Course and Semester Hour Credit Requirements
This portfolio training program will be an integral part of the pre-doctoral education of students in the inter-institutional University of Texas Biomedical Engineering graduate program and the Department of Electrical and Computer Engineering. Students must meet all of the requirements set by either the BME GSC or the ECE GSC for graduation from our institution. Students admitted to the imaging portfolio program from other departments must satisfy all of the degree requirements set forth by their departmental GSC. The imaging curriculum described in this portfolio program will count towards the course requirements for Ph.D. students in BME or ECE. Students continuing for a PhD in BME or ECE must also write a research proposal (NIH format), pass an oral qualifying exam, and successfully defend their doctoral research. The average length of the doctoral graduate educational experience is five years.
All imaging science portfolio trainees are required to take the following four core classes.
Note: One of the guidelines for portfolio programs is that students take courses offered by at least two departments outside their major department. The four core classes include one BME class, two ECE classes, and one CS class. Cross-listed BME numbers have been assigned to the classes for record keeping. BME is by nature interdisciplinary and this portfolio program includes faculty from outside the department and from UTMDACC, UTHSCH, UTMB and USAF HEDO. For these reasons, the directors of this program believe that the intention of the graduate school guideline is met through the courses listed below.
1. Biomedical Imaging Modalities: Principles, Systems, and Applications
BME 381J.3 Biomedical Imaging Modalities: This course is an introduction to biomedical imaging systems. The main objective of this course is to expose the students to the world of biomedical imaging with an emphasis on principles, systems and applications of each modern biomedical imaging modality. For each imaging modality, the following approach is used: 1) describe basic physics; 2) describe hardware and software; 3) analyze advantages and limitations; 4) discuss biomedical and clinical applications. The first module covers ultrasound imaging including grayscale ultrasonography, Doppler and color flow imaging, and advanced imaging methods. In the second part of the course, optical imaging methods such as optical coherence tomography and confocal microscopy are introduced. In the third module, the focus is on nuclear magnetic resonance covering magnetic resonance imaging (MRI) and MR spectroscopy. Finally, X-ray and nuclear medicine imaging including computed tomography (CT), single photon emission computed tomography (SPECT), and positron emission tomography (PET) are presented. Fundamental similarities between the imaging equations of the different modalities are stressed, and fundamental differences between different modalities are discussed. Throughout the course, theoretical studies are accompanied with experimental and numerical laboratory exercises. Concepts of signal and image processing are introduced. This course will be taught by Dr. Stanislav Emelianov.
2. Image Processing
BME 381J.7 (ECE 371R) Digital Image and Video Processing: This coure is a comprehensive introduction to the broad field of image and video processing. Topics presented in this class include: image types, imaging geometry, image acquisition, imaging devices, image representation, binary image processing, histogram manipulations, point operations, algebraic and geometric image transformations, two-dimensional discrete Fourier Transform, sampling theorem, spatial convolution, linear image filtering, linear image enhancement, linear image restoration (deconvolution), nonlinear image filtering and enhancement, image quality assessment, image compression, digital image analysis, edge and line detection, Hough transform, spline representations of curves, computational stereopsis and digital video processing, including video compression by current and emerging standards such as MPEG and H.26X. This course requires a substantive project that will employ data collected in the focused modality elective. This course will be taught by Dr. Al Bovik.
3. Modeling and Visualization
BME 395T (CS 395T )Multi-Scale Bio-Modeling and Visualization: Biomedical modeling and visualization has roots in medical illustration and communication for the health sciences, with branches of application to mathematical modeling and computer simulation of artificial life. This course emphasizes computational image processing, and modeling algorithms with emphasis on spatial realism, and the programmatic use of simulation and visualization to quantitatively depict "how things work" at the molecular, cellular, tissue, and organ level scales. Computational methods include multi-scale geometry representations, image filtering, contrast enhancement, segmentation, fusion, boundary and finite element meshing, spline interpolants and approximants and, their use in integral and differential equation solving, quadrature and cubature formulas, volumetric contouring, volumetric rendering, volumetric texture-based image and geometry composition, combinatorial, topological and integral/differential metric quantitation. Practical exercises on computational domain and physiological modeling and visualization at multiple scales shall be drawn from cardiology (heart, cardiac tissue, myocytes, ion-channels), neurology (brain, spinal cord, neurons, Schwann cells, neurotransmitters), and their interactions (synaptic transmission at the neuro-muscular junction). This course requires a substantive project that will employ data collected in the focused modality elective. This course will be taught by Dr. Chandra Bajaj.
4. Informatics
BME 383J.7 (ECE 380L.10) Data Mining for Biomedical Imaging Informatics: Imaging creates vast amounts of data. Imaging scientists have now begun applying multi-media data mining techniques to discover and extract pieces of information useful not only for data interpretation and diagnosis but also to provide feedback on data gathering and modeling. Often these are not stand-alone systems but serve as aids to human experts to help them sift through vast amounts of data and focus on the most significant pieces of information. Effective data mining, as opposed to data dredging, requires an understanding of concepts from exploratory data analysis, pattern recognition, machine learning, heterogeneous databases, parallel processing, and data visualization, in addition to knowing the problem domain. Topics include the iterative data mining process, pre-processing, cleaning, reduction, feature extraction and visualization, clustering/segmentation, classification and decision support (Bayesian vs. neural vs. knowledge based approaches such as CBR), and ensembles (combining multiple models for accuracy and robustness). This course requires a substantive project that will employ data collected in the focused modality elective. This course will be taught by Dr. Joydeep Ghosh.
In addition, students are encouraged to take the following three elective classes:
1. Focused Elective in Instrumentation, Devices, and Contrast Agents
Students will select a course focused on a single modality from a list of classes covering a broad range of imaging modalities: optical microscopy, MRI, OCT, SEM. TEM, US, etc. By the conclusion of this course, students will have imaging data that they can analyze in the project portions of the following courses in the core. Students may elect to take more than one of these courses based on their research interests. A list of the many practical experience electives is included in Appendix A
2. Ethics seminar
BME 197E Professional Development and Professional Responsibilities for the Imaging Scientist. Lectures will be presented by participating faculties. The VaNTH modules on professional responsibility and the ethical conduct of research will provide a forum for discussion. This seminar will have a practical emphasis on situations the trainees are likely to encounter in their careers. Seminar topics will include: plagiarism, review process for scientific papers and grants, entrepreneurship, patents, intellectual property, scientific fraud and misconduct, and career development. We will adopt the successful format developed in the molecular imaging IGERT. The seminar will be presented as a three-day symposium just before the beginning of formal classes each semester. Attendance is mandatory for trainees and optional for other BME students. This course is organized by Drs. Rylander and Markey.
3. Research Seminar
BME 197R Imaging Research Seminar. This seminar will meet weekly to discuss recent advances in biomedical imaging and to provide trainees with guidance in their development as imaging scientists. The seminars will be coordinated with the BME seminar series, presented interactively via video conferencing to all UT BME sites, and open to all BME students (not just imaging science trainees). Five guest speakers per semester will present state-of-the-art imaging research and discuss career opportunities. Imaging trainees in their second year are required to present a seminar on their own research. While all of the speakers will be encouraged to include professional development topics in their presentations as appropriate, some of the seminars will be specifically devoted to this issue. For example, the FIC will give two seminars per year on topics such as recent developments in instructional technologies. As part of the research seminar course, trainees will create an electronic portfolio using Polaris, the College’s online portfolio tool (http://www.polaris.engr.utexas.edu/). We will also use the seminar series to highlight diversity issues and our associated professional responsibilities. For example, we will encourage our trainees to consider research opportunities on improving the quality of life for people with disabilities by ensuring that at least one research seminar speaker per year addresses such topics. This course is organized by Drs. Rylander and Markey.
The BME and ECE core classes provide the math and physiology background for the proposed training program in imaging science. For example, the Biostatistics course in the BME core provides students with a foundation in hypothesis testing and statistical modeling that is extended in the data-mining course in the imaging sciences core. The Analysis of Biomedical Engineering Systems - Parts I and II classes cover math modeling and fields, forces, and flows in physiological systems. The four imaging sciences courses are the minimum required by this portfolio training program. Students may wish to take additional classes depending on their research projects.
Additional Requirements
Trainees in the proposed program will be BME or ECE students; and hence, the advising system for imaging science trainees will generally follow the standard procedure for their home department, with a few additional guidelines. Imaging science students will have a primary BME or ECE supervisor with expertise in one of the four core imaging science areas (instrumentation, devices, and contrast agents; image processing; modeling and visualization; or imaging informatics) and a co-supervisor with expertise on the practical application of imaging for scientific or clinical discovery. In addition to the two supervisors, the dissertation committee must have at least one member outside the BME or ECE departments and at least one member from each of the core imaging science areas. Most BME and ECE students do not organize their committees until they are nearly ready for the qualifying exam (around the end of their second year). In comparison, imaging science trainees will identify their co-supervisors and advising committee during the supervisor selection period. While changes to the dissertation committee will be common as students’ interests evolve, we feel it is critical that the trainees have a group of faculty they can consult from the beginning of their studies, given the scope and inter-disciplinary nature of the proposed training program.
An important innovation of the training program over the typical graduate experience is that during the summer between the first and second year, the trainees will have the opportunity to perform a research internship. The intention of the internship experience is to provide the trainees with hands-on experience on an imaging science application within an organizational structure that is different from an engineering college. Examples include working with a basic scientist (e.g., chemist) in a natural sciences department, a basic or clinical scientist in a medical facility, a basic or clinical scientist in industry, or basic or clinical scientist in a government facility.
The second summer will be a primary research experience during which the trainees will prepare the research proposal for their qualifying exam to be taken before the third year. This research experience may be with a basic scientist, clinical scientist, or an industrial internship. After the two year portfolio program, the students typically work on their research projects full time for about three years. We expect most students to complete the Ph.D. within five years. Students are expected to publish their research in peer-reviewed journals and attend and present posters/papers at national meetings like SPIE Medical Imaging.