主页

Bo Xiao
Research Assistant
Signal Analysis and Interpretation Lab (SAIL)
Department of Electrical Engineering
University of Southern California

RTH320, USC
skype: bob_xiaobo
E-mail:


EDUCATION

Aug. 2009 - present

PhD student
SAIL Lab, Department of Electrical Engineering, USC
Advisor: Prof. Shrikanth Narayanan

Sep. 2007 - Jul. 2009

Graduate School
New Generation Network Lab
Department of Electronic Engineering
Tsinghua University
Advisor: Prof. Yongfeng Huang

Sep. 2003 - Jul. 2007 Undergraduate
Department of Electronic Engineering
Tsinghua University
B.E. Degree Major: Information and Electronics Engineering

ACADEMIC EXPERIENCES

Mar.2005 - Mar.2007
Selected Undergraduate Projects/Courses
  • Study on iLBC (internet Low Bit-rate Codec) speech codec, Student Research Training Program
  • 3-level pipeline CPU model design and simulation, Computer Principle
  • Recognition of 3 digits Chinese speech, Speech Signal Processing
Mar.2007 - Jul.2007
Diploma Project  
  • Research on information hiding communication system based on streaming media (e.g. VoIP)
  • Design and Implementation of VoIP softphone based on SIP supporting IPv4/IPv6 on Windows OS using VC 2005
Aug.2007- Jul.2008 Selected Graduate Project/Course
  • Survey on Network Coordinates, Advanced Network Technology, highest score in class
  • Video Signal Processing, second highest score in class
  • Network Security, highest score in class
  • Survey on MIMO channel capacity, Information Theory
  • AdaBoost face detection, Pattern Recognition
  • Adaptive filtering, Modern Signal Processing
Aug.2007 - Jun. 2009
Graduate Research, Team Leader  
  • Support: National Science Foundation of China (NSFC) / National High Technology Development Program of China

  • Optimized Information Hiding in Low Bit-rate Speech:
      Traditional QIM(Quantization Index Modulation) embedding divided the codebook into 2 parts, each labeled '0' and '1', but overlooked the way of partition, which may introduce intolerable distortion in case the partition is not proper.

      In our work we used Graph Theory to prove that it is always feasible to guarantee that every codeword is partitioned into the opposite group of its nearest neighbor.
      As the result, we find a minimum bound of maximum distortion in data embedding.

  • Covert Communication System of Information hiding over VoIP:
      Layered structure is designed for adaption to diverse embedding algorithms, carriers, speech codec and functions.
      The layer over TCP/IP application layer is called Execution and Adaption, whose function is actual embedding into VoIP data and extraction from received data.
      The second layer is called Transmission Management, whose function is to guarantee the accurate secret data transmission and provide interface for application.
      The top layer is called Information Hiding Application, whose function is providing useful application based on the covert communication.

      Specific protocol is design for secret message retransmission in transmission management layer, using ARQ-like scheme.
      The carrier is not reused. Instead, the secret message frame is re-embedded in new coming voice data.

  • Improved Reversible Data Hiding in Image:
      Traditional Histogram Shifting algorithm was carried out in the whole image. This method desires the histogram to be concentrated with a sharp peak.

      Considering the locally correlated nature of image, the total capacity is enlarged if the shifting is carried out in every small block rather than the whole image. However, some blocks may not available for embedding, or change into unavailable after embedding, making the receiver unable to find which blocks are used in carrying messages.

      To address this problem, a scheme of capacity reserving is designed to provide the information of the blocks to receiver. Though such information increases overhead bookkeeping, the real capacity is increased 2 to 5 times due to embedding in small blocks.

  • Other works in the team:
      Information Hiding in RTP/RTCP protocols
      Detection of Information Hiding in middle bit-rate speech stream

PUBLICATIONS

  • Xiao Bo, Huang Yongfeng, Shanyu Tang. An Approach to Information Hiding in Low Bit-rate Speech Stream. IEEE Global Communications Conference (GLOBECOM'08), New Orleans, USA, Dec. 2008.
  • Xiao Bo, Huang Yongfeng. Modeling and optimizing of the information hiding communication system over streaming media. Journal of Xidian University (Natural Science). 2008 Vol.35, No.3, P.554-558. (In Chinese)

HONORS/AWARDS

2006.9 Scholarship for Excellent Academic Performance, Tsinghua University
2007.7 Excellent Graduate Thesis, Tsinghua University
2008.11 Intel Scholarship, first class in Tsinghua University

SOCIAL ACTIVITIES

Sep. 2004 - Aug. 2005
Secretary of Youth League branch of Class EE34
Aug. 2006 - Mar. 2007
Practice Group of Youth League of Department EE
Sep. 2006 - Nov.2006
Producer and Director of EE34 Class Movie for Graduation
Jul. 2005 - Jun. 2009
ZIJING Volunteer in Dept.EE, serving over 100 hours

HOBBIES

Swimming, Table tennis, Reading History/Novel, Chinese poem, Movie

Ċ
Bo Xiao,
2008年12月15日 上午1:39
Ċ
Bo Xiao,
2008年12月15日 上午1:39
Comments