Research Interests
Present Research and Development Activities
Ultrasound of the Breast
Breast cancer is one of the leading causes of death from cancer.
Early detection is widely believed to reduce breast cancer mortality by allowing intervention at an earlier stage of cancer progression.
Screening [X-ray] mammography has secured a place as the gold standard routine health maintenance procedure for women – a mature technology
that provides high-quality images in the majority of patients. However, conventional mammography does not detect all breast cancers,
including some that are palpable, and as many as three-quarters of all breast lesions biopsied because of a suspicious
finding on a mammogram turn out to be benign. The purpose of this research is to find alternative solutions for early detection of breast cancer.
Image on the right is from:
Whole-breast ultrasound tomography shows promise
Ultrasound Beamforming and Ultrasound Tomography ("UST")
The conventional ultrasound approach is driven by the need for real-time data acquisition and display. Therefore, some of the complex physics associated with propagation of sound waves is traded off. One of the tradeoffs corresponds to the usage of straight-ray theory, a basic approximation of the true physics of acoustic wave propagation, which is only valid for purely homogenous media. A second important tradeoff is the assumption of a two dimensional geometry in which only the directly backscattered reflections are collected. The purpose of this research is to implement beamforming and tomography approaches that “undo” the trade-offs of conventional ultrasound, leading to a marked increase in the signal-to-noise ratio, while reducing artifacts and yielding higher quality images for greater clinical “sensitivity”. Furthermore, the signals that propagate through the anatomy, and which are never reflected back, contain additional information.
Previous Research and Development Activities
Segmentation of Ultrasound Images:
Medical image segmentation algorithms, applicable to various scenarios were researched. For example, segmentation in ultrasound images of fetal body from surrounding fluid, segmentation of the fetal body from placenta or uterus [usually a very difficult task], segmentation of fetal spine and segmentation of an incomplete fetal head. An important component of the research was focused on an advanced edge-detection research, a specific approach that handles well a speckled environment of ultrasound images. Model based segmentation has also been studied for specific needs.
Fetal Weight Determination with Ultrasound:
The clinical importance of knowing the fetal weight during pregnancy is due to its relation to the development of the fetus. The weight according to the gestational age and the rate of fetal weight increase can indicate normal development or detect growth abnormality. Weight estimation at term is a very important parameter in the determination of cesarean section delivery. This study was focused on developing apparatus and methods for measuring the weight of a fetus in utero using 3-dimensional ultrasound. The work included the research on automating fetal weight calculation based on image processing technologies.
3-dimensional and 4-dimensional ultrasonic imaging:
The study and creation of ultrasonic imaging systems by a variety of technologies, including motorized transducers and free-hand 2-dimensional transducers with inertial [gyroscopic] sensors.
The Heart in Ultrasound Imaging:
Three-dimensional fetal echocardiography may improve prenatal screening for fetal heart syndromes. We researched and reported together with clinical collaborators that three dimensional ultrasound diagnoses of fetal cardiac malformations are feasible, and may allow generalists to improve the sensitivity and specificity of their screening evaluation. We have also researched the implementation of intra-heart ultrasound imaging of adults, using accurate electormagnetic positional sensors for obtaining three-dimensional ultrasound of the interior of the heart.
Guidance of Minimally Invasive Surgical Procedures using Real-Time 3D Ultrasound:
The objective of this project was to develop a 3D ultrasound technology that will assist the physicians in performing minimally invasive procedures safer, faster and more cost effectively than they are today. This requires advances in both the ultrasound probe and three dimentional (3D) display technologies. These technologies were explored in two forms. One is a 3D laparoscopic probe that is motorized, allowing the physician to acquire 3D without requiring him to scan the transducer manually as well as enabling continuous 3D acquisition and display. A prototype probe was developed and successfully used in animal trials with a target imaging system. The second research aspect was on a real-time (4D) probe for needle guidance.
Thermo Field Dynamics:
The application of the Non-Equilibrium Thermo Field Dynamics to molecular systems has been studied. A new Wigner phase space distribution function has been introduced for thermal quasi particles currently in use in Thermo Field Dynamics, thereby permitting a new way for the numerical investigation of quantum and thermal fluctuations.
Fast Reactions in Liquids:
Stochastic dynamics has been used in the interaction picture to define energy transfer in collisions within a liquid. Langevin equations of motion with minimum number of physical parameters have been constructed and numerically integrated. The important influence of the liquid's "velocity friction" on the energy transfer efficiency has been studied numerically.
Time-dependent Solution of the Liouville von Neumann Equation: Non-dissipative and Dissipative Evolution:
A new method for solving the Liouville von Neumann equation has been developed. Fast Fourier Transform (FFT) is used extensively, this transformation preserving all exact commutation relations. The accuracy and convergence properties of the method have been investigated and compared to an exact solvable model problem. Typical non-trivial applications were studied numerically and included thermal relaxation under constraints of selection rules.
Self Consistent Quantal Equations of Motion within the Lie Algebraic Setting: Atom-diatom Collisions:
In this study, a direct and practical new procedure for the computation of density maximal entropy has been implemented. The method determines directly the time evolution of parameters of the density operator using their equations of motion generated via a Lie algebra. Energy transfer to the vibrational motion of the molecule has been determined numerically.
Electron Molecule Scattering:
In this research, the electronic and nuclear motions in the target molecule have been investigated in new ways. The electronic problem has been treated within a many-body theory by perturbation methods and a generalized optical potential. The scattering problem has been treated using the Schwinger variational principle. The behavior of important shape resonances has been investigated in detail using a new and sophisticated library of computational tools.
Theory of the Double Resonance Raman Amplifier:
The use of stimulated resonance Raman emission in the three-level double resonance has been considered. In this project, a new model has been presented for the steady state three-level resonance Raman amplification, based on an algebraic expression for single molecule amplification, combined with a numerical solution of non-linear equations governing macroscopic collinear propagation.
Excited States of Rare-Gas Molecules:
Rare-Gas molecules are chemically unbounded in their electronic ground state. However, many of the electronic excited states are strongly bounded. Their possible usage in UV lasers is of great importance. A new and fast method for obtaining excited-state potentials of rare-gas diatomic molecules has been developed.
Statistical Mechanics of Adsorption of Large Molecules on Charged Electrodes:
The influence of the voltage on the sticking of molecules to an electrical double-layer has been studied. This behavior is of interest in electrochemistry. A new statistical mechanical model has been developed to analyze numerically the experimental findings.