Signal and Information Processing

Transcription

1 Signal and Information Processing Alejandro Ribeiro Dept. of Electrical and Systems Engineering University of Pennsylvania January 11, 2017 Signal and Information Processing Signal and Information Processing 1

2 Presentations Presentations Signals and Information Class contents Signal and Information Processing Signal and Information Processing 2

3 Me Alejandro Ribeiro Associate Professor, Dept. of Electrical and Systems Engineering Levine 362, My research is on statistical signal processing We meet on Towne 100 Mondays, Wednesdays, Fridays 9 am to 10 am Signal and Information Processing Signal and Information Processing 3

4 Lab sessions Four lab sessions 1 hour on Tuesday + 1 hour on Thursday Labs are computer based. We expect you to work on your own. Lab sessions are for guidance But we expect you to be there (mandatory assistance) Labs start next Tuesday Signal and Information Processing Signal and Information Processing 4

5 Prerequisites A lot of things that you already know and a couple you may not Linear algebra Vector matrix notation, systems of linear equations, eigenvalues We will cover them, but it helps if you know Programming in Matlab Needed for Lab assignments. If you know programming you can learn Matlab in one afternoon But it has to be this afternoon Signal and Information Processing Signal and Information Processing 5

6 Midterm, final, and grading First midterm will be in class, during the lecture on Friday, March 3 Worth 26 points Second midterm in class, during the lecture on Wednesday, April 26 Worth 26 points. Cumulative. You will be handing in 13 lab reports. Graded from 0 to 4 No show (0). Poor (1). Fair (2). Good (3). Excellent (4) Maximum possible points: 104 = You pass with 60 points and 15 points in midterm and final (each) C requires at least 70 points. B at least 80. A at least 90 Don t worry. We will work hard. We will learn. We will get A s. Signal and Information Processing Signal and Information Processing 6

7 Textbook Textbook for the class is M. Vetterli, J. Kovacevic, and V. K. Goyal Foundations of Signal Processing, Cambridge University Press, August 2014 Available online We will not follow the order of the book, but I will assign readings Signal and Information Processing Signal and Information Processing 7

8 Difficulty Been told this class has a reputation for been hard and demanding This is false Difficulty ratings are 3.4/4 On the other hand, quality ratings are very good They are 4 out of 4 among students who give four points Two nominations for best college class (me and my mom) A rating of 3.06/4 among all students This is, really, a great class. You will do things that look like magic Also, the class is front loaded. It will become easier after the break. Don t drop it! You will enjoy it. Signal and Information Processing Signal and Information Processing 8

9 Signals and Information Presentations Signals and Information Class contents Signal and Information Processing Signal and Information Processing 9

10 Signals and Information A collection of values (numbers) that are important to us A conversation recording, an image, car fuel efficiency People tend to say signal when talking about physical quantities And information when talking about virtual quantities But not always, and in any event, irrelevant All are functions Signal and Information Processing Signal and Information Processing 10

11 Signals are functions One dimensional signal. Given parameter t specify values x(t) Time can be continuous (left) or discrete (right) Spring day in Philadelphia Spring day in Philadelphia Temperature ( o F) Temperature ( o F) Time of day (hours) Time of day (hours) Continuous time signal function x(t), with t R (reals) Discrete time signal function x(n) with n Z (integers) In either case can have subsets, e.g., t [0, 24] and n = 0,..., 24 Signal and Information Processing Signal and Information Processing 11

12 Two dimensional signals We can also define a signal in 2 dimensions. E.g., an image Gien parameters x and y specify values f (x, y) The parameters may be continuous (left) or discrete (right) Heat propagation in a room Heat propagation in a room Temperature ( o F) x y Temperature ( o F) x y Continuous support signal function f (x, y) with x R, y R, Discrete support signal function f (x, y) with x Z, y R Z Signal and Information Processing Signal and Information Processing 12

13 Multidimensional signals Can have three variables Coordinates of three-dimensional space x, y, x (a field) Two physical dimensions x, y and time t (a video) Or any arbitrary number of dimensions We want to do many interesting things with x(t) All starts from understanding how fast x(t) changes. Bear with me Signal and Information Processing Signal and Information Processing 13

14 Class contents Presentations Signals and Information Class contents Signal and Information Processing Signal and Information Processing 14

15 Lecture schedule Class organized in two blocks Theory and design (sort of) Theory block 20 lectures and 7 labs From Friday 1/15 to Wednesday March 2 (up to Spring break) Midterm coming right after on Friday, March 4 Design block 18 lectures and 6 labs From Monday 3/14 to Friday April 22 (up to the summer break) Cumulative final coming right after on Wednesday, April 27 Signal and Information Processing Signal and Information Processing 15

16 Theory block One dimensional signal processing Discrete signals. Discrete Fourier Transform. Linear systems Continuous signals. Sampling. Various labs converging towards voice recognition Signal and Information Processing Signal and Information Processing 16

17 Design block Use one-dimensional theory to build extensions and design systems Image processing JPEG compression Principal component analysis Face recognition Graph signal processing cancer type classification We begin on Friday... Signal and Information Processing Signal and Information Processing 17