Professional Information

Contact:

Home: 21743 Castleton St., Cupertino, CA 95014.
Tel: (Work) 650-253-0775; (Home) 408-245-2456; (Cell) 408-234-5956

Occupation:

  • Senior Staff Software Engineer, Google (Chrome), Mountain View CA: Jan 10 - current
  • Senior Research Scientist, HP Laboratories, Palo Alto, CA: Aug 99 - Jan 10
  • Software Engineer, IBM Global Services, Bangalore, India: July 93 - Dec 94

Education:

  • Ph.D., ECE, University of California, Santa Barbara: July 1999
  • M.S., ECE, University of California, Santa Barbara: Dec 1995
  • B.Tech. (Hons.), E & ECE, Indian Institute of Technology, Kharagpur, India: June 1993.

Skills/Interests:

  • Video and Image Coding
  • Video, Image and Signal Processing
  • Information Theory

Awards and Honors:

  • Senior Member of IEEE
  • Co-author of top 10% paper award in Multimedia Signal Proc. Workshop, Rio, Brazil, Oct 2009.
  • Recipient of the IEEE Student Paper Award at the IEEE International Conference on Image Processing, Chicago, Illinois, October 1998.
  • Recipient of the Int. Symposium on Microwaves & Communications Award over the period 1992-93. 
  • Recipient of the National Talent Search Scholarship from NCERT (India), over the period 1987-92. 
  • All-India 1st rank in Math and Physics in ISC (Higher secondary school exam – 'A' level equivalent, 1989); All-India 10th rank in ICSE (Secondary school exam – 'O' level equivalent, 1987).

Recent Professional Activities:

  • Associate Editor of IEEE Transactions on Circuits and Systems for Video Technology (2016 -)
  • Associate Editor of IEEE Transactions on Image Processing (2012-15)
  • Associate Editor of SPIE Journal of Electronic Imaging (2012-2015)
  • Area Co-chair: Several ICIP conferences
  • Industry Co-chair: ICIP 2017, Beijing.
  • Technical program co-chair IMMERSCOM 09, Program Committee Member IMMERSCOM 07
  • Co-advisor of a recently graduated PhD student at University of Brasilia, Brazil
  • Regular reviewer of IEEE/SPIE journals and conferences.
  • Session chairs: Video Coding in ICASSP 09, Authentication and Cryptography in ICIP 06.
  • MPEG Ad hoc Group co-chair on Core Experiments, and Streaming Adaptations for MPEG-21.

Talks:

  • Keynote speaker at SPIE Stereoscopic Displays and Applications XXIII Conference, Burlingame, CA, Jan 2012, on "The Past, Present and Future of YouTube 3D."
  • Keynote speaker at 3DCine Workshop, at CVPR Conference, Providence, RI, June 2012,  on: "Monoscopic to stereo video conversion of YouTube videos."
  • Several Google workshops and invited talks.

Granted Patents

  • 47 granted patents, several pending

Publications:

  • Research Gate profile: https://www.researchgate.net/profile/Debargha_Mukherjee

Research and Engineering Projects (selected):
  • Video codec research: Vp9, Av1. [Google]
  • 3D support on Youtube, including an innovative service for 2D to 3D conversion using depth map generation, followed by depth image based rendering. [Google]

  • Release of new formats for Youtube, such as 2K/4K resolutions, WebM formats based on Vp8 video codec, as well as the first formats with multichannel audio. [Google]

  • Developed new adaptive bitrate control strategies for Youtube videos, that make the best rate-distortion-computation trade-off based on cloud computation resources. [Google]

  • Lossless recompression algorithm for JPEG images without use of arithmetic coding. [Google]

  • Proposed and managed funded collaborations between HP Labs and University of Brasilia on next generation video coding between 2005–present, between HP Labs and Zhejiang University on video coding between 2006–08, and HP Labs and Tsinghua University on image classification between 2007–08. [HP]

  • Initiated and led project on distributed audio-visual signal processing, view selection, view interpolation, and single- and multi-view video coding, for next-generation teleconferencing systems, based on multiple ad hoc cameras and microphones. [HP]

  • Developed a robust and practical mixed resolution framework for reversed complexity Wyner-Ziv video coding that can be readily supported on existing codecs as an additional mode. The new mode implemented on H.263+ and recently on H.264, shows promising results. Theoretical analysis of optimal parameter selection for Wyner-Ziv coding has also been conducted. [HP]

  • Co-developed an error-resilient video transmission and error-concealment mechanism called RECAP using a low-resolution reduntant bit-stream in conjunction with Reference Picture Selection.

  • Developed a scalable complexity H.264/AVC video encoder and a mechanism for coding delay control. [HP]

  • Developed Structured Scalable Metaformats (SSM), a comprehensive framework for fully format-agnostic adaptation of scalable media formats for supporting heterogenity of networks and terminals. SSM was proposed for standardization in MPEG-21 Part 7 (Digital Item Adaptation), and several of its components were eventually adopted in the standard after 2+ years of extensive travel as part of the standardization process.  [HP]

  • Developed algorithms for key generation, authentication and progressive encryption for generalized scalable bit-streams. [HP]

  • Developed a scalable image sequence compression scheme based on DCT and wavelet coding of motion compensated residuals, for image based 3D rendering of images for object visualization. [HP]

  • Developed a multimode, rate-controlled compound document compression scheme called GRAFIT, that guar antees compression by a tunable factor in the range 4-12:1 for all images in a single pass, at complexity significantly lower than that of JPEG and other existing codecs. This codec was implemented in ASIC and is currently used in a very wide range of HP printer products.[HP]

  • Developed a MRC-based compound document compression scheme, using JPEG with arithmetic coding or JPEG2000 for coding the image layers, and JBIG for the mask layer. [HP]

  • Developed the Vector Set Partitioning (VSPIHT) algorithm and several variations of it for wavelet based image and video coding, based on magnitude ordering of wavelet vectors followed by classified successive refinement vector quantization (VQ). Two new VQ techniques – VLVQ and LT-CSVQ – are invented. [UCSB, Ph.D. Dissertation]

  • Developed a generic algorithm for hiding different kinds of secure data in image and video hosts, using transform domain vector perturbations. Systems hiding speech in video, and video in video, were implemented. [UCSB]

  • Developed an algorithm for optimal mode and quantizer selection in the P-frames of the H.263-family of multimode video coders. An optimized encoder produces bit streams that are compliant with the H.263 standard, but is more efficient in the rate-distortion sense than the H.263 Test Model (TMN). [UCSB]

  • Developed a region-based video codec, using edge-flow spatial segmentation, hierarchical affine region matching for motion estimation followed by spatio-temporal region merging, affine motion compensa tion, and DCT based residual error coding. [UCSB]

  • Developed a very low complexity, high quality CELP speech coder operating between 12-16 Kb/s, using backward and inverse filtered residual and tree-structured multitap adaptive codebooks. [Texas Instruments]


Comments