Machine Vision and Pattern Recognition

Catalog description

Covers the field of computer vision, whose broad goal is to create algorithms and systems for processing of visual signals (e.g., images, videos) for low-level, mid-level, and high-level perceptual tasks. Presents the broad principles and techniques for devising computer vision algorithms starting from understanding the imaging process for a pin-hole camera; understanding lenses, image-statistics such as gradients and edges, 3D structure estimation, motion estimation, illumination modeling to perceptual tasks such as shape recognition, texture modeling, face recognition, activity recognition, and scene recognition. The class is a mixture of in-class lectures and discussions, and individual and group projects.