DISSERTATION
Development of an Emotional Scene Classification System using Spatial Frequency
Master’s Thesis — Advisor: Prof. Narayanan Srinivasan June 2015 – May 2016
Investigated the role of spatial frequency in classifying affective scenes and classified the images that belong to three categories of affectiveness (pleasant, unpleasant, and neutral). Used IAPS -International Affective Picture System database.
Genetic Algorithm-Based Mental Stress Assessment from HRV Analysis
Bachelor’s Thesis — Advisor: Prof. G.Ranganathan Sep 2012 – April 2013
Devised a method to assess mental stress using HRV analysis. Used fuzzy C-Means clustering to handle the uncertainties and employed a genetic algorithm for optimization. Used MIT-BIH Arrhythmia database and NASA Task Load Index.
COURSE PROJECTS & PRESENTATIONS
Perception:
Threshold determination task using 2 AFC paradigm
Reaction-time measurement task using cueing paradigm
Sound wave generation and frequency discrimination – pure tone and harmonics
Principles of Neuroscience:
Long-Term Potentiation – Physiological evidence of the Hebbian learning
Computational models of cognition:
Choosing a model
Cognitive Neuroscience & Neuropsychology:
Mechanisms of face perception
Psycholinguistics:
Evolution of language - summarization
Computational models of grammatical development
Computational Neuroscience:
Neural Decoding
Action:
Central Pattern Generators
TERM PAPERS
Attention:
Is Consciousness a module or an Integral part of Attention?
Culture & Cognition:
Cultural Variations in Scene Perception
Philosophy of Mind:
Functionalism: Framework of Mind
OTHER RESEARCH PROJECTS:
Assessing Agency Measurement and Temporal Changes: A Cross-Country Analysis of India and Vietnam
Young Lives Longitudinal Study | Presented to: Prof. Antje Von Suchodoletz, NYU Abu Dhabi Nov 2022
Conducted exploratory data analysis and Confirmatory Factor Analyses to examine the factor structure of the agency measure. Analyzed changes in the agency over time and visualized patterns of change in India and Vietnam.
CNN with Primary Visual Cortex front-end: Activations in Clean and Perturbed Images
Neuromatch Academy Project Aug 2021
Examined the robustness of the VoneBlock (a hybrid neural network that shares the properties of primary visual cortex V1) in handling perturbed images and performed Representational Similarity Analysis (RSA).
Correspondence between CNN and Human Visual Areas for Categorical Representation
Neuromatch Academy Project — Mentor: Prof. Jing Xu, University of Georgia July 2021
Evaluated the effectiveness of CNN-AlexNet in reflecting the fMRI activity in the visual cortex (V1-LOC) for natural and artificial scene stimuli. Analyzed the differences using Representational Dissimilarity Matrices (RDM) and permutation testing.
fMRI Data Analysis of auditory perception task with bi-syllabic words April 2016
Brain Imaging Course Project | Supervisor: Prof. Chandrasekhar Pammi
Detection, Estimation, and Brain Connectivity – fMRI model setup, parameter estimation, and inference in SPM 12, including construction of the design matrix in the GLM, linear and non-linear methods to detect the brain activity maps, estimation of the hemodynamic response (HDR), functional connectivity of the brain structures by using nonlinear cross-correlation analysis, and effective connectivity by Granger causality approach.
Integration of HRV, Wavelet transforms & Neural Network for ECG Arrhythmias Classification
Supervisor: Prof. G. Ranganathan July –Nov 2012
Developed a cardiac arrhythmia classifier based on the statistical analysis of the ECG signals of normal and abnormal heart rhythm. Extracted features for classifying arrhythmias from the wavelet decomposition technique. Compared the classifier using multilayer perceptron (MLP) and radial basis function (RBF) neural networks.
Development of E-Passport scheme using Fingerprint Biometrics Mar-Apr 2013
Developed a biometric e-passport system with a fingerprint scanner using embedded technology to match individual fingerprints with the database for more accurate verification and identification of an individual. Designed using micro-controller (PIC), LCD, matrix keyboard, EEPROM, USART, and Visual Basic 6.0
Automated Line Tracer Robot using ATMEGA 8 microcontroller Sep 2011
Developed an automated machine, capable of traversing a black curve on a white background, by interfacing sensors and power actuators to a microcontroller.