Avinash Sharma

To apply my academic experiences and interpersonal skills to build a research career in the field of Computer Vision, Robotics.

Computer Vision, Pattern Recognition, Robot Navigation, Image Processing, Machine Learning.

• Master of Science by Research
  Expecting to finish by June' 2008 from IIIT-H,Hyderabad, India, with CGPA: 8.98/10, Adviser Dr. Anoop Namboodiri. Area of specialization : Computer Vision and Pattern Recognition.


• Bachelor of Engineering
  Institute of Engineering & Technology, Devi Ahilya University Indore (May 2005) with 77.69% aggregate.


• Senior Secondary
  Government Bal Vinay Mandir, Indore, Madhya Pradesh (MP board 2001) 84.66% (PCM: 87.33%).


• Secondary
  Government Bal Vinay Mandir, Indore, Madhya Pradesh (MP board 1999) 81.80% (ScM: 93.50%).

         Three dimensional objects are characterized by their shape, which can be thought of as the variation in depth over the object, from a particular view point. These depth variations are lost during the process of imaging. Image based algorithms try to characterize the intensity variation in the image for recognition, thus neglecting most of the 3D information. On the other hand shape based algorithms recovers the 3D shape and directly use it for recognition, which is computational intensive and error prone. Our problem is to find a 2D recognition method, which also uses the shape information of the object shape, without explicitly recovering the 3D shape of the object. That is we want to avoid the high computational cost while achieving high accuracies. The method should be robust to view variation, occlusion and invariant to scale and position of the object. It should also handle the problems of specular surfaces and texture-less surfaces.
         We propose the use of structured lighting patterns, which we refer to as Projected Texture, for the purpose of object recognition. The depth variations of the object induces deformations in the projected texture, and these deformations encode the shape information. The primary idea is to view the deformation pattern as a characteristic property of the object and use it directly for classification instead of trying to recover the shape explicitly. To achieve this we need to use an appropriate projection pattern and derive features that summerciently characterize the deformations. We have derived features for three class of objects : (i) 3D objects with arbitrary pose. (ii) 3D objects with fixed pose and (iii) 3D textured surfaces. The above mentioned approaches have been implemented and tested on various datasets collected as well as available on the Internet. The analysis and comparative results demonstrate significant improvement over the existing approaches, in terms of accuracy and robustness.

• Projected Texture for Hand Geometry based Authentication, Proceedings of the CVPR Workshop on Biometrics, (CVPRW 2008), June 28, Anchorage, Alaska, USA. IEEE Computer Society 2008.
  Avinash Sharma, Nishant Shobhit, Anoop Namboodiri


• Projected Texture for Object Classification, to appear in Proceedings of the European Conference on Computer Vision (ECCV 2008), October 12-18, Marseille, France.
  Avinash Sharma, Anoop Namboodiri


• Object Category Recognition with Projected Texture, to appear in Proceedings of the Indian Conference on Vision, Graphics and Image Processing (ICVGIP 2008), December 16-19, Bhubaneswar, India.
  Avinash Sharma, Anoop Namboodiri


• Frequency Domain Visual Servoing using Planar Contours, to appear in Proceedings of the Indian Conference on Vision, Graphics and Image Processing (ICVGIP 2008), December 16-19, Bhubaneswar, India.
  Visesh Chari, Avinash Sharma, Anoop Namboodiri, C.V. Jawahar

• Attended International Joint Conference on Artificial Intelligence 2007 (IJCAI) as student volunteer.
• Attended Asian Conference on Computer Vision 2006 (ACCV) as student volunteer.
• Attended Indian Conference on Computer Vision, Graphics and Image Processing 2006 (ICVGIP).
• Attended Indo-Israel Workshop on Computer Vision 2008 (WCV).

• Selected and participated in Intel India Student Research Contest 2005.
• Participated in paper presentation at Computer Society of India Gaziabaad chapter on topic "web based Automated placement system".
• Stood II in paper presentation contest on "Digital Watermarking" Confluence 2k4, a National Level students technical symposium, NIT Raurkela.
• Stood II best final year project competition in Auroreole-2005 held at GEC Jabalpur 2005.
• Stood II in quiz competition in Auroreole-2005 held at GEC Jabalpur 2005.
• Qualified Semi-final round of C programming contest,Confluence 2k4,NIT Rourkela.

• Worked as a research assistant during my Masters, from May 2006 onwards, at the Center for Visual Information Technology, IIIT-H.
• Teaching Assistant for Engineering Mathematics in Fall 2006.

Programming Languages	C, C++, CUDA, Java, COBOL
Operating Systems	GNU/Linux, Windows 95/98/ME/2000/XP
Programming Environments	GNU/GCC, Microsoft Visual Studio (VC++), Matlab
Scripting Languages	Perl, Shell, Java Script
Internet	html, xml, CGI ,PHP, CSS stylesheets
Libraries	Qt, OpenCV, Lnknet, Visual Servoing Toolbox, FFMPEG
Databases Technologies	MySQL, Oracle 8i(SQL) ,IBM DB2 and MS Access
Web Technologies	J2EE (EJB's, JSP, Servlet),ASP, J2ME
Application Server	Macromedia JRun4, Weblogic
Development Environment	IBM WSAD, Visual Basic
Multimedia Tools	Macromedia Flash5
Software Engineering Tool	Rational Rose
Other Tools	MS Office, Latex

• IBM DB2 (Universal Database V8.1 Family) IBM Certified Database Associate.

Computational Geometry, Computer Vision , Data Compression, Pattern Recognition ,Digital Image Processing, Machine Learning, Multi Robotics, Advance Problem Solving,

Master's Project

Major Project


• 3D Object Recognition using Projected Texture : 3D object recognition application where deformation in projected pattern encodes the object shape. The following three types of object class were handled in proposed framework :
  
  • Category Recognition with generic pose of 3D object
  • 3D Texture Recognition
  • Recognition of Aligned Rigid objects
  Developed under guidance of Dr. Anoop Namboodiri.
  Webpage
• Fourier Domain based Visual Servoing : We demonstrate how Fourier analysis may be used to obtain straight camera paths in the Cartesian space, do path following and correspondence-less visual servoing. Most importantly, by introducing Fourier techniques, we set a framework into which robust Fourier based geometry processing algorithms may be incorporated to address the various issues in servoing. Work done in association with Mr. Visesh Chari under guidance of Dr. C.V. Jawahar and Dr. Anoop Namboodiri.
• Homography based solution for generation of reference plane at arbitrary depth : A reference plane generation algorithm which can be used for Fourier Transform Profilometry (FTP) for the purpose of 3D reconstruction, in association with Mr. Visesh Chari under guidance of Dr. C.V. Jawahar and Dr. Anoop Namboodiri.
• Body Pose Recognition : Developed a primary level system for Philips Research India which recognizes human body pose with simple motion based tactics. Color marker's were used to capture joints in human body and stick model was tted to simulate the human action. This work was done in association with Mr. Pratyush Bhatt under guidance of Dr. Anoop Namboodiri.

Minor Project


• Implementing Stereo Matching algorithm using graph cut : Implemented as part of computer vision course, implemented graph cut based stereo matching algorithm.
• Markov Localization on Amigobotes : Implemented as part of multi-robotics course, in association with Mr. Rakesh Goyal under guidance of Dr. Madhav Krishna.
• Multiple Robot Collision Avoidance on Amigobotes : Implemented as part of multi-robotics course, in association with Mr. Rakesh Goyal under guidance of Dr. Madhav Krishna.
• DIP Library function implementation on GPU using CUDA : Implemented Background Subtraction function for GPU based DIP Library using CUDA as computer vision course project in association with Ms. Prachi Agrawal under guidance of Dr. C.V. Jawahar.
• Color based person following for Max Robot : Implemented color based person following on MAX Rob as summer project under guidance of Dr. Madhav Krishna.
• Implemented Priority Search Tree for DNA sequence search : Implementation of PST for DNA sequence as course project for Computational Geometry in association with Mr. Naveen Tewari under guidance of Dr. Prosenjit Gupta.

Bachelor's Project

Major Project


• Telemedicine : Developed a prototype for a life saving system that enables the transfer of medical(ECG) data through mobile phone. At client side small hardware was designed which encodes the ECG signals into binary data and pass it to mobile with serial port. At server end an J2EE based interface was designed which automatically handles the incoming call, identify the user, register the data and display it to duty doctor. Project completed at Astute System Technologies, Indore.

Minor Project


• Course Management System : As part of Management Information System of IET DAVV, Course Management System was implemented using EJB.
• e-Placement System : Developed fully automated portal for Institute of Engineering and Technology, DAVV using J2EE.
• Indian Orchids Database : Developed vortal of Indian Orchids for easy retrieval of orchid fauna in India at School of Biotechnology, DAVV, Indore.
• Web Space Management : Application of Client Server Model. It helps the user to efficiently handle the account of Web Space & Domain Name of their clients. Developed for North South Technologies Private Limited, Indore.
• Time-Table Management Assistant : Developed Time Table Management Assistant interface in Visual Basic for Institute of Engineering and Technology, DAVV, Indore.

• Organized NAACH-MANIA . An Inter College Group Dance Competition at IET DAVV.

Reading Novel, Traveling, Social Service.

End of Curriculum Vitae