COURSE TITLE: EECS 395/495 Computational Photography: What and How

CATALOG DESCRIPTION:  Seminar to explore computational photography topics by selected readings, lectures on underlying principles, and self-directed projects.

REQUIRED TEXTBOOK:  None: selected readings, starter code & course notes supplied

COURSE COORDINATOR: Jack Tumblin with Xiang Huang

COURSE DESCRIPTION:

Computational photography combines plentiful low-cost computing, digital sensors, actuators, and lights to escape the limitations of traditional film-like methods. New methods offer unbounded dynamic range and variable focus, lighting, viewpoint, resolution and depth of field; hints about shape, reflectance, and location. Instead of fixed digital snapshots and video playback, computational methods promise direct interactions to explore what we photograph.

COURSE GOALS:

Students will learn a wide range of recent techniques, including mosaicking, panorama stitching with lens distortion correction, high dynamic range imaging and tone mapping, digital photomontage, flash/no-flash imaging, separating direct/indirect illumination with high-frequency illumination, image-based relighting, light field photography, all-focus imaging, coded aperture imaging, event-based time-lapse video and more. We will gain understanding of the underlying principles and mathematical tools involved, including lens basics, radiometry, projective geometry in 2D and 3D, paraxial ray descriptions of 4D light fields and 8-D reflectance fields, SVD, homographies (DLT), bilateral filtering, graph cuts, seam carving, Fourier Slice theorem, etc. We will learn to use some free/open source computational photography tools such as Hugin (panoramas), HDRshop/PFStools (tone map), Voodoo (Camera Tracking), etc. We will rely on class consensus to pursue greatest depth in topics of greatest interest.

PREREQUISITES:

Linear Algebra, Calculus, Matlab. Recommend OpenGL and C/C++. No prior knowledge of optics, graphics, image processing and computer vision are required, but may prove helpful. Most students will need (and want) their own digital cameras for this course. We have only a few (3) for 2-day loans for student projects.

HOMEWORK ASSIGNMENTS:

GRADES: