You are on page 1of 1

ECE 4051 COMPUTER VISION [3 0 0 3]

Total number of Hours: 36


Course objectives

At the end of this course, the student will be able to:


1. Describe image formation using pinhole camera and explain preprocessing techniques.
2. Analyze different local image features such as SIFT, HOG and texture.
3. Contrast different image segmentation methods based on clustering, model fitting and
probabilistic modelling.
4. Explain the geometry of two view and discuss the least square estimation method for
camera calibration with radial distortion.
5. Evaluate various techniques for building classifiers with examples of their use in vision
applications.

Introduction:
Pinhole cameras, Image formation, Sources: Point, line and area sources, Linear filters and
convolution, spatial frequency and Fourier transforms, sampling and aliasing, Filters as
templates, Estimating derivatives of image, edge detection using LoG. [6]

Image features:
Finding corners, Using scale and orientation to build neighborhood, SIFT features, HOG features
Texture- Representing textures, analysis and synthesis using oriented pyramids [8]

Segmentation:
Segmentation using clustering: K-means, mean shift, watershed; Segmentation by fitting a model
(Fitting lines, curves), fitting as a probabilistic inference problem; Segmentation and fitting using
probabilistic methods: Expectation-maximization algorithm. [10]

Geometry of two view and Camera calibration:


Homogenous coordinates, Epipolar geometry, Fundamental Matrix, Image rectification, Camera
parameters, Least square parameter estimation, taking radial distortion into account [4]

Pattern Classification:
Feature selection, Bayes Classifier: using class histograms, using class conditional density,
Perceptron, Multi-layer perceptron, Support Vector machine [8]

TEXT BOOK
1. David A. Forsyth and Jean Ponce, Computer Vision: A Modern Approach, Pearson
Education, 2003

References:

1. Richard Szeliski, Computer Vision: Algorithms and Applications, Springer, 2010


2. Richard Hartley and Andrew Zisserman, Multiple View Geometry in Computer Vision,
2nd Edition, Cambridge University Press, 2004
3. Linda Shapiro and George Stockman, Computer Vision, Pearson Education, 2001

You might also like