and further fortifies this with cryptographic methods that ensure confidentiality, integrity, and authenticity of the entire process of voting, right from the voter's end to the final tallying of votes. Finally, it provides end-to-end security from voter authentication to vote-tallying using symmetric and asymmetric encryption, as well as secure hash functions. The approach is tested against common security threats with simulations and practical implementation on voter impersonation, data tampering, and unauthorized access. Our results show that a high degree of enhancement in the security and reliability of e-voting schemes can be achieved by integrating fingerprint-based MFA with cryptographic protocols. This work is, therefore, a part of the modernization and security effort within the electoral process in India and is hopefully going to be one prospective solution for protecting the democratic rights of its citizens.
Intrusion Detection in Smart Homes: A Comprehensive Review Parveen Sadotra, Pardeep Chouksey, Mayank Chopra, Gaurav Thakur, Malavath Hanmanth Nayak Proceedings 2024 International Conference on Progressive Innovations in Intelligent Systems and Data Science Icpids 2024, 2024
