Medical Electrical Engineering

Aditya Gupta

***, ******* ******, ************, **-27278

407-***-****. ************@*****.***

SUMMARY

Self motivated professional with a strong technical background in image processing, machine learning

and image analysis. Expertise in analyzing and solving complex problems with demonstrated success in

several projects. Strong experience in writing technical papers, reports and research proposals.

Areas of Expertise

Image processing, Machine learning, data classification methods, pattern recognition, writing proposals

and grants, algorithm and tool development in languages C++ (with ITK, VTK, Qt, OpenCV), MATLAB,

VHDL, Verilog.

COMPUTER SKILLS

Tools: MATLAB 2011, R, Mathcad 14, Microsoft Office 2010, PSPICE, Xilinx ISE 9.2i and Ansoft

Maxwell 2D.

Languages: C, C++, VC++, ITK, VTK, ImageJ, OpenCV, Latex, VHDL, Verilog, Assembly

Languages, Visual Basic.

EDUCATION

• University of Central Florida, Orlando, FL

Doctor of Philosophy Ph. D. in Electrical Engineering GPA: 3.8. Dissertation Title – “Signal

Processing of an ECG signal in the presence of a strong static magnetic field (MRI).”

• University of Central Florida, Orlando, FL

Master of Science M.S. in Electrical Engineering GPA: 3.9. Specialization: Digital Signal & Image

Processing, Embedded Systems.

• Visvesvaraya Technological University, India

Bachelor of Engineering B. E. GPA: 3.9. Electronics and Communication Engineering.

RELEVANT EXPERIENCE

Postdoctoral Research Scholar July 2010-Present

Neuro Image Research and Analysis Laboratories, University of North Carolina & Department of

Pediatrics, University of Pittsburgh Medical Center

• For the neurodegenerative Krabbe disease it is extremely crucial to develop a biomarker that would

point the narrow window of time for successful treatment of the patients. Developed algorithms and

tools using C++ with ITK and VTK to perform a white matter tract based analysis on diffusion tensor

MR images. The pipeline includes unbiased diffeomorphic atlas building, DTI tensor registration,

tensor normalization, log-euclidean tensor resampling, statistics and interpretation of DTI parameters.

The developed pipeline now provides crucial information to the pediatric doctors to help determine

the crucial time of treatment and as a marker for disease progression.

• The white matter fiber tracts of brain of patients with hydrocephalus are highly deformed because of the

enlarged lateral ventricles and hence generally not analyzed. Developed a pipeline based on creating

images from full brain tractography, correspondence using SIFT features, Gaussian radial basis function

and diffeomorphic Demons registration. The developed pipeline now provides a potential method of

analysis of the enlarged lateral ventricle cases, with an accuracy of 23% higher than currently

available registration methods.

• Scalar image normalization is a commonly used method prior to registration. Developed a method for

normalization of 3D tensors based on mapping the tensor information to a cumulative distribution

function space. The developed tensor normalization method improves the accuracy of tensor based

registration methods by 9%.

• Developed ideas and preliminary results for novel research projects - analysis of the shape of lateral

ventricles to distinguish between hydrocephalus and atrophy, use of resting state functional connectivity

as a biomarker for white matter demyelinating diseases in addition to DTI, and analysis of DTI for

Turner’s disease. These ideas were translated to research proposals and collaborative work.

Research Associate / Post Doc May 2008-December 2008 & August 2009-May 2010

Computer Vision Lab, University of Central Florida

• Cardiac MRI: Project in collaboration with Orlando Regional Medical Center and Northwestern

University. The infarct and peri-infarct zones are found to be independent predictors of post myocardial

infarction. Developed a fully automatic method to determine the mass and volume of the infarct and

the peri-infarct tissue in the left ventricular wall of the heart from cine MR images. The method uses a

combination of 3D active appearance model and level sets using shape priors to determine the

ventricular walls. Features of intensity, volume, shape and heart wall thickness are used to determine the

infarct and peri-infarct zones. The method achieves a high correlation co-efficient of 91% as

compared to the results with manual tracings.

• Detection and tracking of different cell types when neural stem cells differentiate in 2-D phase

contrast microscopy. Project in collaboration with Bio-molecular Science Center, Burnett College of

Biomedical Sciences, UCF. The detection of the cell types was achieved using methods based on k-

means clustering, level-set algorithm and background subtraction. Cell recognition was achieved using

a SVM with a 32-element feature vector. Detection rate of 78% and a recognition rate of 80.5%

were achieved.

• Brain MRI: Project in collaboration with Florida Hospital. Development of an objective,

computer aided diagnostic (CAD) methodology for automatic computation of brain tumor volume

relevant to clinical decision making. Part of the team that was awarded $400,000 from NIH for a two-

year period.

Doctoral Thesis

• With the extensive use of MRI with higher magnetic field strengths it is important to understand

the effect of these fields on the human body. Developed software package to simulate the effect of the

magnetic field on blood flow in the blood vessels, particularly the aorta. The package outputs the

interaction of magnetic field with blood flow and the effect on the patient’s ECG in MRI environment.

The package involved developing 3D model of the aorta and thorax, volume conductor, computing the

potential induced using magneto fluid dynamics, surface maps and algorithm to transfer the potential

induced on the aorta to a model of the thorax. The method also involved developing methods to

determine cardiac output non-invasively using a pulse oximeter. The package quantitatively explains

the observed elevation of the T wave in the ECG signal and filters this artifact with average accuracy of

74%.

• Developed methods based on wavelet transform and adaptive filters to successfully filter the T wave

artifact with an average accuracy of 83%.

• Developed Continuous Non invasive blood pressure simulator using FlexiForce® pressure sensors,

tubes, data acquisition card and front end circuitry. Also developed methods of filtering motion artifacts

from signals – comb filter, wavelet transform, and adaptive wavelet transform acquired from the proof

of concept portable blood pressure monitor.

ADDITIONAL EXPERIENCE

• Developed the digital circuits’ laboratory on FPGA using Xilinx ISE 9.2i. Project involved training

the lab instructors on VERILOG programming and the procedures involved in the laboratory.

• Developed a GUI software code in VC++ to interpret the output of the frame synchronizer to display

satellite GPS parameters for Indian Space Research Organization. Successfully accomplished the task

of miniaturization of frame synchronizer into ASIC.

• Conducted corporate trainings and taught various courses in biomedical image and signal

processing, pattern recognition, biomedical devices and embedded systems.

Peer Reviewed Publications and Proposals

• Aditya Gupta, Arthur R Weeks, Samuel M Richie – “Simulation of elevated T-waves of an ECG inside

a static magnetic field (MRI).” IEEE Trans Biomed Eng. 2008 July, 55 (7):1890-6.

• Aditya Gupta, Mubarak Shah – “Segmentation of infarct and peri-infarct zones in cardiac MR

images.” In Proceedings of MIAR 2010, pp 31-41.

• Yi Wang, Aditya Gupta, Zhexing Liu, Hui Zhang, Maria L. Escolar, John H. Gilmore, Sylvain Gouttard,

Pierre Fillard, Eric Maltbie, Guido Gerig and Martin Styner – “DTI registration in atlas based fiber

analysis of infantile Krabbe disease.” NeuroImage Volume 55, Issue 4, 15 April 2011, Pages 1577-

1586.

• Aditya Gupta, Maria L. Escolar, Cheryl Dietrich, John H. Gilmore, Guido Gerig, Martin Styner – “3D

Tensor Normalization for Improved Accuracy in DTI Tensor Registration Methods”. 5 th International

Workshop on Biomedical Image Registration 2012, LNCS 7359, pg 170-179.

• Aditya Gupta, Matthew Toews, Ravikiran Janardhana, Yogesh Rathi, John Gilmore, Maria Escolar,

Martin Styner – “Fiber feature map based landmark initialization for highly deformable DTI

registration.” SPIE Medical Imaging 2013.

• Audrey Verde, John-Baptiste Berger, Aditya Gupta, Adroan Kaiser, V W Chanon, C A Boettiger, Casey

Goodlett, Yundi Shi, Guido Gerig – “Atlas Based Fiber Tract Analysis with Application to a Study of

Nicotine Smoking Addiction.” Proceedings of SPIE Medical Imaging 2013.

• Xiujuan Geng, Martin Styner, Aditya Gupta, Dinggang Shen, John H Gilmore, “Multi-contrast

Diffusion Tensor Image Registration With Structural MRI”, IEEE International Symposium on

Biomedical Imaging, 2012.

• Anuja Sharma, P.T. Fletcher, John Gilmore, Maria Escolar, Aditya Gupta, Martin Styner, Guido Gerig -

“Spatiotemporal Modeling of Discrete-Time Distribution-Valued Data Applied to DTI Tract Evolution

in Infant Neurodevelopment,” In IEEE Proceedings of ISBI 2013, pp. (accepted). 2013.

• Mahshid Farzinfar, Cheryl Dietrich, Rachel Smith, Yinpeng Li, Aditya Gupta, Zhexing Liu, Martin

Styner – “Entropy based dti quality control via regional orientation distribution”, IEEE International

Symposium on Biomedical Imaging, 2012.

• NIH proposal: “Computer Assisted Identification and Volumetric Analysis of Enhancing Components.”

The goal of this research is to develop a more accurate and reproducible way to measure the amount of

disease present in patients suffering from brain tumors. Submitted from the Computer Vision Lab, UCF

(PI: Dr. Mubarak Shah) in collaboration with Florida Hospital.

Status: Awarded for a two year period for $400,000 under US Recovery Act in June 2009

• NIH proposal: “Automated live neural stem cell type identification and differentiation analysis.”

The goal of this research is to develop an automated system capable of glial cell identification and

recognition, glial cell differentiation analysis and neuron detection and tracking. Submitted from the

Computer Vision Lab, UCF with Dr Sugaya, Bio-molecular Science Center, Burnett College of

Biomedical Sciences, UCF as the co-PI.

• NIH proposal: “Computer Assisted Analysis of Infarct and Peri-Infarct Areas.” Based on the

development of an objective, computer aided diagnostic (CAD) methodology for automatic

computation of the mass and the volume of the infarct and the peri-infarct tissue relevant for clinical

decision making. Submitted from the Computer Vision Lab, UCF (PI: Dr. Mubarak Shah) in

collaboration with the Orlando Regional Medical Center and Northwestern University.

Contact this candidate