Tufts University
Department of Electrical and Computer Engineering
EE-133: Digital Image Processing
Spring 2008

Table of contents


Problem sets and solutions
 Lecture materials

Overview

The processing and analysis of multidimensional signals is an area of research and practice whose importance is growing rapidly both with the emergence of commodity imaging devices with resolutions undreamed of only as few years ago as well as with the development of state of the art sensors capable of collecting vast quantities of information across space and time.  Whether one is looking to stream higher resolution images and video over a wireless channel, exploit hyperspectral satellite image data to better quantify global climate change, or fuse information from multiple modalities to unlock the inner workings of the mind, fundamental methods in image processing play a crucial role.

In EE 133 this spring, we shall explore the basics of digital image processing.  The first portion of the class will focus on traditional ideas in digital image processing including linear shift invariant filtering, Fourier analysis, histogram methods for image enhancement, wavelets, and compression.  In the middle portion of the class, we shall examine topics more closely linked to use of image processing for problems of computer vision.  Here we will look at nonlinear filtering methods based on mathematical morphology as well as a collection of image segmentation techniques ranging from simple thresholding, iterative clustering, and finally active contours.  T

It is hoped that the class will be accessible and useful to students in all engineering, science, and mathematics disciplines where image processing is employed either as a tool or as a central area of research.  Thus, assignments for the class will offer a blend of theory as well as Matlab-based experimentation.  There will be a final project in the class as well as a single exam.  The only formal pre-requisite is EE 102, Linear Systems, although some familiarity with discrete signals and systems would be helpful.

Times and places

Staff

Prerequisites

If necessary, some time will be spent briefly reviewing linear systems ideas as well as probability.

Class requirements and preliminary grading

Texts

This is the required text for the class.  It is relevant for some, but by no means all of the material we will be covering.
Material for various parts of the class will also be drawn from the following three texts.  

Links

Preliminary Schedule


Date Topic Distributed/due
Jan. 17 Introduction and overview
Jan. 22 Fourier transforms in 1D and 2D
Jan. 24 Fourier transforms in 1D and 2D
Jan. 29 NO CLASS
Jan. 31 Image filtering: 2D convolution
Feb. 5 Image filtering: Smoothing and edge extraction
Feb. 7 Image filtering: Frequency domain
Feb. 12 Image filtering: Frequency domain
Feb. 14 Image enhancement
Feb. 19 Multiscale processing: Wavelet transform of images
Feb. 21 No Class
Feb. 26 Multiscale processing: Wavelet and denoising
Feb. 28 Multiscale processing: Denoising
March 4 Midterm
March 6 Compression: Intro
March 11 Compression: Lossless
March 13 Compression: Lossy
March 15 Morphology
March 25 Morphology
March 27 Morphology
April 1 Segmentation
April 3 Segmentation
April 8 Segmentation
April 15 TBD
April 17 TBD
April 22 TBD
April 24 TBD