GIRISH MALLYA UDUPI

*** **** **** ******, #*** Phone no.: 312-***-****

Chicago, IL - 60616 Email: xavj9d@r.postjobfree.com

Objective

Looking for a challenging opportunity in the field of Signal Processing (Image/Speech/Sound) to apply my knowledge, skills and potential; to hone them and become an asset to the company.

Summary

• About 3 years of experience in image processing, DSP, and applied optimization of which 2 years are in research involving image processing, linear algebra, applied optimization and machine learning.

• Sound experience with using/programming in MATLAB for communication, signal processing, image processing, computer vision, sound processing, applied optimization and digital filter design.

• Thorough understanding of random signal processing and statistical signal processing techniques.

• Experience with working with images obtained from magnetic resonance imaging (MRI).

• Knowledge of multispectral MRI techniques such as T2-weighted MRI, DCE MRI, DWI MRI.

• Knowledge of B-splines and their application for image deformation.

• Good understanding of classification techniques such as discriminant analysis and support vector machines (SVM).

• Knowledge of dimensionality reduction methods like principal component analysis (PCA).

• Knowledge of programming with C/C++.

• Experience with programming with VHDL.

• Experience conducting brainstorm sessions and presenting ideas in a clear manner.

• Excellent communication and analytical skills.

• Potential to become a team performer and ability to learn new technology and systems quickly.

Education

• Master of Science (M.S.), Electrical Engineering, December 2010 (GPA 3.87/4.00)

Illinois Institute of Technology (IIT), Chicago, IL

Technical Skills

Programming Languages: MATLAB, C/C++, Python, VHDL, Intel 8085/8086, Intel 8051 Microcontroller

Professional Softwares: Adobe Photoshop, Audacity, Goldwave, MS Word/Excel/Powerpoint, Windows Movie Maker

Platforms: Windows 98/XP/Vista, Ubuntu

Experience

Research Assistant, Medical Imaging Research Center, IIT, Chicago, IL Jan 2009 – Dec 2010

Project: Image registration

Experience with using Matlab

Experience with using B-splines

Description: In medical imaging, it is often necessary for experts to use information from images obtained from different modalities. Typically, the images are fused with different weighting parameters to get a single image. Since the images would be from different times and different imaging systems, it is often necessary to register them before integration. We used a quadratic transformation to deform one image; then used conjugate gradient optimization method to minimize the squared error between the transformed image and the image it had to be registered with. This gave us the optimum transformation co-efficients.

Environment: Windows Vista, Matlab

Project: Localization of prostate tumor using multi-spectral magnetic resonance imaging (MRI)

Experience with using Matlab, along with the Image Processing and Statistics Toolboxes

Studied different MRI modalities like T2-weighted MRI, Dynamic-Contrast-Enhanced (DCE) MRI, Diffusion-Weighted-Imaging (DWI) and the estimation of different pharmacokinetic parameters from the images

Description: MRI has proved to be more accurate in localizing prostate tumors. But a single type of MRI is not sufficient for reliable tumor localization. Hence, use of multiple MRI-derived datasets has emerged as a promising alternative. Different MR images present different information and fusing these different information has proved to be useful. We investigated the role of texture in prostate cancer localization with multispectral MRI data.

Environment: Windows Vista, Matlab

Thesis: Prostate cancer localization using texture with multi-spectral MRI

Independently wrote codes for extraction of texture features from multispectral MRI data and for classification of pixels based on the features

Wrote and defended my Master’s thesis on the usefulness of using texture for prostate cancer localization under the guidance of Dr. Imam Samil Yetik

Experience with using Matlab, along with the Image Processing and Statistics Toolboxes

Experience with using LATEX for typesetting of documents

Description: Investigated the role of texture in prostate cancer localization with multispectral MRI. Texture features were extracted using Gray-level Co-occurrence Matrices and Local Binary Pattern Operators. Classification and segmentation using the features was carried out using linear discriminant analysis, and pixels were labeled tumor or non-tumor. Multispectral MRI data of twenty-two patients with biopsy-confirmed prostate cancer was used for evaluation purposes. Performance was evaluated using leave-one-out cross validation and statistical measures such as sensitivity, specificity, dice co-efficient and area under the receiver-operator characteristic curves were computed.

Environment: Windows Vista, Matlab

Tutor, Academic Resource Center, IIT, Chicago, IL March 2009 – April 2010

• Responsible for tutoring and helping electrical engineering undergraduate students with courses such as signals and systems, digital systems, circuit analysis

• Also helped students with Matlab

• Carried out brainstorm sessions with student groups

Project 1: Digital audio encoder/decoder

Used polyphase Discrete Fourier Transform filter banks in the encoder and decoder

Carried out quantization and inverse quantization

Independently designed and tested the whole encoder/decoder system

Description: The aim was to implement lossy compression of audio signals by variable encoding of different frequency bands based on the power in the band. In the encoder, the signal was first divided into sub-bands using a polyphase analysis Discrete Fourier Transform filter bank, and then the outputs of each band were quantized differently to get a compressed file. In the decoder, the compressed file was first inverse quantized and then fed to a polyphase synthesis Discrete Fourier Transform filter bank to estimate the original sound signal. A compression ratio of up to 3:1 was achieved without significant loss of sound quality.

Environment: Windows Vista, Matlab

Project 2: Linear Prediction

Implemented the Levinson-Durbin and the Schur algorithms for forward linear prediction

Independently designed and tested the predictor system

Description: The linear predictor was an implementation of the Levinson-Durbin and Schur algorithms for forward linear prediction. The predictor was tested using sample sound signals assuming some samples as unknown. The performance was evaluated by calculating minimum mean square errors.

Environment: Windows Vista, Matlab

Project 3: Detection of an object in a noisy scene

Derived a generalized likelihood ratio test to detect the presence of an object in a noisy scene

Independently implemented the test and detection mechanism

Description: The aim was to design a detector which detects objects in noisy images obtained using photon-counter-based imaging systems. A generalized likelihood ratio test was derived for this purpose using the maximum likelihood threshold. The presence/absence of the object was tested for every possible spot in the given image.

Environment: Windows Vista, Matlab

Project 4: Classification of bacteria

Derived a likelihood ratio test to classify bacteria into two types based on observations made under a microscope

Independently wrote the code to carry out the classification

Description: The aim was to construct a decision rule to classify two types of bacteria using brightness and cell diameter measurements obtained from a microscope. A maximum likelihood test was constructed and receiver-operator characteristic curves were used to evaluate performance.

Environment: Windows Vista, Matlab

Programming Analyst, Cognizant Technology Solutions, Bangalore, India Nov 2007 – July 2008

• Created and published reports using SQL Server Reporting Service (SSRS) for client.

• Wrote stored reporting procedures using SQL Server.

• Prepared the analysis document for development of the reports.

• Developed reports and prepared defect trackers for the same.

• Generated test cases for developed reports.

• Tested the frontend (UI) of the reports extensively and logged the defects.

• Fixed defects in reports and deployed them to the web.

• Coordinated the activities of a sub-team of four members (including myself) in the project team.

Other Projects:

Transistor-level simulation of a 16-bit Multiply-Accumulate (MAC) unit

• As part of the curriculum in my B.E. program, 8th semester, I, along with three others took up the project to simulate the working of the MAC unit

• The unit was designed block level using VHDL and tested for functional accuracy

• Then, using SPICE, every transistor of the unit was described and the unit was simulated.

Note: Reference will be provided upon request.

Contact this candidate