ece 180: image and video restoration · 2020. 9. 15. · ece 188: image and video restoration...
TRANSCRIPT
ECE 188: Image and video restoration
Course outline
I Introduction to inverse problems in image/video restoration contexts:denoising, deblurring, super-resolution, tomography, compressed sensing
I Fundamentals of linear/local filtering:maximum likelihood, spatial averaging, heat equation, low-pass and Wiener filtering
I Basic of non-linear filtering:signal adaptation, maximum a posteriori, wavelets and sparsity, non-locality, patches
I Towards advanced filtering:dictionary learning, convex and non-convex optimization, parameter selection
I Lab work covering the implementation of such techniques in Matlab.
0 0.2 0.4 0.6 0.8 1
−0.2
0
0.2
0.4
0.6
0.8
1
Position index i
Val
ue x
i
Interval ± σVector x0
Threshold λ
=
+?
Prerequisites
I Linear algebra (MATH 18)
I Differential calculus (MATH 20C)
I Probability and statistics (ECE 109)
I Fourier transform (ECE 161A)
I Basics of optimization (ECE 174)
I Matlab programming
Project – blind restoration challenge
I Analysis of a corrupted corpus of images
I Mathematical modeling of the inverse problem
I Elaboration of a restoration technique
I Implementation in Matlab
I Detailed technical report with bibliography
I Evaluation: originality/performance/quality