HI 5323 Image Processing
This course provides a broad and practical introduction to the major techniques employed in image processing and pattern recognition: dilation and erosion, segmentation and thresholding, denoising, direct space filter kernels, Fourier-based filters, matching and morphing, artificial neural networks, self-organizing maps, principal component analysis. The course will be useful for graduate students in biomedical computing who wish to learn state of the art data in mining and image vision techniques.
We will cater both to biological/clinical and quantitatively trained students. A background in at least one quantitative discipline (physics, chemistry, mathematics, computer science) at college level with solid background in geometry (ideally: vector calculus) is desirable. Knowledge of at least one programming language (ideally: C or C++) and UNIX, or willingness to acquire necessary skills.
By the end of the semester, the student will have had the opportunity to meet the following objectives:
- Characterize the role of image processing and pattern recognition in concurrent biomedical imaging research, and in robotics and engineering applications.
- Describe the functionality, advantages, and limitations of standard computing strategies used in data mining and image vision.
- Acquire a working knowledge of freely available software and algorithms to carry out independent research projects.
- Explore the possibilities for biomedical imaging to assist in the process of determination, analyzing, evaluating, displaying, and retrieving of 2D and 3D data in a research or industry laboratory environment.
- Develop pieces of software and computer scripts that serve as template for own future research work.