Hari Kalva

Professional Preparation:

• B.Tech., Electronics and Communications Engineering, S.V. University, Tirupathi, India, 1991
• M.S.C.E., Computer Engineering, Florida Atlantic University, 1994
• M.Phil., Electrical Engineering, Columbia University, 1999
• Ph.D., Electrical Engineering, Columbia University, 2000

Appointments

• 8/11/03 – present, Assistant Professor, Dept. of Computer Science and Engineering, Florida Atlantic University
• 11/01 – 07/03, Consultant, Mitsubishi Electric Research Labs, Cambridge, MA
• 01/00 – 11/01, Co-founder and V.P. of Engineering, Flavor Software Inc., NY
• 01/95 – 08/96, Research Staff Associate, ADVENT Project, Columbia University

Related Publications:

Publication summary: 2 books, 7 book chapters, 19 journal papers, 47 conference papers;
Patents: 7 issued, 12 pending; MPEG and DAVIC contributions: 26

1. G. Fernandez-Escribano, H. Kalva, P. Cuenca, and L. Orozco-Barbosa, “Very Low Complexity MPEG-2 to H.264 Transcoding Using Machine Learning,” Proceedings of the ACM Multimedia 2007, Santa Barbara, CA, October 2006, pp. 931-940.
2. H. Kalva and B. Petljanski, “Exploiting the directional features in MPEG-2 for H.264 intra transcoding,” IEEE Transactions on Consumer Electronics, vol.52, no.2, May 2006, pp. 706-711.
3. H. Kalva and L. Christodoulou, “Using machine learning for fast intra MB coding in H.264,” Visual Communications and Image Processing (VCIP) 2007, IS&T/SPIE Symposium on Electronic Imaging 2007, January 2007.
4. H. Kalva, R. Shankar, T. Patel, and C. Cruz, “Resource estimation methodology for multimedia applications,” Proceedings of the 14th Annual ACM/SPIE Multimedia Computing and Networking Conference (MMCN'07), San Jose, CA, Jan 2007.
5. J.B. Lee and H. Kalva, “An Efficient Algorithm for VC-1 to H.264 Transcoding in Progressive Video Compression,” Proceedings of the IEEE International Conference on Multimedia & Expo (ICME) 2006, pp. 53-56.
6. D. Socek, H. Kalva, S. Magliveras, O. Marques, D. Culibrk, abd B. Furht, “A Permutation–Based Correlation Preserving Encryption Method for Digital Video,” Proceedings of the International Conference on Image Analysis and Recognition ICIAR 2006, Povoa De Varzim, Portugal, September 2006.
7. H. Kalva, L. Christodoulou, L. Mayron, O. Marques, and B. Furht, “Design and Evaluation of 3D Video System Based on H.264 View Coding,” International Workshop on Network and Operating Systems Support for Digital Audio and Video (NOSSDAV 2006), Newport, Rhode Island, May 22.23, 2006.
8. H. Kalva and B. Furht, “Complexity Estimation for the H.264 Coded Bitstreams,” The Computer Journal, Oxford University Press, Vol 48, No. 5, Sept. 2005, pp. 504-513.

Summary of Current Research Projects:

Video Transcoding: Video transcoding is the process of converting video from one format to another (e.g, MPEG-2 to H.264). We are interested in complexity reduction in multi format video transcoding. We have developed algorithms that reuse the information from the input video decoding stage and accelerate the encoding stage. Our algorithms are based on structured supervised machine learning reduce the complexity significantly. Our current work addresses transcoding to/from MPEG2, H.264, H.263, VC-1, and VP-6.

Low Complexity Video Encoding: H.264 is a highly efficient and complex video codec. The complexity of the codec makes it difficult to use all its features in lower complexity mobile devices. We are developing machine learning based approaches to reducing the encoding complexity. Determining the macro block coding mode requires substantial computational resources in H.264 video encoding. We reduce MB mode computation from a search operation, as is done in the encoders today, to a classification.

Object Encryption for Privacy in Video Surveillance: The key concern with the increasing use of video surveillance is the fact that private citizens, who are not suspects, are being recorded and recordings archived through the use of video surveillance systems. A solution to the problem is selective encryption of objects (e.g, faces, vehicle tags) in surveillance video. Objects in a video can be encrypted to ensure privacy and still allow decryption for legitimate security needs at anytime in the future. We are developing compression algorithm independent solutions to this problem. This allows the use of standard video encoders and decoders and also enables smart-cameras that output encrypted video.