Parag Havaldar

Parag Havaldar

Software Architect – Sony Imageworks


Many recent movies have been shot in stereo (termed native stereo) using two cameras. Using two cameras on set rarely gives stereo “out of the box” as expected and further processing is often required to correct the images. Additionally, augmenting computer generated imagery and visual effects into native stereo images have significantly more stringent requirements across all departments in a traditional post production pipeline. This talk will survey the common problems and possible solutions when working with native stereo imagery that includes the addition of computer generated elements. Case studies include Oz the Great and Powerful and The Amazing Spider-Man.


Parag Havaldar, software architect at Sony Pictures Imageworks, leads the research and development efforts for stereoscopic imagery and motion capture. He has led the stereo group to help design the facilities first native stereo pipeline which was used on recent projects such as OZ THE GREAT AND POWERFUL and THE AMAZING SPIDER-MAN.

Previously Havaldar led the performance capture group, where he developed technologies using computer vision facial action coding system to successfully animate digital faces from markerless video data and traditional motion capture data. The system was used on Imageworks’ performance capture films, GREEN LANTERN, WATCHMEN, ALICE IN WONDERLAND, BEOWULF, SPIDER-MAN 3 and MONSTER HOUSE.

Havaldar first joined Sony Pictures Imageworks in 2003 as senior software engineer where he helped to write deformers used in rigging set-ups, create custom Inverse Kinematic solutions for animation and develop other software tools around the modeling, rigging, animation, and rendering pipelines.

Prior to his work at Imageworks, Havaldar was a consultant in MPEG4 as a compression and systems expert, where he was part of the first commercialization of the open standard. He has also worked extensively in computer games to develop game tools used on the Nintendo and Sega platforms. His earlier work on computer vision technologies began at the Institute of Robotics and Intelligent Systems at the University of Southern California.

Currently, Havaldar is also a part time professor at the University of Southern California in the Computer Science department. He authored a graduate level textbook in the Multimedia and Graphics area and has written numerous technical papers and presented at many international conferences. Havaldar has also been awarded many patents for instereoscopic imagery and performance capture.

Havaldar holds a BS in Computer Science and Engineering from the Indian Institute of Technology as well as an MS and PhD in computer Science from the University of Southern California with a specialization in computer graphics and computer vision.