EEC106 - Image Processing And Computer Vision

4 units - Spring Quarter

Lecture: 3 hours

Laboratory: 3 hours

Prerequisite: course 150B

Grading: Letter; homework (15%), lab (25%), midterm (20%), final (40%)

Catalog Description: Imaging geometry; transforms and sampling; enhancement, restoration, and conversion; image compression; time-varying image analysis; elementary pattern recognition; segmentation; multiresolution analysis.

Course Outcomes:
Students who have successfully completed this course should:

Course Outcomes Program Outcomes
Understand the mathematical and statistical principles forming the basis for methods of image processing and computer vision. 1a
Be able to apply mathematical and statistical principles to solve problems in image processing and computer vision involving image geometry, transforms and sampling enhancement, converison, compression, pattern recognition, segmentation, and multiresolution analysis. 1b
Be able to develop algorithms to solve a problem, to determine which algorithm is the most appropriate to apply, to develop and debug Matlab-based algorithms to solve problems, and to interpret the results. 2a, 2b
Be able to apply image processing design methodologies and software tools, and to choose among alternative approaches to analysis and design, including space-domain, frequency-domain, point processing, edge-based, region-based, or multiresolution. 3a, 3b
Be able to design and implement an algorithm to achieve specified performance in filtering for noise reduction, detection of edges, compression, and segmentation. 4a
Be able to develop the technical specifications required to accomplish a desired function of noise reduction, edge detection, compression, or segmentation 4b

Expanded Course Description:

  1. Geometric Imaging Models
    1. Perspective-Projective Transform
    2. Stereo (Multiple Camera) Vision
    3. Geometric Transformations
    4. Applications
  2. Transforms and Sampling (2-D)
    1. Linear Systems and Discrete Transforms
    2. Convolution and Correlation
    3. Sampling Effects in the Transform Domain
    4. Temporal Sampling Effects
  3. Enhancement, Restoration, and Conversion
    1. Discrete Linear Operators
    2. Nonlinear Operators
    3. Edge Detection
    4. Conversion of Grayscale to Binary Images (thresholding and halftoning)
  4. Image Compression
    1. Elementary Compression
    2. Transform Coding
    3. Differential Techniques
  5. Time-Varying Image Analysis
    1. Difference Images
    2. Moving Edge Detection
    3. Optical Flow
  6. Elementary Pattern Recognition
    1. Statistical Pattern Recognition
    2. Syntatic Pattern Recognition
  7. Segmentation
    1. Edge-based Approaches
    2. Region-based Approaches
    3. Matching for Recognition
  8. Multiresolution Analysis
    1. Hierarchical (Pyramidal) Analysis
    2. Scale-space Filtering
    3. Quadtrees
Textbook: R. C. Gonzalez and R. E. Woods, Digital Image Processing , Addison Wesley, 1993.

Computer Use: Matlab and the Matlab Image Processing Toolbox on the ECE instructional workstations are used for all laboratory projects.

Laboratory Projects:

  1. Image processing with Matlab
  2. Linear and nonlinear filtering for noise reduction
  3. Enhancement and detection of edges
  4. Image compression
  5. Image segmentation
Engineering Design Statement:
The laboratory projects examine design at both the level of individual functions (e.g., the design of 2-D filters) and the system level. In addition to design methodology, criteria for the selection of algorithms appropriate to a given application are stressed. Designs are implemented in software, via a visual programming language which emphasizes a modular approach to problem solving. Projects do not have unique solutions, and students are encourage to explore alternative approaches.

Professional Component: Engineering Depth, Laboratory
Engineering Science: 2 units
Engineering Design: 2 units

Instructors: Reed, Ford, Levy
Prepared by: Reed, Ford
Revised: 6/01