Research & Publications
17 peer-reviewed papers by Vanshaj Awasthi across IEEE and Springer, spanning medical AI diagnostics, deep learning, neuroimaging, brain–computer interfaces, and bioelectromagnetics. Two IEEE Best Paper Awards.
All publications
- Deciphering Molecular Subtypes in Uterine Corpus Endometrial Carcinoma
IEEE · Bioinformatics / ML
Leveraging machine learning to uncover distinct molecular signatures and histological markers for personalized therapeutic strategies.
- Interpretable Deep Learning for Knee Osteoporosis Screening
IEEE · Medical AI / Deep Learning
Employs ResNet-18 with feature attention for accurate osteoporosis detection from knee X-rays with clinical interpretability.
- A Disaster Prediction Ensemble Classifier Using Artificial Intelligence
IEEE (MoSICom 2024) · Crisis Management / AI
Introduces the Ensemble B-Classifier Network (DE BCN) for improved disaster forecasting, leveraging multi-model integration and Mustard Twin Swarm…
- GenAI-Enhanced Brain-Computer Interface for Assistive Communication
IEEE (ICEIL 2024) · BCI / Generative AI
Proposes a GAN-augmented BCI framework that breaks communication barriers for individuals with physical disabilities, achieving a 20% improvement in…
- Automated Alzheimer's Disease Detection via 3D ResNet and Hippocampal Atrophy Mapping
IEEE · Neuroimaging / Deep Learning
A 3D-ResNet backbone with multi-scale feature fusion achieves 97.31% classification accuracy in early-stage Alzheimer's detection from high-resolution…
- CLAHE-Augmented MRI Dementia Classification via Soft-Voting Ensemble with Gradient-Based Analysis
Springer Nature · Medical Imaging / Ensemble Learning
Soft-voting ensemble (ResNet-50, DenseNet-121, ConvNeXt, Swin Transformer, ViT) with CLAHE preprocessing achieves AUC-ROC 0.99/0.97 across 87,000+ MRI…
- AI-Optimized Wireless Power Transfer Systems for Implantable Medical Devices: Enhancing Energy Efficiency in Parkinson's Neurostimulation
Springer Nature · Biomedical Engineering / WPT
AI-driven WPT framework achieving >85% energy transfer efficiency, 250 mW stable delivery at 10 mm, ≤1.5°C temperature rise, and 1,200+ hours in-vitro…
- Exploring the Therapeutic Potential of Electromagnetic Field Exposure from WPT Systems in Neural Regeneration Post-Spinal Cord Injury
Springer (2nd Symposium on Smart, Sustainable, and Secure IoT) · Neural Regeneration / EMF Therapy
Controlled EMF exposure from WPT demonstrates 40% improvement in movement performance vs. untreated controls, with enhanced axonal growth, myelin repair…
- Attention-Augmented MobileNetV2 for MRI-Based Brain Tumor Classification: Cosine Annealing and Advanced Metrics
Springer LNNS (ICACIT 2024) · Neuro-Oncology / Deep Learning
Attention mechanisms integrated into MobileNetV2 with cosine annealing and stratified k-fold validation for glioma, meningioma, and pituitary tumor…
- Synergizing GAN-Driven Synthetic Data Pipelines with Deep Neural Networks for Enhanced Breast Cancer Diagnosis
Springer LNNS (ICACIT 2024) · Oncology / Generative AI
GAN-generated synthetic histopathology images augment CNN training to achieve 99.17% accuracy, 100% recall, MCC 0.9801, and ROC-AUC 0.9998 for benign vs…
- Priority-Driven MAC Protocol Design for IoT-Enabled Wireless Body Area Networks
Springer LNEE (FLAME 2024) · IoT / Wireless Networking
Priority-aware MAC protocol for WBANs reduces signal collisions, cuts energy consumption, and improves reliability for real-time IoT health monitoring…
- Reinforcement Learning and Spatio-Temporal GNNs for Alzheimer's Disease Progression Prediction with VAE-Based Data Imputation
IEEE ICETET-SIP 25 · Neuroimaging / GNN — 🏆 Best Paper Award
Spatio-temporal GNNs + reinforcement learning + VAE-based imputation for Alzheimer's progression modeling across 2,149 ADNI MRI scans. 98.2% accuracy…
- Hybrid RBC Morphology Analysis and Diagnostic Framework for β-Thalassemia Using SEBlock-CBAM Enhanced MobileNetV2 and TabNet
IEEE ICCSAI 2025 · Hematology / Medical AI
Hybrid diagnostic pipeline for β-thalassemia: attention-augmented MobileNetV2 for RBC morphology + TabNet for tabular blood data, with Optuna and…
- Hybrid 3D CNN and ResNet Deep Transfer Learning for High-Resolution Hippocampal Atrophy Mapping and Automated Alzheimer's MRI Diagnosis
ETASR – Engineering, Technology & Applied Science Research (Scopus Q2) · Neuroimaging / Transfer Learning
3D U-Net hippocampal segmentation fused with multi-scale ResNet achieves 97.31% accuracy and 92.84% DSC. Validated on ADNI + OASIS datasets with Grad-CAM…
- Neuro-Oncology Reimagined: Tailored Prognosis for Brain Tumors Using Adaptive Machine Learning
IEEE · Neuro-Oncology / Precision Medicine
Adaptive ML with Random Forest, SMOTE, and algorithmic fairness achieves 98.36% brain tumor prognosis accuracy, integrating multi-omics data with bias…
- Survey on Asclepius: A User-Centric Chatbot Bridging the Gap in Healthcare Integrated with VPN Technology
IEEE AECE 2024 · Healthcare AI / NLP
Asclepius: an AI healthcare chatbot achieving 89% diagnostic accuracy via supervised decision trees, VPN-secured communications, and facial recognition…
- Analysis and Impact of Electromagnetic Field Leakage in Wireless Power Transfer on the Central Nervous System: A Neural Network Approach
IEEE Sustainable Energy & Future Electric Transportation 2024 · Bioelectromagnetics / Safety — 🏆 Best Paper Award
Neural network modeling of EMF absorption from WPT systems on CNS tissue, informing bioelectromagnetics safety guidelines. Best Paper Award at IEEE…