Class (44 hours), lab (12 hours), project (50 hrs) and outside preparation Students are required to attend class and lab for a total of 6 hrs/week
Syllabus Students will learn basic tools for analysing 1-D and 2-D signals in the space and in the frequency domains. Particular attention will be devoted to filters, to deal with noise attenuation and feature enhancement. Dynamic filters will also be considered. The course will also cover low level vision topics, including image feature extraction, image segmentation,image registration, and image matching. Students will be involved in project activities. CONTENT: Systems - Systems: Input/Output signal, Response of a System - Systems Properties: Linearity, Time-invariance, Causality 1D signals - Complex numbers, Periodic Functions, Complex Functions, Trigonometric Polynomial - Fourier Series - Fourier Transform - Noise - Sampling, Sampling Theorem - Convolution Theorem - Filters - Kalman Filter - Wavelets 2D signals - Greyscale Images - Color images - Histogram - 2d Fourier Transform - Spatial filters - Image features (corner, edge, ridge) - Image matching - Image similarity measures
Signals and Systems Oppenheim et al. - Signals and Systems Bertoni et al. - Introduzione all’elaborazione dei segnali (UniMi) (in Italian)
Digital Signal Processing Orfanidis - Introduction to Signal Processing Oppenheim et al. - Discrete-Time Signal Processing
Signal & Image processing Gonzales Woods - Digital Image Processing Mallat - A wavelet tour of signal processing
Ricevimento: Appointment by email
Ricevimento: Send a mail to patrizia.boccacci@unige.it to make an appointment
The final evaluation will take into account: (1) class attendance, (2) planned homework, (3) project discussion (4) oral dissertation
The project should be written clearly, complemented with working code and it should show that the student has fully understood the topic. Examples on different real scenarios are encouraged.
The oral examination consists in a discussion of the project and of the topics taught in class.
