Utpal Roy
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Utpal Roy has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Soumen Roy
University of Calcutta
6 shared publications - Devadatta Sinha
Department of Computer Science and Engineering, University of Calcutta, Kolkata, India
5 shared publications - Rounak Biswas
Visva-Bharati University
5 shared publications - Anandarup Roy
Sarojini Naidu College for Women
4 shared publications - Jayati Mukherjee
Computer Science and Engineering, Academy of Technology, Hoogly, India
3 shared publications - Soumen Saha
University of Calcutta, Kolkata, India
3 shared publications
Latest Publications
- 15 June 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Pattern Recognition and Artificial Intelligence
Brain Tumor and Haemangioma Discrimination Framework Using Unsupervised Feature Selection
Nowadays, in medical imaging, discrimination of brain tumors and haemangiomas is a major challenge due to tumor heterogeneity, overlapping vascular structures, and limitations in existing segmentation approaches. To address these issues, a Brain Tumor and haemangioma discrimination framework is proposed that integrates unsupervised feature selection with a multi-stage pipeline. First, the Dense Feature Reuse and Volumetric U-Net (DFR-V-Net) is employed to segment the 3D MRI scans, enhancing feature representation while reducing redundancy in tumor boundaries. Next, a Weighted Voxel Upscaling method is introduced to correct partial volume effects and improve voxel-wise resolution, facilitating more accurate separation of tumor and normal tissues. The enhanced volumetric features are then processed using a Contiguous Factor Disentangled Variance Autoencoder (CFD–VAE), which isolates vascularization and blood flow-based characteristics that differentiate haemangiomas from malignant tumors. Finally, a Binary Gaussian-Initiated Self-Organizing Map (SOM) classifier performs unsupervised discrimination of tumor categories by modeling Gaussian cluster priors across intensity gradients and vascular orientations. The experimental validation on the proposed framework achieved 99.7% accuracy, 99.8% precision, 99.7% recall, 99.5% F 1-score, and 98.4% Dice Similarity Coefficient (DSC), which is significantly better than other conventional methods. The integration of voxel-level correction, disentangled vascular feature learning, and Gaussian-based unsupervised classification collectively enhances diagnostic reliability. This proposed framework provides a clinically practical solution for the precise and rapid differentiation of brain tumors and haemangiomas in MRI-based diagnosis.
- 15 June 2026
- Published by World Scientific Pub Co Pte Ltd in International Journal of Pattern Recognition and Artificial Intelligence
Brain Tumor and Haemangioma Discrimination Framework Using Unsupervised Feature Selection
Nowadays, in medical imaging, discrimination of brain tumors and haemangiomas is a major challenge due to tumor heterogeneity, overlapping vascular structures, and limitations in existing segmentation approaches. To address these issues, a Brain Tumor and haemangioma discrimination framework is proposed that integrates unsupervised feature selection with a multi-stage pipeline. First, the Dense Feature Reuse and Volumetric U-Net (DFR-V-Net) is employed to segment the 3D MRI scans, enhancing feature representation while reducing redundancy in tumor boundaries. Next, a Weighted Voxel Upscaling method is introduced to correct partial volume effects and improve voxel-wise resolution, facilitating more accurate separation of tumor and normal tissues. The enhanced volumetric features are then processed using a Contiguous Factor Disentangled Variance Autoencoder (CFD–VAE), which isolates vascularization and blood flow-based characteristics that differentiate haemangiomas from malignant tumors. Finally, a Binary Gaussian-Initiated Self-Organizing Map (SOM) classifier performs unsupervised discrimination of tumor categories by modeling Gaussian cluster priors across intensity gradients and vascular orientations. The experimental validation on the proposed framework achieved 99.7% accuracy, 99.8% precision, 99.7% recall, 99.5% F 1-score, and 98.4% Dice Similarity Coefficient (DSC), which is significantly better than other conventional methods. The integration of voxel-level correction, disentangled vascular feature learning, and Gaussian-based unsupervised classification collectively enhances diagnostic reliability. This proposed framework provides a clinically practical solution for the precise and rapid differentiation of brain tumors and haemangiomas in MRI-based diagnosis.
- 10 February 2026
- Published by Springer Nature
- p. 571-583
Automatic Speech Recognition Model for Low Resource Kudmali Language
- 10 February 2026
- Published by Springer Nature
- p. 571-583
Automatic Speech Recognition Model for Low Resource Kudmali Language
- 31 January 2026
- Published by Springer Nature
- p. 221-235
NLP Tools and Resources to Revitalize Potentially Vulnerable Low Resourced Kudmali Language: A Brief Overview
- 31 January 2026
- Published by Springer Nature
- p. 221-235
NLP Tools and Resources to Revitalize Potentially Vulnerable Low Resourced Kudmali Language: A Brief Overview
- 19 September 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of Quantum Information
- Vol. 23 (07)
Quantum noise-driven randomness for password generation: Analysis via simulation and FakeProvider
This study introduces a quantum-enhanced approach to password generation using two distinct quantum randomness sources: classical measurement of superposition states and depolarizing noise. These sources overcome the shortcomings of classical pseudo-random number generators. Notably, this work pioneers integrating depolarizing noise as a controlled entropy source alongside superposition, offering a novel alternative to conventional quantum randomness methods. Quantum circuits, implemented via Qiskit, were executed on AerSimulator (idealized simulation) and FakeProvider (hardware-inspired noise model) backends, producing binary sequences through XOR and Concatenation modes. These sequences were converted into passwords of lengths 16 and 25 across datasets of 10K, 100K, and 500K, using Binary Character Mapping and Randomized reseeding methods. Binary entropy’s randomness was verified by the NIST Statistical Test Suite, with all p-values above 0.01. Generated password strength was evaluated through Clustering (Levenshtein distance), Entropy (3.84 to 4.40 bits per character), Hamming Distance, Uniqueness (100 percent), and ZXCVBN scoring (Score 4). Results from both backend-based password datasets were nearly identical, confirming simulated and hardware-based quantum randomness as a reliable and scalable alternative to systems. This work positions quantum noise-driven password generation as a robust cryptographic tool with the potential for real-world security applications.This research investigates the effectiveness of quantum noise, derived from superposition states and depolarizing channels, as a high-entropy source for generating cryptographically secure passwords. Quantum circuits implemented on the AerSimulator and FakeProvider backends produce random binary sequences, which are processed through chaotic mixing and rejection sampling to ensure uniform randomness. These sequences are rigorously evaluated using Shannon entropy, min-entropy and the NIST statistical test suite, alongside password strength assessments via tools like ZXCVBN. Results indicate exceptional randomness, with an average Shannon entropy of 3.84 ± 0.4 bits per character and 100% uniqueness across datasets of 20,000 passwords. The FakeProvider backend slightly outperforms the AerSimulator, achieving full compliance with NIST tests, while both produce passwords with top-tier strength scores (ZXCVBN Score 4). Comparative analyses demonstrate that quantum noise surpasses classical pseudo-random number generators in entropy density and resistance to brute-force and dictionary-based attacks. This study validates quantum noise as a superior entropy source, offering significant potential for enhancing password security in cryptographic applications and paving the way for quantum-enhanced cybersecurity solutions.
- 19 September 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of Quantum Information
- Vol. 23 (07)
Quantum noise-driven randomness for password generation: Analysis via simulation and FakeProvider
This study introduces a quantum-enhanced approach to password generation using two distinct quantum randomness sources: classical measurement of superposition states and depolarizing noise. These sources overcome the shortcomings of classical pseudo-random number generators. Notably, this work pioneers integrating depolarizing noise as a controlled entropy source alongside superposition, offering a novel alternative to conventional quantum randomness methods. Quantum circuits, implemented via Qiskit, were executed on AerSimulator (idealized simulation) and FakeProvider (hardware-inspired noise model) backends, producing binary sequences through XOR and Concatenation modes. These sequences were converted into passwords of lengths 16 and 25 across datasets of 10K, 100K, and 500K, using Binary Character Mapping and Randomized reseeding methods. Binary entropy’s randomness was verified by the NIST Statistical Test Suite, with all p-values above 0.01. Generated password strength was evaluated through Clustering (Levenshtein distance), Entropy (3.84 to 4.40 bits per character), Hamming Distance, Uniqueness (100 percent), and ZXCVBN scoring (Score 4). Results from both backend-based password datasets were nearly identical, confirming simulated and hardware-based quantum randomness as a reliable and scalable alternative to systems. This work positions quantum noise-driven password generation as a robust cryptographic tool with the potential for real-world security applications.This research investigates the effectiveness of quantum noise, derived from superposition states and depolarizing channels, as a high-entropy source for generating cryptographically secure passwords. Quantum circuits implemented on the AerSimulator and FakeProvider backends produce random binary sequences, which are processed through chaotic mixing and rejection sampling to ensure uniform randomness. These sequences are rigorously evaluated using Shannon entropy, min-entropy and the NIST statistical test suite, alongside password strength assessments via tools like ZXCVBN. Results indicate exceptional randomness, with an average Shannon entropy of 3.84 ± 0.4 bits per character and 100% uniqueness across datasets of 20,000 passwords. The FakeProvider backend slightly outperforms the AerSimulator, achieving full compliance with NIST tests, while both produce passwords with top-tier strength scores (ZXCVBN Score 4). Comparative analyses demonstrate that quantum noise surpasses classical pseudo-random number generators in entropy density and resistance to brute-force and dictionary-based attacks. This study validates quantum noise as a superior entropy source, offering significant potential for enhancing password security in cryptographic applications and paving the way for quantum-enhanced cybersecurity solutions.
- 1 September 2025
- Published by Bentham Science Publishers Ltd. in Recent Advances in Computer Science and Communications
- Vol. 18 (6) , 1-13
Efficacy of Keystroke Dynamics-based User Authentication in the Face of Language Complexity
Introduction: This study investigates the impact of language complexity on Keystroke Dynamics (KD) and its implications for accurate KD-based user authentication system performance in smartphones. Methods: This research meticulously analyzes keystroke patterns using 160 volunteers, including both frequently typed and infrequently typed texts. Our analysis of 12 anomaly detection algorithms reveals that a simple text-based KD system consistently outperforms its complex counterpart with superior Equal Error Rates (EERs). Results: As a result, the Scaled Manhattan anomaly detector achieves an EER of 2.48% for simple text and an improvement over 2.98% for complex text. The incorporation of soft biometrics further enhances algorithmic performance, emphasizing strategies to build resilience into KD-based user authentication systems. Conclusion: Throughout this study, the importance of text complexity is emphasized, and innovative pathways are introduced to strengthen KD-based user authentication paradigms.
- 1 September 2025
- Published by Bentham Science Publishers Ltd. in Recent Advances in Computer Science and Communications
- Vol. 18 (6) , 1-13
Efficacy of Keystroke Dynamics-based User Authentication in the Face of Language Complexity
Introduction: This study investigates the impact of language complexity on Keystroke Dynamics (KD) and its implications for accurate KD-based user authentication system performance in smartphones. Methods: This research meticulously analyzes keystroke patterns using 160 volunteers, including both frequently typed and infrequently typed texts. Our analysis of 12 anomaly detection algorithms reveals that a simple text-based KD system consistently outperforms its complex counterpart with superior Equal Error Rates (EERs). Results: As a result, the Scaled Manhattan anomaly detector achieves an EER of 2.48% for simple text and an improvement over 2.98% for complex text. The incorporation of soft biometrics further enhances algorithmic performance, emphasizing strategies to build resilience into KD-based user authentication systems. Conclusion: Throughout this study, the importance of text complexity is emphasized, and innovative pathways are introduced to strengthen KD-based user authentication paradigms.
- 29 May 2025
- Published by Springer Nature in Advances in Computational Intelligence
- Vol. 5 (2) , 1-26
A predictive contrivance for recognising traits in keystroke dynamics
- 29 May 2025
- Published by Springer Nature in Advances in Computational Intelligence
- Vol. 5 (2) , 1-26
A predictive contrivance for recognising traits in keystroke dynamics
- 15 March 2025
- Published by Springer Nature
- p. 276-286
Unsupervised Feature Selection for High-Dimensional Data Using Clustering and Multi-Objective Optimization
- 15 March 2025
- Published by Springer Nature
- p. 276-286
Unsupervised Feature Selection for High-Dimensional Data Using Clustering and Multi-Objective Optimization
- 5 February 2025
- Published by Springer Nature
- p. 35-45
Exploring the Limits of Classical Random Number Generation and Unveiling Quantum Alternatives
- 5 February 2025
- Published by Springer Nature
- p. 35-45
Exploring the Limits of Classical Random Number Generation and Unveiling Quantum Alternatives