ECE 418
Introduction to Image and Video Processing
Menu: Course View Options
Displaying course information from Spring 2014.
Section  Type  Times  Days  Location  Instructor 

NB1  LAB    Minh Do 

NL1  LEC  1230  1350  T R  245 Everitt Lab  Minh Do 
Web Page  http://courses.engr.illinois.edu/ece418 

Official Description  Concepts and applications in image and video processing; introduction to multidimensional signal processing: sampling, Fourier transform, filtering, interpolation, and decimation; human visual perception; scanning and display of images and video; image enhancement, restoration and segmentation; digital image and video compression; image analysis. Laboratory exercises promote experience with topics and development of C and MATLAB programs. Course Information: Prerequisite: ECE 310; credit or concurrent registration in one of ECE 313, STAT 400, IE 300, MATH 461; MATH 415; experience with C programming language. 
Course Prerequisites  Credit in ECE 310 Credit or concurrent registration in STAT 410 or ECE 313 or STAT 400 or MATH 461 or IE 400 Credit or concurrent registration in MATH 415 
Course Directors 
Pierre Moulin

Detailed Description and Outline 
To introduce students to both the fundamentals and emerging techniques in image and video processing. Topics:

Texts 
Lecture Notes (required) Recommended text: 
Course Goals 
This course is an elective. It introduces the fundamentals of image and video processing to seniors and has two dual goals: (1) provide students with a broad set of skills in this rapidly growing area, and (2) prepare students for further indepth study of this field. 
Instructional Objectives 
A. By the time of Exam No.1 (after 10 lectures), the students should be able to do the following: 1. Compute a twodimensional (2D) Fourier transform in both discrete and continuous spatial coordinates, and implement the 2D discrete Fourier transform. (a,b,k,m) 2. Identify an appropriate sampling resolution, given the 2D spectrum of a continuous image. (a,b,k,m) 3. Perform 2D finiteimpulseresponse (FIR) filtering of images. (a,b,k,m) 4. Design 2D decimation and interpolation schemes. (a,b,k,m) 5. Understand basic properties of the human visual system. (a,b,c,k,m) 6. Quantitatively evaluate image quality based on Frei and Baxter’s color vision model. (a,b,c,k,m) 7. Determine appropriate specifications for a chargecoupleddevice (CCD) camera. (a,b,c,e,k,m) 8. Determine appropriate specifications for the amplitude and spatial resolutions of an image display system based on the characteristics of the images and the contrast sensitivity and spatiotemporal modulation transfer function of the human visual system. (a,b,c,e,h,i,k,m) 9. Design appropriate specifications for a color quantization system based on the color vision properties of the human visual system. (a,b,c,e,h,i,k,m) 10. Design appropriate gamma correction techniques. (a,b,c,e,k,m) 11. Perform halftoning and error diffusion operations. (a,b,c,e,k,m) B. By the time of Exam No.2 (after 20 lectures), the students should be able to do all of the items listed under A, plus the following: 1. Know basic features of digital video standards for highdefinition television (HDTV), standarddefinition television (SDTV), videoconferencing (CIF) and videophones (QCIF). (a,b,c,e,f,h,i,j,k,m) 2. Select an appropriate format for various video applications. (a,b,c,e,h,i,j,k,m) 3. Design a point operation to improve contrast or modify the dynamic range of grayscale and color images. (a,b,c,e,h,i,j,k,m) 4. Perform histogram equalization and histogram specification on an image. (a,b,c,e,h,i,j,k,m) 5. Design an edgeenhancing linear filter. (a,b,c,e,h,i,j,k,m) 6. Design lowpass and median filters for reducing the noise level in an image without drastically affecting image contents, and select an appropriate filter depending on noise characteristics. (a,b,c,e,h,i,j,k,l,m) 7. Derive and implement the linear Wiener restoration filter, given a measurement model and image and noise statistics. (a,b,c,e,h,i,j,k,l,m) 8. Derive and implement an adaptive Wiener filter. (a,b,c,e,h,i,j,k,l,m) 9. Estimate the entropy of a sequence of statistically independent symbols. (a,b,c,e,h,i,j,k,l,m) 10. Apply the concept of entropy to estimate the bit rate of a lossless image coder. (a,b,c,e,h,i,j,k,l,m) 11. Perform runlength coding of a binary image and bit plane encoding of a grayscale image. (a,b,c,e,h,i,j,k,l,m) 12. Perform lossless predictive coding of a grayscale image. (a,b,c,e,h,i,j,k,l,m) 13. Derive the KarhunenLoeve transform based on secondorder image statistics. (a,b,c,e,h,i,j,k,l,m) 14. Perform lossy predictive coding and transform coding of images. (a,b,c,e,h,i,j,k,l,m) 15. Know basic features of the JPEG compression standard (baseline, lossless, progressive and hierarchical modes) and wavelet image compression. (a,b,c,e,h,i,j,k,l,m) C. By the time of the Final Exam (after 28 lectures), the students should be able to do all of the items listed under A and B, plus the following: 1. Perform motion compensation of video sequences using meansquarederror and meanabsoluteerror block matching criteria, and full or fast search techniques. (a,b,c,e,h,i,j,k,l,m) 2. Perform motioncompensated predictive coding of video using forward, backward, and bidirectional predictive methods. (a,b,c,e,h,i,j,k,l,m) 3. Select an appropriate encoding method for each macroblock in a video sequence. (a,b,c,e,h,i,j,k,l,m) 4. Know basic features of MPEG2, MPEG4, and H.264 video compression standards. (a,b,c,e,h,i,j,k,l,m) 5. Implement an edge detection algorithm. (a,b,c,e,h,i,j,k,l,m) 6. Represent a region boundary using chain codes and Fourier descriptors, and evaluate the effects of geometric image transformations on Fourier descriptors. (a,b,c,e,h,i,j,k,l,m) 7. Select appropriate features for image segmentation. (a,b,c,e,h,i,j,k,l,m) 