ECE 66100 - Computer Vision

Credits: 3

Areas of Specialization(s):

Computer Engineering

Counts as:

Normally Offered: Fall - even years

Catalog Description:
This course deals with how an autonomous or a semi-autonomous system can be endowed with visual perception. The issues discussed include: vision psychophysics, image representation, edge detection, region-based segmentation, camera modeling, stereo vision, pose calculation, object recognition, optical flows, visual tracking, color vision, and beginning concepts of computational geometry. Students are expected to implement vision algorithms through programming assignments.

Required Text(s): None.

Recommended Text(s): None.

Lecture Outline:

Weeks Topic
1.0 1. Vision Psychophysics A. Vision as an inverse problems B. Assumptions in human visual perception
0.5 2. Connectivity and Distance Functions
1.0 3. Image Representation and Data Structure A. Run-length B. Quadtree C. MAT D. Chain-code E. Crack-code F. Skeleton
0.5 4. Border Following and Thinning
0.5 5. Component Labeling
1.5 6. Edge Detection A. Robert's B. Prewitt C. Sobel D. Laplacian E. LoG Operator - bias problems - false edges - fixes for these problems F. Edge Thinning G. Ridge Tracking
0.5 7. Hough Transformation A. Extraction of straight lines B. Extraction of circles
1.0 8. Region-based Segmentation A. Split-and-merge algorithm B. Samet's neighbor finding algorithm
0.5 9. Camera Modeling A. The pin-hole model B. The two-plane model
2.0 10. Stereo Vision A. Epipolar geometry B. Constraints C. Rectification
1.0 11. Pose Calculation A. Pose estimation from point correspondences B. Pose estimation using quaternions
1.0 12. Object Recognition A. Subgraph isomorphism B. Range Data - segmentation of range maps
1.0 13. Optic Flows and Analysis of Time-varying Imagery
0.5 14. Visual Tracking
1.5 15. Color Vision A. The trichromatic theory of color perception B. Color representation by RGB, HIS, and XYZ spaces C. Additive (RGB) and subtractive (CMYK) colors D. Object detection and tracking by color
1.0 16. Computational Geometry A. Transformation Groups - Affine - Similarity - Equiaffine - Euclidean - Projective - Relationships of the five groups B. Binary and greyscale morphology