Course catalog description: Introduction to computer vision and robotics. Image formation and analysis. Rigid body and coordinate frame transformations. Low-level vision and edge detection. Models for shading and illumination. Camera models and calibration. 3-D stereo reconstruction. Epipolar geometry and fundamental matrices. Motion estimation.
Credits and contact hours: 3 credits; 1 hour and 20-minute session twice a week, every week
Pre-Requisite courses: 14:332:345, 14:332:346
Co-Requisite courses: None
Topics Covered:
- Computer Vision Overview, robot kits
- Extension of One-Dimensional Signal Processing to Two-Dimensions, Convolution, Image filtering, Discrete Fourier transforms, Sampling theory
- Linear Algebra, Basic principles
- Numerical Methods, Least squares estimation, Singular value decomposition
- Image Analysis, Image pyramids, Image features, Edge and corner detection
- Rigid body transformations, rotation, translation, homogeneous coordinates
- Coordinate frame transformations
- Image Formation and Camera Models, Perspective projection, Homography
- Camera Calibration
- Stereo Vision, Point correspondences, Epipolar geometry
- Three-Dimensional reconstruction, Generating depth maps
- Motion, Optical flow, Affine motion models, Image stabilization
- Feature-based object recognition using statistical inference
- Appearance-based modeling, Eigenspace methods, Object recognition
Textbook: E. Tresso and A. Verri, Introductory Techniques for 3-D Computer Vision, Prentice-Hall; R. Szeliski. Computer Vision: Algorithms and Applications, Springer; D. Forsyth and J. Ponce, Computer Vision: A Modern Approach, Prentice-Hall.
Other supplemental material: Learning Open CV, Bradsky & Kaehler, O’Reilly; MatLab: Student Version, Current Edition, The MathWorks, Inc., J. Knudsen