Mehdi Salehifar

Mehdi Salehifar

Staff Research Engineer

Hi! I’m a staff research engineer at LG Electronics at San Diego, California, where I’m researching and developing new technologies for next generation video coding.

I obtained my B.Sc. degree in electrical and computer engineering from University of Tehran, Iran, in 2012 and my M.Sc. and Ph.D. degrees in electrical and computer engineering from the University of California, Santa Barbara (UCSB), in 2014 and 2017 respectively.

Prior to joining LG, I was a graduate research/teaching assistant at UCSB, where I was working in signal compression lab under supervision of Prof. Rose. My research interests include video/audio compression, signal processing, general quantization theory, and information theory. I also worked in LG Electronics as a video researcher intern as well as at Qualcomm as an audio researcher.

I’m a winner of several titles in national and international mathematic competitions.

download CV

Latest publication and patent:

M. Salehifar, M. Koo, S. Kim, J. Lim, “Apparatus on Reduced Transform for Video Coding”, US Provisional Application Number 62599020.

M. Salehifar, T. Nanjundaswamy, and K. Rose, “On Quantizer Design to Exploit Common Information in Layered Coding of Vector Sources,” to be submitted to IEEE Transaction on Signal Processing.

Education.

  • 2012-2017

    Doctor of Philosophy (Ph.D.) in Electrical Engineering (Communication and Signal Processing), University of California Santa Barbara (UCSB)

    Thesis: Common Information and Hidden State Dependencies in Layered Source Coding

  • 2012-2014

    Master of Science in Electrical Engineering (Communication and Signal Processing), University of California Santa Barbara (UCSB)

  • 2008-2012

    Bachelor of Science in Electrical Engineering, Telecommunication, University of Tehran, Iran

    Minor in Abstract Mathematics

graduation

Research.

Common Information in Layered Coding

A novel fundamental design technique for a common information based layered coding framework is proposed

A novel fundamental design technique for a common information based layered coding framework is proposed, wherein jointly designing common and individual layers’ vector quantizers overcomes the limitations of conventional scalable coding and non-scalable coding, by providing the flexibility of transmitting common and individual bit-streams for different quality levels.

Common Information in Layered Coding

Compression of Hidden Markov Sources

A new fundamental source coding approach for hidden Markov sources is proposed

A new fundamental source coding approach for hidden Markov sources is proposed, based on tracking the state probability distribution, and is shown to be optimal. Practical encoder and decoder schemes that leverage the main concepts are introduced.

Compression of Hidden Markov Sources

Low Complexity Transform for Video Compression

A new low complexity transformation for residual is proposed for next generation video coding

A new low complexity transformation for residual is proposed for next generation video coding, wherein mapping the data form higher rank space to data in a lower rank space (Reduced Transform) in adjunct to applying another new proposed transform, i.e., Selective Transform, dramatically reduces multiplication and memory complexity, with virtually no loss in performance.

Low Complexity Transform for Video Compression

Entropy Coding in Video Compression

New residual and data index entropy coding for next generation video compression is proposed

New residual and data index entropy coding for next generation video compression is proposed, based on implicit decoder side index extraction. Practical design of the proposed entropy coder yields to significant performance improvement.

Entropy Coding for Video Compression

I'm passion about data compression, video/audio coding, signal processing, general quantization theory, and
information theory.

Serving as Reviewer

As a member of IEEE, I'm an active reviewer for journal and conference papers in signal compression, video/audio coding and
information theory.

As a member of IEEE, I'm an active reviewer for journal and conference papers in signal compression, video/audio coding and information theory. I have served as reviewer for IEEE Transaction on Multimedia (TMM), IEEE International Symposium on Information Theory (ISIT), IEEE International Conference on Image Processing (ICIP)

Publications & Patents.

common information

Common Information in Layered Coding



On Quantizer Design to Exploit Common Information in Layered Coding of Vector Sources

to be submitted to IEEE Transaction on Signal Processing

Quantizer Design for Exploiting Common Information in Layered Coding

International Conference on Acoustics, Speech and Signal Processing (ICASSP), March 2016

Joint Design of Layered Coding Quantizers to Extract and Exploit Common Information

IEEE Data Compression Conference, March 2016

Hidden Markov Modle

Compression of Hidden Markov Sources



Layered Coding of Hidden Markov Sources

to be submitted to IEEE Transaction on Signal Processing

On Scalable Coding of Hidden Markov Sources

International Conference on Acoustics, Speech and Signal Processing (ICASSP), March 2016

On Optimal Coding of Hidden Markov Sources

IEEE Data Compression Conference, March 2014

Transform

Low Complexity Transform for Video Compression



Apparatus on Reduced Transform for Video Coding

US Patent App. 62599020

Apparatus on Selective Transform

US Patent App. 62609270

Apparatus on Methodized of Scanning Pattern for Patterned Data

US Patent App. 62599024

Entropy

Entropy Coding for Video Compression



Apparatus on Generalized Residual Coding with NSST Index

US Patent App. 62599022

Apparatus on Adaptive Secondary Transform Indicator Coding

US Patent App. 62599021

Professional Experiences.

Staff Research and Standard Engineer

Senior Research and Standard Engineer

LG Electronics, USA | July 2017-Present

  • Research and developing new technology for next generation video coding standard

Graduate Research Assistant

Department of Electrical Eng., UC Santa Barbara, CA | Sep. 2012 – July 2017

  • Exploit common information in layered coding and designing optimum coding scheme for real audio and video signal
  • Designed and optimized encoding and decoding system for scalable and individual coding of Hidden Markov Sources

Interim Engineering Intern

Qualcomm Incorporated, USA | Sep. 2016 – Dec. 2016

  • Developed and evaluated new algorithms for stereo/multi-channel scalable audio coding and implemented them in next generation audio coder

Multimedia Research Intern

LG Electronics (Mobile Research Lab), USA | June 2016 – Sep. 2016

  • Developed new tools for inter prediction video coding and implemented in HEVC and JVET-D0085 contributor

Undergraduate Researcher

Department of Electrical Eng., Tehran University, Tehran, Iran | April 2011- Aug. 2012

  • Used fractional Fourier transform in cognitive radio for OFDM’s PAPR reduction
  • Used Cyclotomic fields and Galois extensions for approximating unit before DSP processor

Awards.

  • 2017

    Dissertation Fellowship from ECE Department of UCSB

  • 2012-2017

    Full Assistantship from UC Santa Barbara for Masters and Ph.D. Program

  • 2012

    International Mathematical Competition: Honor of Mention

  • 2012

    Iranian National Mathematics Competition: Gold Medal

  • 2011

    Iranian National Mathematics Competition: Bronze Medal

  • 2008-2012

    Four Year Scholarship from Iranian National Elite Organization.

  • 2008

    Iranian National University Entrance Exam: Ranked 105 among 350,000 Participants

  • 2007

    Iranian Mathematics Olympiad: Silver Medal

  • 2007

    Iranian High School Chess Competition: Gold Medal

Teaching.

Contact.