HTML Preview Computer Engineer page number 1.
Saurabh Gupta
Senior Year B.Tech Student +91 9911592327
Indian Institute of Technology Delhi http://www.cse.iitd.ernet.in/
~
cs1070185
Academic Details
Year Degree Institute Percentage/CGPA
2007- BTech in Computer Science Indian Institute of Technology CGPA = 9.938/10
Present and Engineering Delhi Institute Rank 1
2007
Class XII Bal Bharati Public School 90.2%
CBSE-AISSCE Rohini, Delhi 99% in Computer Science
2005
Class X Bal Bharati Public School 93.4%
CBSE-AISSE Rohini, Delhi 100% in Mathematics
Objective
To pursue graduate studies in computer science and engineering, leading to a career in research. I am
interested in computer vision and machine learning.
Major Projects
•
Improving Performance with MKL in SVMs Dr. Manik Varma, MSR India
Dr. Prateek Jain, MSR India
Summer Internship and B.Tech Project May 2010 - Present
Support Vector Machines learn linear classifiers from labelled training data to predict labels for previ-
ously unseen test data. Kernels with SVM’s allow to learn non linear classifiers, and hence lead to much
higher classification accuracies in practice. We are learning non linear kernel combinations to give
us even better classification accuracies and are seeing a possibility of a significant performance increase.
Also, looking at new regularizers (which take into account the model complexity) for use in the MKL
objective. Working towards submitting this to one of ICML, NIPS or ECCV.
•
Unsupervised Video Surveillance Prof. Subhashis Banerjee, CSE, IITD
Student Undergraduate Research Award (SURA - 2009) May 2009 - October 2009
This project involved the development of a novel technique to detect unusual activities in videos in an
unsupervised manner. We used pLSA with an appropriate choice for features on video clips to obtain
clustering of similar videos, and any new activity different from the ones learnt was flagged as unusual.
We also extended the idea to multi person unusual activity detection.
Other Projects
•
Captcha Reader Prof. Prem Kalra
Digital Image Analysis Course Project November 2010
I made a simple captcha reader to read the captcha in use for the institute web mail. It involved letter
segmentation and training in a semi supervised manner using a KNN classifier.
•
Using Structural Information for Object Recognition Prof. Subhashis Banerjee
Computer Vision Course Project April 2010
I investigated various feature vectors and techniques and frameworks for incorporating structural infor-
mation in object recognition tasks.
•
Non Photo-realistic Rendering Prof. Prem Kalra
Digital Image Analysis Course Project September 2010
Implemented color quantization, dithering, image segmentation and edge enhancement (by edge linking,
differential thickening and smoothing) to give a painterly look to photo-realistic images.
•
Knowledge Management Prof. Subhashis Banerjee
Independent Project January 2010 - October 2010
Conceptualized new web 2.0 based IT services such as Interest Group Discussion Forums, Campus