Shridhar Ravikumar email

University of Bath, United Kingdom

Topic: Enhancing performance driven facial animation using machine learning approaches
Master of Science in Computer Science, University of Southern California, USA May 2009
Relevant Coursework: 3D Graphics and Rendering, Computer Vision,

Analysis of Algorithms, Game Engine Development, Advanced Game Projects,

Animation and Simulation
Bachelor of Engineering in Computer Science, University of Pune, India July 2006

CURRENT RESEARCH PROJECTS 2013 – 2016
Markerless performance capture from monocular video

• 
  Automatic landmark detection and optical flow to solve for blendshape weights in 2D.
  
• 
  Second pass using regression to map from input video to blendshape weights.
  

• 
  Using latent variable models to learn the dynamics of an individual's facial motion.
  
• 
  Modifying the dynamics of a performance to simulate identity.
  

• 
  Unconstrained photometric stereo for reconstructing static geometry of faces.
  
• 
  Tracking geometry through a video sequence by optimizing for scene flow parameters.
  

• 
  Learning a 3D Morphable model of identity from multiple 3D scans of neutral faces.
  
• 
  Learning a mapping from difference in identity space to expression space to generate
  

• 
  Ravikumar S, Davidson C, Kit D, Campbell N, Benedetti L, Cosker D. Reading Between the Dots:
  

Graphics Interface (Victoria, Canada, 2016), (In Press).

• 
  Development of a system for optimizing and solving animation parameters for 3D facial
  

• 
  Development of a pipeline for automated personalized blendshape rig generation.
  

A test bed for integrating all Virtual Human related technologies that allow users and

developers to prototype their own virtual humans including speech recognition, vision,

NLP, behavioral-animation, speech synthesis, rendering engine.

A mixed-reality, interactive experience with multiple virtual humans in a controlled

environment with support for both visual perception of user actions using infrared cameras

and verbal inputs using speech recognition and processing.

A procedural animation system for generating runtime animations on skeleton models interfaced

with multiple rendering engines including Ogre3D, Unity3D, Gamebryo.
Software Engineer Intern - Sony Computer Entertainment America (3 months) 2008
Game: “NBA ’09 - The Inside” for the Play Station Portable (PSP)

Development of in-game achievement system for tracking NBA awards

Bug fixes in game before release


MSc PROJECTS (University of Southern California)
Advanced Game Project ‘SkyDie’ - Lead Engineer (http://shridharravikumar.com/videos/)

Developed a networked kite simulation game on the Ogre3D graphics engine using the Wii-mote

Developed custom physics modules including thread and kite physics
IPhone Game ‘AquariYum’ - Engineer (http://shridharravikumar.com/videos/)

• 
  Game play programming
  
• 
  Artificial Intelligence
  

C, C++, C#, Java, Python, MATLAB, XML, HTML, JavaScript, Perl, PHP, Shell-scripting, SQL

Maya, Visual Studio, Subversion, Perforce, Git, Microsoft Office

Linux, Windows, Mac OS X, Unix, iOS

OpenCV, DirectX, OpenGL, FaceFX, MS SAPI, Festival TTS, NumPy, SciPy, LibSVM

Ogre3D, Unity 3D, Gamebryo

• 
  Won the best speaker award at the University of Bath inter-departmental postgraduate research
  

• 
  Supervised a final year undergraduate student for his final year project – Generating a statistical