Sagnik De

Sagnik De

Machine Learning Engineer

Posts by Collection

portfolio

publications

A Novel Human Stress Level Detection Technique Using EEG

Published in IEEE NMITCON 2023, 2023

A majority-voting ensemble of SVM, KNN, and Naive Bayes classifiers detects four levels of mental stress from frontal-lobe EEG features with 93.85% accuracy.

Recommended citation: Konar, D., De, S., Mukherjee, P., & Halder Roy, A. (2023). "A Novel Human Stress Level Detection Technique Using EEG." IEEE NMITCON.
Download Paper

SLiTRANet: An EEG-Based Automated Diagnosis Framework for Major Depressive Disorder Monitoring Using a Novel LGCN and Transformer-Based Hybrid Deep Learning Approach

Published in IEEE Access, 2024

Recommended citation: S. De, A. Singh, V. Tiwari, H. Patel, G. N. Vivekananda and D. Singh Rajput, "SLiTRANet: An EEG-Based Automated Diagnosis Framework for Major Depressive Disorder Monitoring Using a Novel LGCN and Transformer-Based Hybrid Deep Learning Approach," IEEE Access, vol. 12, pp. 173109-173126, 2024
Download Paper

Maestro: A Robust Multi-Head Attention Enhanced CNN Architecture for Heat-Induced Stress Recognition Using EEG Signals

Published in IEEE CSITSS 2024, 2024

MAESTRO combines convolutional and multi-head attention blocks to classify heat-induced stress from EEG into Acute, Chronic, and Control categories with 98.88% accuracy.

Recommended citation: De, S., Pavuluri, S., Sayyad, A., & Gupta, A. K. (2024). "Maestro: A Robust Multi-Head Attention Enhanced CNN Architecture for Heat-Induced Stress Recognition Using EEG Signals." IEEE CSITSS.
Download Paper

ParViT: A modified Vision Transformer architecture for Parkinson’s Disease identification using EEG signals

Published in IEEE ICSSES 2024, 2024

ParViT applies STFT-based time-frequency features to a modified Vision Transformer for Parkinson’s Disease identification from EEG, reaching 98.25% accuracy.

Recommended citation: De, S., Sayyad, A., Kotian, H., & Gupta, A. K. (2024). "ParViT: A modified Vision Transformer architecture for Parkinson's Disease identification using EEG signals." IEEE ICSSES.
Download Paper

A Quantum Machine Learning framework for Driver Drowsiness Detection using Biopotential Signals and Head Movement Analysis

Published in IEEE ICWITE 2024, 2024

A CNN + Attention-based Quantum LSTM hybrid model that detects driver drowsiness levels from EEG, facial EMG, pulse rate, and head movement with up to 99% accuracy.

Recommended citation: De, S., & Gupta, A. K. (2024). "A Quantum Machine Learning framework for Driver Drowsiness Detection using Biopotential Signals and Head Movement Analysis." IEEE ICWITE.
Download Paper

talks

teaching