Peshala Pahalawatta Staff Engineer, Image Technology Dolby Laboratories Inc.
Experience
Dolby Laboratories Inc—Burbank, CA Staff Engineer, Image Technology October 2007 - Present Research and development in image and video technology. Investigate advanced techniques for obtaining objective video quality metrics to evaluate video compressed under varying encoding conditions. Investigate advanced video compression techniques. Investigate error resilient video encoding techniques, and decoder error concealment techniques. Investigate next generation 3D video coding and delivery schemes. Write internal and external documentation, and patent applications, related to research and development. Give internal and external presentations on technology innovations. Hitachi America Ltd—San Jose, CA Summer Intern: Wireless Systems Research Laboratory May 2007 – September 2007 Researching QoS techniques for HDTV video streaming over WLANs. Implementing forward error correction (FEC) and error concealment techniques to minimize errors due to packet losses. Implemented algorithms for real-time experimentation and demonstrations in existing 802.11 testbed at WSRL. Gained experience with MPEG-2 transport streams, RTP/UDP, multithreaded programming in C, and network programming. Gained familiarity with 802.11 MAC/PHY including QoS extensions in 802.11e. Showed significant improvement in packet error rates and video quality can be achieved using proposed techniques. Northwestern University—Evanston, IL Research Assistant: Image and Video Processing Laboratory September 2004 – May 2007 Collaborated with Motorola Network Advanced Technology Group in Arlington Heights, IL, on project related to multiuser video streaming over emerging wireless networks such as HSDPA, IEEE 802.16. Developed content-aware packet scheduling techniques for streaming multiple pre-encoded video streams on-demand to multiple clients over a wireless network. Devised a practical scheme that uses the gradients of a content-dependent utility function to determine the scheduling and resource allocation across multiple users. Investigated implementation issues, such as, error robust encoding strategies, buffer delay, and packet fragmentation at the MAC layer. Explored the effect of decoder-side error concealment of lost packets on QoS of video. Showed that significant improvement in video quality (in terms of first and second order QoS statistics) can be achieved using content-aware scheduling technique. Explored the use of objective perceptually relevant video quality metrics in addition to the commonly used PSNR metric. Extended the algorithm to scalable video codecs. Implemented algorithms within the H.264/AVC and H.264 Scalable Video Coding extension reference codecs using C/C++. Northwestern University—Evanston, IL Research Assistant: Image and Video Processing Laboratory September 2001–September 2004 Collaborated with Motorola Labs in Schaumburg, IL on project related to object tracking using multiple imaging sensors. Implemented real-time background segmentation demo in C using USB camera on Linux PC. Used spatio-temporal Markov Random Field method to obtain smooth background segmentations for moving object detection. Designed a multi-scale feature detection and extraction algorithm, to be used for wide baseline stereo camera calibration. Developed an algorithm that used a non-linear Kalman filter tracker for optimal sensor selection as the tracked object moved through the sensor field. Lafayette College—Easton, PA Honors Thesis Research: Electrical Engineering Department August 1999–May 2000 Developed a web-based handwritten character recognition system. Tested the performance of various algorithms including Hopfield neural networks, Hough transforms for line detection, and vector quantization algorithms. Performed image scaling and rotation to obtain affine invariant templates for recognition. Implemented the system to operate as a web-based application.
Education
Northwestern University—Evanston, IL Ph.D., Electrical Engineering, December 2007 Thesis: Downlink Packet Scheduling and Resource Allocation for Multiuser Video Transmission over Wireless Networks Cabell Fellowship awarded for first year of graduate study Northwestern University—Evanston, IL M.Sc., Electrical Engineering, December 2002 Thesis: Collaborative Video-Based Target Tracking Lafayette College—Easton, PA B.Sc., Honors in Electrical Engineering, Minor in Computer Science, May 2000 Thesis: A Web-Based Handwritten Character Recognition System Elected to Tau Beta Pi, Eta Kappa Nu, and Phi Beta Kappa honor societies (1999)
Publications
HOME