Week 1	Introduction to recognition, matched filtering
Topics	Recognizing objects and object classes in images. Invariance with respect to deformations, occlusion, viewpoint and lighting conditions. Detection and recognition using matched filters. Designing multi-dimensional matched filters using principal component analysis.
Homework	Prove facts on matched filtering. Experiment on a simple recognition problem using matched filtering. Come up with some new idea for recognition.

Week 2	introduction to constellation modelling, detection theory
Topics	Limits of image-based approaches. Modelling geometry explicitly. Ideas on achieving translation-rotation-scale invariance in geometry and photometry. Modelling background noise and threshold design. Representation versus discrimination. Elements of learning theory and decision theory. Gaussian classifiers. Linear discriminant analysis. Fisher linear discriminants.
Homework	Building and experimenting with n-dimensional matched filters. Construct ROC curves. Compare 1D and nD matched filters' performance.

Week 3	Probabilistic constellation modelling
Topics	Single-object single feature. Modelling the detectors' output. Modelling the probability of the background events with Poisson densities. Probability of object present. Labelling. Ratios of probabilities of events. The labelling problem is local in the features.
Homework	Using Fisher discriminants. Gaussian classifiers. Representation vs discrimination.

Week 4	Probabilistic constellation modelling cont'd
Topics	Extension of constellation models to objects composed of multiple features. Comparing the likelyhood of partial constellation hypothesis to full-constellation hyp.. Including the detectors' response in the model. Discussion of options for obtaining translation invariance. Brief introduction to generic feature detectors: the Lucas-Tomasi-Kanade feature detector; the Foerstner interest operator for finding conrners and circles.
Homework	Building an object-recognition program using generic feature detectors and constellation models.

Week 5	Feature detectors
Topics	Tuned detectors. Support vector machines. PCA-trained multidimensional features. Neural network feature detectors.
Homework

Week 6	Feature Detectors
Topics	Generic detectors. Filter banks. Corner, point, edge detectors.
Project

Week 7	Automatic model-building
Topics	Expectation Maximization (EM). Automatic model building.
Project

Week 8	Overview of the literature
Topics	Recognition from contours, color histogram, principal components.
Project

Week 9	Recognition in the human visual system
Topics
Project

This is a tentative syllabus. It will be updated every week or so. Check this page often.