Title: INTRODUCTION TO IMAGE PROCESSING
Digital images. Sampling and quantization of images. Color, stereo and video images. Arithmetic operations, gray scale manipulations, distance measures, connectivity. Image transforms. Linear and nonlinear filters. Image enhancement. Image restoration: degradation models, inverse filtering. Image segmentation. Image representation and description techniques.
Coordinator: Bülent Sankur, Professor of Electrical Engineering
Goals: This course aims to introduce the students to the understanding and processing of digital images. Thus concepts and tools from digitization of images via sampling and quantization to their manipulations via arithmetic, logical operations and linear operators are covered. Intermediate level representations using contour extraction and segmentation algorithms and concomitant image representations are investigated. Images in multimedia environments are discussed in the context of compression and statistical models.
At the end of this course, students will be able to:
Be able to manipulate and process images in the computer.
Have an understanding of the nature and statistical characterization of images.
Execute various low-level and some intermediate-level algorithms for image enhancement, restoration and compression
Be capable of combining image processing tools for solving vision problems
Understand compression algorithms and their role in multimedia.
Textbook: R.C.Gonzalez, R.E.Woods, Digital Image Processing, Addison-Wesley, 2008.
1) W.K.Pratt, Digital Image Processing, Wiley, 1991.
2) A. Jain, Two-Dimensional Signal and Image Processing, 1991
Prerequisites by Topic:
Signals and systems
Familiarity with Matlab
1. Introduction: Needs for digital processing of images. Types of images. Imaging requirements (one week)
2. Elements of digital image processing: Image acquisition. Image storage and databases. Image display. Image communication. (one week)
3. Pixels: Sampling and quantization. High resolution imaging. Pixel relationships. Data structures for images. (one week)
4. Binary Images: Binarization techniques. Morphological operations. Opening, closing, skeletonization, thinning. Morphological filtering. (one week)
5. Image Enhancement: Image enhancement principles. Point processing. Histogram equalization. Spatial filtering. Frequency-domain enhancement. Homomorphic filtering. Ranking operations. Median filtering. (two weeks)
5. 2D Linear Systems: Linear shift invariant systems. The Fourier transform. 2-D DFT and FFT algorithms. (one week)
6. Image Transforms and Filtering: Correlation. Separable kernel transforms. Walsh transform, Hadamard transform, Haar, slant, cosine transforms. Optical realizations. (one week)
7. Edge Detection: Edge, line, contour, arc, boundary. Derivative based methods. Marr-Hildreth paradigm. Edge detection performance. (two weeks)
8. Segmentation: Binarization and multithresholding. Measurement space based methods. Region growing methods. Edge based methods. Physics based methods. (one week)
9. Image Coding: Compression of images. Vector quantization. Predictive coding. Transform coding. Subband coding schemes. (three weeks)
10. Color: Color fundamentals. Color models. Color image processing. Pseudo-color. (one week)
Course Structure: The class meets for three lectures a week, each consisting of two 50-minute sessions. 8-9 sets of homework problems are assigned per semester. There are two in-class mid-term exams and a final exam. Each student must also prepare a term report, develop a software to bring into realization a realistic image processing solution.
Computer Resources: Students are encouraged to use MATLAB to solve their homework problems..
Laboratory Resources: None.
1. Project sets (50%)
2. Final (25%).
3. Term project (25%)
(a) Apply math, science and engineering knowledge. This course requires linear system theory, applies notions of transform domain processing and some elementary probability concepts. It requires some understanding of devices for image acquisition and an understanding of the relevance of multimedia.
(c) Design a system, component or process to meet desired needs. The students have to design several low-level image processing algorithms, from morphological tools for noise removal to image enhancement, from image interpolation for resizing to edge extraction and image segmentation, and to image compression algorithms.
(d) An ability to function on multi-disciplinary teams. The students are teamed in groups of two, sometimes of three in their term project to carry out complementary parts of the project.
(d) An ability to communicate effectively. The students must present effectively their term papers; their presentation performance is a determining factor in the final grade. Similar concerns are valid for the EE490 Senior Project.
(k) Use of modern engineering tools. Students use Matlab and a number of MATLAB packages for their homework assignments.
Prepared By: Bülent Sankur