Time-Frequency Analysis for Ultrasonic Echo Detection and Parameter Estimation

Event Date: 05 Dec 2018 (Wednesday) - 05 Dec 2018 (Wednesday)

Kinesis Building, Seminar Room 1, Level 5

Time :02:00 - 03:00

Dr. Jafar Saniie

Filmer Endowed Chair Professor and IEEE Life Fellow 
Chair, Department of Electrical and Computer Engineering,
Director of Embedded Computing and Signal Processing Research Laboratory,
 Illinois Institute of Technology

Ultrasonic testing and imaging applications often resort to signal modelling, parameter estimation and data analysis for aberration detection, pattern recognition, and classification.  In this study we present a unified time-frequency analysis and echo parameter estimation to characterise ultrasonic non-stationary signals for imaging and diagnosis.  In particular, highly complex and interfering patterns of scattered signals are decomposed into elementary chirp echoes (i.e. chirplets) for analysis. The decomposition of complex signals facilitates a systematic, tractable and quantitative approach to correlate the estimated chirplets to the actual physical characteristics of the objects and their embedded environment that generates the reflected and scattered ultrasonic echoes.  Identifying the source of ultrasonic signals has a broad range of medical and industrial ultrasonic testing applications including tissue characterisation for medical diagnosis, flow measurements, nondestructive testing, quality control in multi-component/composite materials, and real-time structural health monitoring for integrity assessment and safe use. 

Ultrasonic signal decomposition is a challenging task and requires determining critical components including: i) the design and suitability of the signal model; ii) the complexity of the model in terms of the number of parameters; iii) the development of efficient parameter estimation algorithms; and  iv) above all, a clear understanding  of how the  source of the signal contributes to the complexity of the signal which can be deciphered into sensible constraints for  accurate parameter estimation with diagnostic values.  

To pursue these objectives, this speaker will be presenting several analytical tools stemming from the generalised time-frequency (GTF) distributions.  GTF distributions not only represent powerful tools for signal analysis but also reveal efficient methods for noise reduction and data compression.  This study highlights the link between several time-frequency distributions including Chirplet Transform, Wigner-Ville Distribution, Short-time Fourier Transform, Split-Spectrum Processing, Choi-Williams Distribution, and Generalized Ambiguity Function.  These transformation methods are potentially viable for unraveling ultrasonic signals consisting of many interfering echoes into a structure that leads to the mathematical steps needed for coherent algorithms to estimate echo parameters.

In this research a family of GTF distributions has been explored. Analytical equations have been derived not only to improve the signal-to-noise ratio of ultrasonic signals but also to provide stable and efficient computational algorithms for echo parameter estimation.  Numerous experimental and simulated ultrasonic scattering signals and signals consisting of multiple interfering echoes are studied to evaluate the accuracy and practicality of the developed echo parameter estimation algorithms.  Results will be presented to address and compare the parameter estimation accuracy and the computational efficiency of various time-frequency distributions.

Dr. Jafar Saniie (IEEE Life Fellow for contributions to ultrasonic signal processing for detection, estimation and imaging) received his B.S. degree in Electrical Engineering from the University of Maryland in 1974. Following that, he was conferred his M.S. degree in Biomedical Engineering in 1977 from Case Western Reserve University, Cleveland, Ohio, and his Ph.D. degree in Electrical Engineering in 1981 from Purdue University, West Lafayette, Indiana. In 1981, Dr. Saniie joined the Department of Applied Physics, University of Helsinki, Finland, to conduct research in photothermal and photoacoustic imaging. Since 1983, he has been with the Department of Electrical and Computer Engineering at Illinois Institute of Technology where he is the Department Chair, the Filmer Endowed Chair Professor, and Director of the Embedded Computing and Signal Processing (ECASP) Research Laboratory. Dr. Saniie's research interests and activities are in ultrasonic signal and image processing, software defined ultrasonic communications, statistical pattern recognition, estimation and detection, data compression, time-frequency analysis, embedded digital systems, digital signal processing with field programmable gate arrays, and ultrasonic nondestructive testing and imaging. He has over 330 publications and has supervised 35 Ph.D. dissertations to completion.

Dr. Saniie was a Technical Program Committee member of the IEEE Ultrasonics Symposium since 1987 (The Chair of Sensors, NDE and Industrial Applications Group, 2004-2013), the Lead Guest Editor for the IEEE UFFC Special Issue on Ultrasonics and Ferroelectrics (August 2014) and the IEEE UFFC Special Issue on Novel Embedded Systems for Ultrasonic Imaging and Signal Processing (July 2012) and the  Associate Editor of the IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control since 1994. He was also the General Chair for the 2014 IEEE Ultrasonics Symposium in Chicago, and served as the Ultrasonics Vice President of the IEEE UFFC Society (2014-2017). 


Dress code: Office attire
Registration is free
Click here to register

For further enquiries, please contact Dr. Yao Kui (

Last update : 1/11/2019 9:04:19 AM