Computer Vision, Pattern Recognition

RISHABH BARDIA

***, **** **** ****, **********, 15219, PA

Phone: 206-***-****, E-Mail: *********@*****.***

Interests

My main interests are application of machine learning techniques on image and neural data analysis.

Education

Carnegie Mellon University

Masters of Science in Biomedical Engineering, August 2013 – December 2014

Overall QPA 3.95/4.00

Courses in Pattern Recognition, Neural Signal Processing, Cognitive Videos, Computational Methods and

Computer Vision.

Vellore Institute of Technology

Bachelor of Technology in Biomedical Engineering July 2009 - July 2013

Overall GPA 3.28/4.00

Courses in Digital Signal Processing, Medical Image Analysis, Artificial Neural Networks.

Academic Projects

Soft biometrics:

Nose images for person identification when ocular region or mouth is occluded.

Implemented histogram of oriented gradients for feature extraction and trained a support vector machine for

classification.

The verification rate at 1% false accepts rate was 0.70. The database used was collected from a people

enrolled in Pattern Recognition Fall 13.

Face detection:

Implemented a cascade of classifiers based on correlation with a face template, skin color Gaussian model

and local binary patterns features classified with support vector machines.

The classifier was trained on half of the FDDB dataset and tested on the other half which gave on an

average the F score of 0.40.

Cancer cell classification:

Extracted texture features from liver cells (2 classes) and thyroid cells (3 classes) using Gabor filters having

8 orientations and 5 scales and trained a mixture of Gaussian model for classification.

Work Experience

Mitsubishi Electric, Ofuna, Kanagawa, Japan. 5/28/2014 – 8/5/2014

Face detection and facial expression analysis on uncontrolled data using convolution neural networks.

CyLab Biometrics Center, Pittsburgh, PA, USA. 1/15/2014 – 5/2/2014

A feasibility study of using a nose image as a biometric using Advanced Correlation Filters with Dr. Marios

Savvides. We show that it is comparable to face recognition on a subset of FRGC database

Batra Hospital and Medical Research Center, New Delhi, India. 6/10/2012 – 7/9/2012

Brain tumor segmentation on MRI images.

Skills

MATLAB: Image processing, application of machine learning for object detection and recognition.

C++: Open CV for computer vision applications.

Python: Image processing.

Contact this candidate