Pintu Chandra Paul
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Pintu Chandra Paul has a total of 19 co-authors. Here are the top co-authors by number of shared publications:
- Amena Begum
Department of Information and Communication Technology, Comilla University, Cumilla 3500, Bangladesh
2 shared publications - Anup Majumder
Department of Computer Science and Engineering, Jahangirnagar University, Savar 1342, Bangladesh
2 shared publications - Jun Zhuang
Boise State University
2 shared publications - Kashmi Sultana
Department of Information and Communication Technology, Comilla University, Cumilla 3500, Bangladesh
2 shared publications - Maqsudur Rahman
Department of Computer Science, Boise State University, Boise, ID 83706, USA
2 shared publications - Nahida Akter
Department of Information and Communication Technology, Comilla University, Cumilla 3500, Bangladesh
2 shared publications
Latest Publications
- Open Access
- 30 September 2026
- Published by FOREX Publication in International Journal of Electrical and Electronics Research
- Vol. 14 (3) , 842
Error Signal Based Method for Real Time Pitch Extraction from Noisy Speech
Pitch estimation efficiency always gets affected by two major phenomena. The first one is additive noise which by nature are not correlated with the source speech signal. The second one is the vocal tract effect which creates unnecessary peaks within the signal waveform, resulting inaccurate detection of the pitch value. This paper proposes a Modified Weighted Autocorrelation Function (MWAF) executed in the Linear Predictive Coding (LPC) residual domain, instead of directly executing over the noisy source speech signal. Converting the noisy source speech signal to LP residual helps eliminating the vocal tract effects and emphasizing the necessary effects of glottal excitation. The Autocorrelation Function (ACF) and the Circular Average Magnitude Difference Function (CAMDF) are directly applied on this residual. The proposed MWAF is then calculated by multiplying the ACF with the reciprocal of the CAMDF, which enhances the periodic components as well as reduces the noise induced false peaks. Experimental results were evaluated on the KEELE and NTT databases across White, Babble, Train and HF Channel noise conditions. The result shows that our proposed algorithm reduces the Gross Pitch Error (GPE) by an average of 40.34% compared to baseline methods such as BaNa, PEFAC and WAF. Furthermore, the algorithm achieves an overall computational gain of 53.44% compared to BaNa, PEFAC and WAF, offering a stable solution for pitch tracking in degraded acoustic conditions.
- 30 September 2026
- Published by FOREX Publication in International Journal of Electrical and Electronics Research
- Vol. 14 (3) , 842
Open AccessError Signal Based Method for Real Time Pitch Extraction from Noisy Speech
Pitch estimation efficiency always gets affected by two major phenomena. The first one is additive noise which by nature are not correlated with the source speech signal. The second one is the vocal tract effect which creates unnecessary peaks within the signal waveform, resulting inaccurate detection of the pitch value. This paper proposes a Modified Weighted Autocorrelation Function (MWAF) executed in the Linear Predictive Coding (LPC) residual domain, instead of directly executing over the noisy source speech signal. Converting the noisy source speech signal to LP residual helps eliminating the vocal tract effects and emphasizing the necessary effects of glottal excitation. The Autocorrelation Function (ACF) and the Circular Average Magnitude Difference Function (CAMDF) are directly applied on this residual. The proposed MWAF is then calculated by multiplying the ACF with the reciprocal of the CAMDF, which enhances the periodic components as well as reduces the noise induced false peaks. Experimental results were evaluated on the KEELE and NTT databases across White, Babble, Train and HF Channel noise conditions. The result shows that our proposed algorithm reduces the Gross Pitch Error (GPE) by an average of 40.34% compared to baseline methods such as BaNa, PEFAC and WAF. Furthermore, the algorithm achieves an overall computational gain of 53.44% compared to BaNa, PEFAC and WAF, offering a stable solution for pitch tracking in degraded acoustic conditions.
- Open Access
- 23 June 2026
- Published by MDPI AG in Journal of Imaging
- Vol. 12 (7) , 275
Benchmarking Barren Plateau Mitigation Strategies in Quantum Neural Networks on Standard and Medical Image Datasets
Barren plateaus (BPs) pose a major trainability challenge for quantum neural networks (QNNs) by causing gradients to concentrate near zero as circuit size, depth, or expressibility increases. This study presents a comparative benchmark of 10 BP mitigation strategies across six qubit settings (2, 4, 8, 12, 16, and 20) and three datasets of increasing complexity: Iris, MNIST, and MedMNIST. The evaluated methods include eight initialization-based strategies (Beta, Gaussian, Uniform Norm, CNN-based initialization, He-normal, He-uniform, Xavier-normal, and Xavier-uniform), one model-based variational encoder, and one optimization-based time-nonlocal Fourier parameterization. Experiments were implemented using PennyLane 3.10 and PyTorch 2.5 with simulator backends. We evaluate trainability using gradient variance and training loss, and we clarify that the benchmark analyzes simulated QNN optimization behavior rather than hardware-noise-resilient or noisy-label learning. Across the tested two-layer circuit configurations, the mitigation strategies maintained measurable gradient variance and stable loss reduction, suggesting that severe barren plateau behavior was not observed under the benchmark conditions. CNN-based and Beta initialization showed strong empirical behavior in variance retention and convergence speed, while Gaussian initialization was comparatively weaker in higher-dimensional settings. The study provides a reproducible benchmark structure for comparing BP mitigation behavior and identifies important limitations related to circuit depth, hardware noise, feature encoding, and classification performance that should be addressed in future QNN benchmarking.
- 23 June 2026
- Published by MDPI AG in Journal of Imaging
- Vol. 12 (7) , 275
Open AccessBenchmarking Barren Plateau Mitigation Strategies in Quantum Neural Networks on Standard and Medical Image Datasets
Barren plateaus (BPs) pose a major trainability challenge for quantum neural networks (QNNs) by causing gradients to concentrate near zero as circuit size, depth, or expressibility increases. This study presents a comparative benchmark of 10 BP mitigation strategies across six qubit settings (2, 4, 8, 12, 16, and 20) and three datasets of increasing complexity: Iris, MNIST, and MedMNIST. The evaluated methods include eight initialization-based strategies (Beta, Gaussian, Uniform Norm, CNN-based initialization, He-normal, He-uniform, Xavier-normal, and Xavier-uniform), one model-based variational encoder, and one optimization-based time-nonlocal Fourier parameterization. Experiments were implemented using PennyLane 3.10 and PyTorch 2.5 with simulator backends. We evaluate trainability using gradient variance and training loss, and we clarify that the benchmark analyzes simulated QNN optimization behavior rather than hardware-noise-resilient or noisy-label learning. Across the tested two-layer circuit configurations, the mitigation strategies maintained measurable gradient variance and stable loss reduction, suggesting that severe barren plateau behavior was not observed under the benchmark conditions. CNN-based and Beta initialization showed strong empirical behavior in variance retention and convergence speed, while Gaussian initialization was comparatively weaker in higher-dimensional settings. The study provides a reproducible benchmark structure for comparing BP mitigation behavior and identifies important limitations related to circuit depth, hardware noise, feature encoding, and classification performance that should be addressed in future QNN benchmarking.
- Open Access
- 28 May 2026
- Published by MDPI AG in Journal of Imaging
- Vol. 12 (6) , 232
Systems-Level Support for Hybrid Quantum-Classical Learning: A Systematic Review with a Medical Imaging Translation Lens
Hybrid quantum-classical learning pipelines combine conventional accelerators, quantum runtimes, and quantum processing units (QPUs), creating scheduling, memory, isolation, encoding, and deployment challenges that are not captured by application-level quantum machine learning surveys alone. This paper presents a systematic review of runtime and systems mechanisms for hybrid quantum-classical workloads, with medical imaging used as a translation lens rather than as an exclusive inclusion boundary. Following a PRISMA-aligned review process, we screened 364 records and synthesized 40 studies published between 2020 and 2025. Each study was coded by systems layer, application grounding, noisy-label relevance, and evaluation maturity. The coding shows that the corpus combines direct medical evidence with broader transferable systems evidence: 8 studies directly evaluated medical data, 12 were medically motivated, and 20 were generic systems studies. Across the corpus, the strongest support concerns hybrid orchestration, qubit/resource allocation, classical–quantum data movement, and container-based reproducibility, whereas evidence remains limited for realistic clinical operation, end-to-end remote-QPU workflows, multi-tenant isolation, and noisy-label retraining loops. We contribute an evidence map, a direct/indirect/interpretive evidence distinction, and cross-layer design guidelines for future hybrid quantum-classical imaging pipelines in regulated settings.
- 28 May 2026
- Published by MDPI AG in Journal of Imaging
- Vol. 12 (6) , 232
Open AccessSystems-Level Support for Hybrid Quantum-Classical Learning: A Systematic Review with a Medical Imaging Translation Lens
Hybrid quantum-classical learning pipelines combine conventional accelerators, quantum runtimes, and quantum processing units (QPUs), creating scheduling, memory, isolation, encoding, and deployment challenges that are not captured by application-level quantum machine learning surveys alone. This paper presents a systematic review of runtime and systems mechanisms for hybrid quantum-classical workloads, with medical imaging used as a translation lens rather than as an exclusive inclusion boundary. Following a PRISMA-aligned review process, we screened 364 records and synthesized 40 studies published between 2020 and 2025. Each study was coded by systems layer, application grounding, noisy-label relevance, and evaluation maturity. The coding shows that the corpus combines direct medical evidence with broader transferable systems evidence: 8 studies directly evaluated medical data, 12 were medically motivated, and 20 were generic systems studies. Across the corpus, the strongest support concerns hybrid orchestration, qubit/resource allocation, classical–quantum data movement, and container-based reproducibility, whereas evidence remains limited for realistic clinical operation, end-to-end remote-QPU workflows, multi-tenant isolation, and noisy-label retraining loops. We contribute an evidence map, a direct/indirect/interpretive evidence distinction, and cross-layer design guidelines for future hybrid quantum-classical imaging pipelines in regulated settings.