a. Geometry
i. Ideal perspective projection
ii. Hint of stereo
iii. Real lenses and cameras
b. Image as continuous function
i. Example linear operations - convolution
ii. General image properties: brightness, contrast frequency
iii. Noise vs signal - SNR
iv. Example: SNR vs Contrast
Digression: Human vision
c. Digitization
i. Sampling in space
ii. Quantization, dynamic range
iii. Integer vs continuous values
d. Vector images
i. color
ii. z-buffer
c. Pre-processing
a. Global Photometric
i. Gain normalization
ii. Histogram equalization
b. Geometric
i. Re-sampling
ii. Sub-sample vs interpolation
iii. Interpolation methods
c. Pyramids
i. Gauss reduction -theory
ii. Nyquist rates
iii. Burt pyramid
iv. Quad trees
d. Deformations/mappings
i. Notion of operator
ii. Planar models: Affine, perspective
e. Filtering
i. Blurring
ii. Other local operators - reduce information
iii. Non-linear: median
iv. Structure sensitive enhancement (Freeman)
f. Edge detection
i. Canny, Marr &Hildreth,
d. Binary Vision
a. Generation
i. Thresholds – absolute and dynamic
b. Binary images, morphology
i. dilation and erosion
ii. Morphological Matching
iii. skeletons
c. Region properties
i. Euler numbers,
ii. Moments
e. Finding structure
a. Hough transform
i. Lines
ii. Generalized
b. RANSAC
c. Other machine vision techniques
d. Case study: Face recognition
f. Image motion
a. Optic flow - brightness constraint
i. Kanade etc method
ii. Hierarchical algorithms
iii. Parametric methods
g. 3D Vision
a. Projective geometry
i. Calibration
ii. Essential and Fundamental matrices
b. Stereo
i. Random-dot
ii. SSD
iii. Correlation, occlusion, DP
c. Radiometry and 3D vision
i. Surface reflectance
ii. Shape from shading
iii. Photometric stereo
h. Camera and structure motion
a. Egomotion in velocity space
b. Recovery of egomotion
i. Translation only - FOE
ii. Full - Heeger and Jepson??
i. Model-based recognition
a. 3D models
i. Goad's algorithm
ii. Interpretation trees
b. Class based models
i. Active shape models
j. Video Analysis
a. Representing video
i. Layers
ii. MJPEG, MPEG
b. Tracking
i. Kalman methods
ii. CONDENSATION-like methods
iii. Structural hacks
c. Sequence analysis
i. HMMs
ii. Gesture recognition
d. Watching people
i. Limb tracking with weak models
ii. Limb tracking with strong models