Princeton University

School of Engineering & Applied Science

Image Enhancement in Optics and Ultrasound Using Nonlinearity

Speaker: 
Jen-Tang Lu
Location: 
001 Sherrerd Hall
Date/Time: 
Wednesday, May 10, 2017 - 9:00am to 11:30am

Abstract
Images have played a crucial role in our life and they are everywhere: from newspapers to Facebook, from microscope to telescope, and from photography to medical imaging. While imaging pushes technology forward, conventional imaging systems suffer many limitations including limited resolution, a trade-off between contrast and resolution, and limited sensitivity to phase. In this dissertation, we overcome these limits by developing new imaging schemes using nonlinear effects (where waves or photons interact with each other) and computational imaging techniques. The methods are demonstrated in optical imaging and medical ultrasound imaging.
 
In optical imaging, we focus on phase imaging, which plays an important role in a variety of fields, including astronomy, biomedical imaging, and material science. Specifically, we use nonlinear wave propagation (where the wave speed depends on its intensity) to enhance phase imaging, showing that the conservation of energy and momentum (phase-matching conditions) leads to restriction on the dynamics, and thus to constraints in the computation used to reconstruct the phase. The method can be much faster and more sensitive than conventional approaches using linear wave propagation.
 
Although nonlinear effects play a key role in image enhancement, to date, nonlinear optics has relied on physical media (e.g. crystals) to work. These materials are all constrained by their physical properties, such as frequency selectivity, limited time behavior, and fixed nonlinear response. To address these issues, we use an electro-optic spatial light modulator to produce nonlinear effects without a physical medium. This "digital" method greatly facilitates nonlinear optics, because it works for any light intensity, has a fast response time, operates over a broad bandwidth, is dynamically adjustable, and can generate nonlinearity with any arbitrarily mathematical form.
 
The last part of the thesis focuses on medical ultrasound imaging, which is a popular and non-invasive method of imaging inside the body but generally suffers low resolution, low contrast, and noise. We use the same ideas of nonlinear wave mixing to enhance ultrasound imaging. We propose an imaging processing method, which leverages mutual information between different ultrasound signals, to create a composite image with better contrast, higher resolution, lower noise, and more tissue-specific response.