Usha Desai
Research Subjects
Usha Desai has a total of 88 co-authors. Here are the top co-authors by number of shared publications:
- Ravichander Janapati
School of CSAI SR University,Warangal,India
4 shared publications Viswaprakash Babu
Kaveri University,Department of Electrical and Electronics Engineering,Telangana,India
4 shared publicationsGhousia Sanober Sabreen
Ballari Institute of Technology and Management,Department of Electronics and Communication,Ballari,India
3 shared publicationsAkash Saxena
School of Engineering and Technology, Central University of Haryana,Department of Electrical Engineering,India
2 shared publicationsAkshaya D. Shetty
S.E.A College of Engineering Technology,Dept. of ECE,Bengaluru,India
2 shared publications- Ankit Vijayvargiya
School of Technology, Dhirubhai Ambani University
2 shared publications
Latest Publications
- Open Access
- 1 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Source-Constrained Claim Verification and Risk-Adaptive Retrieval for Evidence-Grounded Healthcare Search: Development and Benchmark Study (Preprint)
BACKGROUND Access to reliable healthcare information depends on finding relevant sources and accurately representing what they establish. Agentic search systems can retrieve information from provider registries, biomedical literature and interoperable health records, but a valid citation does not necessarily support the claim attached to it. Healthcare information seekers increasingly use conversational AI to find providers and understand biomedical evidence. Retrieval-augmented systems can connect generated answers to external sources, yet the presence of a citation does not establish that it supports the statement beside it. This distinction matters when a system combines provider registries, research articles and health data records, because each source supports different kinds of claims. OBJECTIVE Evaluating claim-level evidence verification and evidence-sufficiency decisions separately and together, while measuring unsupported claims, answer completeness and appropriate abstention. METHODS The Source-Constrained Claim Graph (SCCG) and Risk-Adaptive Retrieval and Abstention Controller (RARC) is implemented and tested and their effectiveness is reported. RESULTS The primary outcome depends on rate of unsupported factual claims; secondary outcomes includes claim-to-source accuracy, appropriate abstention, answer completeness and expert-rated usefulness. The existing platform passed 11 of 11 acceptance cases with Gemini and 10 of 11 with Gemma; the remaining Gemma case ended in an interoperability workflow timeout. Both paths passed all applicable deterministic citation, retrieval and unsupported-request checks. Median response times were 10.60 and 167.41 seconds, respectively, under different runtime conditions. CONCLUSIONS This study examined an open-source healthcare search platform using 11 acceptance cases covering provider discovery, biomedical research, health information, interoperability and unsupported requests. The Gemini path passed all 11 cases, while the Gemma path passed 10; its remaining interoperability case ended in a runtime timeout. Both paths completed the applicable citation, retrieval and rejection checks. These findings show that the existing platform followed its expected workflows in this small benchmark, but they do not establish whether every answer claim was supported by its cited evidence. These results describe the existing platform; the proposed evidence-sufficiency and claim-verification components have not yet been evaluated.
- 1 October 2026
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessSource-Constrained Claim Verification and Risk-Adaptive Retrieval for Evidence-Grounded Healthcare Search: Development and Benchmark Study (Preprint)
BACKGROUND Access to reliable healthcare information depends on finding relevant sources and accurately representing what they establish. Agentic search systems can retrieve information from provider registries, biomedical literature and interoperable health records, but a valid citation does not necessarily support the claim attached to it. Healthcare information seekers increasingly use conversational AI to find providers and understand biomedical evidence. Retrieval-augmented systems can connect generated answers to external sources, yet the presence of a citation does not establish that it supports the statement beside it. This distinction matters when a system combines provider registries, research articles and health data records, because each source supports different kinds of claims. OBJECTIVE Evaluating claim-level evidence verification and evidence-sufficiency decisions separately and together, while measuring unsupported claims, answer completeness and appropriate abstention. METHODS The Source-Constrained Claim Graph (SCCG) and Risk-Adaptive Retrieval and Abstention Controller (RARC) is implemented and tested and their effectiveness is reported. RESULTS The primary outcome depends on rate of unsupported factual claims; secondary outcomes includes claim-to-source accuracy, appropriate abstention, answer completeness and expert-rated usefulness. The existing platform passed 11 of 11 acceptance cases with Gemini and 10 of 11 with Gemma; the remaining Gemma case ended in an interoperability workflow timeout. Both paths passed all applicable deterministic citation, retrieval and unsupported-request checks. Median response times were 10.60 and 167.41 seconds, respectively, under different runtime conditions. CONCLUSIONS This study examined an open-source healthcare search platform using 11 acceptance cases covering provider discovery, biomedical research, health information, interoperability and unsupported requests. The Gemini path passed all 11 cases, while the Gemma path passed 10; its remaining interoperability case ended in a runtime timeout. Both paths completed the applicable citation, retrieval and rejection checks. These findings show that the existing platform followed its expected workflows in this small benchmark, but they do not establish whether every answer claim was supported by its cited evidence. These results describe the existing platform; the proposed evidence-sufficiency and claim-verification components have not yet been evaluated.
- 25 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 421-428
AI Secured Simulation of Wireline Blueprints
- 25 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 421-428
AI Secured Simulation of Wireline Blueprints
- 9 June 2026
- Published by Taylor & Francis
- p. 337-355
Exploring the Impact: A Comprehensive Survey of IoT Applications in Healthcare
- 9 June 2026
- Published by Taylor & Francis
- p. 337-355
Exploring the Impact: A Comprehensive Survey of IoT Applications in Healthcare
- 9 June 2026
- Published by Taylor & Francis
- p. 87-123
A Comprehensive Study on Human Factors in Cyber-sustainability
- 9 June 2026
- Published by Taylor & Francis
- p. 87-123
A Comprehensive Study on Human Factors in Cyber-sustainability
- Open Access
- 1 May 2026
- Published by Elsevier in Intelligence-Based Medicine
- Vol. 14
Reinforcement learning and GAN-augmented autoencoders for H&E-Stained bone biopsy image analysis
- 1 May 2026
- Published by Elsevier in Intelligence-Based Medicine
- Vol. 14
Open AccessReinforcement learning and GAN-augmented autoencoders for H&E-Stained bone biopsy image analysis
- Open Access
- 30 March 2026
- Published by Elsevier in Journal of Orthopaedic Reports
- Vol. 5 (4)
Non-parametric analysis of lumbar disc space variations in lower back pain patients
- 30 March 2026
- Published by Elsevier in Journal of Orthopaedic Reports
- Vol. 5 (4)
Open AccessNon-parametric analysis of lumbar disc space variations in lower back pain patients
- 11 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1708-1713
Hybrid VMD–ICA Framework for Artifact-Resilient Single-Channel EEG in Driver Drowsiness Detection
Fatigue is a significant contributor to road accidents in real-time drowsiness systems, which are reliable, and necessary. Nevertheless, the single channel electroencephalography (EEG) recorded by consumer-grade headsets characterize low signal-to-noise ratio because of ocular, muscular and motion artifacts, which restricts the use in real-world Brain-Computer Interface (BCI) systems. This paper introduces a hybrid signal-processing framework using Variational Mode Decomposition (VMD) and Independent Component Analysis (ICA) to enhance single-channel EEG quality. In this paper, raw EEG from Sleepy Driver Brainwave dataset is acquired into finite number of intrinsic modes of VMD to represent the oscillatory part in frequency bands. These modes considered as pseudo-multichannel signals and subjected to ICA to isolate and eliminate artifact-dominated signals. The reconstructed EEG feature extracted using power spectral density and entropy measures, and the drowsiness classification is performed with Support Vector Machine (SVM) and Convolutional Neural Network (CNN) models. The experimental findings indicates, a substantial gain in signal quality, and average SNR increase by 4.6 dB and classification rate by 9.2% over the traditional filtering and ICA approach. Hybrid VMD-ICA method provides effective maintenance of cognitive rhythm integrity and elimination of non-neural artifacts makes a strong and lightweight drowsiness detection with single-channel EEG. Architecture provides a viable future of intelligent transportation and human-machine interfaces for efficient neuroergonomic systems.
- 11 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1708-1713
Hybrid VMD–ICA Framework for Artifact-Resilient Single-Channel EEG in Driver Drowsiness Detection
Fatigue is a significant contributor to road accidents in real-time drowsiness systems, which are reliable, and necessary. Nevertheless, the single channel electroencephalography (EEG) recorded by consumer-grade headsets characterize low signal-to-noise ratio because of ocular, muscular and motion artifacts, which restricts the use in real-world Brain-Computer Interface (BCI) systems. This paper introduces a hybrid signal-processing framework using Variational Mode Decomposition (VMD) and Independent Component Analysis (ICA) to enhance single-channel EEG quality. In this paper, raw EEG from Sleepy Driver Brainwave dataset is acquired into finite number of intrinsic modes of VMD to represent the oscillatory part in frequency bands. These modes considered as pseudo-multichannel signals and subjected to ICA to isolate and eliminate artifact-dominated signals. The reconstructed EEG feature extracted using power spectral density and entropy measures, and the drowsiness classification is performed with Support Vector Machine (SVM) and Convolutional Neural Network (CNN) models. The experimental findings indicates, a substantial gain in signal quality, and average SNR increase by 4.6 dB and classification rate by 9.2% over the traditional filtering and ICA approach. Hybrid VMD-ICA method provides effective maintenance of cognitive rhythm integrity and elimination of non-neural artifacts makes a strong and lightweight drowsiness detection with single-channel EEG. Architecture provides a viable future of intelligent transportation and human-machine interfaces for efficient neuroergonomic systems.