Harmonic Deconvolution Harmonic Deconvolution in in

Ultrasound Vibro-Acoustic imagesUltrasound Vibro-Acoustic images

Alexia Giannoula

Communications group, Dept of Electrical & Computer Engineering, University of Toronto

Elastography

Changes in elastic properties of soft tissue have been often attributed to the presence of disease or abnormal structures

Most techniques in elasticity imaging or elastography involve:» tissue excitation by an external or internal force» detection of the tissue motion or displacement

• Using Ultrasound, magnetic resonance (MR), acoustic/optical methods

Hard inclusion/tumor: smaller displacementB-mode elastogram

Acoustic Radiation Force A way to excite directly a target inside the body is through the

use of the radiation force of ultrasound

Advantages:» Non-invasive (external) excitation» Highly-localized radiation stress field (leads to increased precision)

The radiation force mainly depends on:» The type of propagating medium (lossless/lossy, viscoelastic fluid etc.)» Mechanical properties of the target object» Geometry of the target object

Ultrasound Vibro-Acoustography (USVA)

1. Two CW beams at slightly different frequencies interfere in the focal zone2. A modulated ultrasound field is generated at the “beat” frequency Df (low)3. A highly-localized dynamic (oscillating) radiation force is produced4. In response to the force (stress field), the object vibrates at the same Df5. Vibration acoustic emission Detected by hydrophone/laser vibrometer

• Detection sensitivity: few nanometers• Image resolution: PSF~700μm

USVA X-Ray Photo

Proposed Deconvolution Scheme (I) Usually a blur is observed around the object

» Due to the sidelobe effects of the system PSF

Apply separate deconvolution to the fundamental and second harmonic signals recorded by the hydrophone

Higher-harmonics arise due to tissue nonlinearities» Harmonic imaging better resolution and less noise/blur

Second harmonicFundamental

Proposed Deconvolution Scheme (II)

First and Second-harmonic image formation:

? ξ(r1) object function

Each PSF hi represents the response of a point target to the radiation force Fi (i=1,2)

Obtain 2 deblurred images ξ1, ξ2

Fuse the outputs based on the differentattenuations: ξ = α1 ξ1 + α2 ξ2

Form h1, h2 Filter each χ1(r), χ2(r) with the inverse PSFsFind F1, F2