Image Processing and Analysis

Lecture 20: Carmichael discusses three main ways of obtaining medical imaging data: CT (Computed Tomography) scans, MRIs (Magnetic Resonance Imaging) and PET (Positron Emission Tomography)

Lecture 19: Carmichael discusses some problems that can arise in range data-based detection and how these problems can be fixed by creating transformation-consistent clusters.

Lecture 18: Carmichael discusses advantages of local shape representations and two methods for creating them: the segmentation method and the vertex-based method.

Lecture 17: The instructor discusses ways of storing and searching 3D models. Explains database querying, geons, object signatures and shape histograms.

Lecture 16: Carmichael discusses two approaches (mesh-based and volumetric) for combining multiple meshes to form a single closed surface.

Lecture 15: In the second lecture on mesh alignment, Carmichale explains nonrigid alignment and how to accomplish this process using a technique called deformable registration.

Lecture 14: In the first of two lectures, Carmichale discusses rigid alignment of meshes as well as the metrics and transformation models involved.

Lecture 13: The instructor discusses using mesh smoothing to remove noise from three-dimensional data. Also includes Gaussian smoothing and mesh shrinkage.

Lecture 12: Carmichael explains how to obtain an image with three dimensions. The unit covers two principles for doing so: time of flight and triangulation.

Lecture 11: Carmichale explains the use of invariants for dense and sparse matching, as well as some various kinds of invariants.

Lecture 09: Carmichael explains why it is useful to study the textures of an image and methods for detecting them in images.

Lecture 10: The instructor explains using neighborhood operations to access different pixels. Also covers implementation of a correlation filter.

Lecture 08: The second of two methods is presented for grouping pixels in an image. The lecture covers deformable contours, parameterizations and gradient descent.

Lecture 07: One method for grouping pixels in an image is presented. Carmichael discusses pairwise coherence, cluster modeling and modeling with metric spaces.

Lecture 06: Carmichael explains how to go about detecting certain objects within an image if those objects cannot be outlined with just rectangles.

Lecture 05: Carmichael discusses finding certain rectangular objects within an image, and explains various methods for doing so, such as PCA (principal component analysis) and dimensionality reduction.

Lecture 04: What it means to have a region of an image be 'salient' and the algorithms for finding such areas.

Lecture 03: The lecture covers edge and corner detection using the Canny and Harris corner detector methods.

Lecture 02: Introduction to Fourier analysis, as well as the subject of wavelets.

Lecture 01: An introduction to image processing and analysis; covers image processing on a broad scope as well as course logistics.