An introduction to statistical learning uic

Prerequisite: BIOEР’В 485 or permission of the instructor.

Stochastic object models will be introduced and applied in various validation tasks.

Unsupervised clustering and cluster analysis of extracted image features will be introduced. Principal component analysis will be described and applied to various image processing tasks. The effects of inter-feature correlation upon prediction confidence will be discussed. Feature selection and dimensionality will be discussed in terms of diagnostic task performance. Cross-validation of image-derived decisions and some common methods of assessing model robustness will be shown. Multiple linear regression will be applied to neuroimaging data and some common methods of assessing model robustness shown. The random forest algorithm will be introduced and used on a number of large- and small-data tasks. Support vector machines will be introduced and applied to image classification and interpretation tasks. The extraction of both low- and high-order spatial features at multiple scales will be demonstrated and employed throughout the course. Histogram transforms and the fundamental properties of image texture will be introduced and revisited throughout the course. Analyses will be done in Python using the Scikit-learn package and all homework assignments comprise statistical analyses of biomedical image data in real decision scenarios. In this course, several applications of statistical learning to biomedical image data will be covered in depth from first principles. Furthermore, the extraction of image features and the characterization of modality-dependent background noise can be particularly challenging in typical biomedical scenarios.

Concurrent enrollment in BIOEР’В 485 and BIOEР’В 580 is required.Р”В«IOEР’В 484 Statistical Analysis of Biomedical Images credit: 3 or 4 Hours.Р”В«iomedical image data often come in extreme numbers: there is either so many of them that humans can't analyze them in reasonable time (e.g., three-dimensional light sheet microscopy data) or they are few, highly varied and of limited spatial and intensity resolutions (e.g., positron emission tomography scans). Positron emission tomography (PET) will be covered and specific clinical issues discussed in terms of reconstruction algorithm and parameter choices. The physics of magnetic resonance imaging will be presented and related to specific mathematical issues of image reconstruction and under-sampled measurement space. X-ray computed tomography will be covered in depth and in the context of the imaging science principles presented the co-requisites this also includes practical concerns about computing resources and modern GPU-based computing. General imaging and detection principles common to all computational modalities will be covered in context of specific biomedical imaging scenarios. In this course, computational image reconstruction techniques will be developed and employed across a broad range of radiographic, magnetic resonance, and nuclear imaging modalities. The frontier of biomedical imaging is computed imaging where multi-dimensional images must be reconstructed from measured data that is otherwise not meaningful to human observers. BIOEР’В 483 Biomedical Computed Imaging Systems credit: 3 or 4 Hours.

An introduction to statistical learning uic