Day 1 – Monday, 21st of March
Lectures to be held in Päätalo Auditorium D11, Tampere University, City Centre campus
Signal Processing
8:50 – 9:00 Opening words, overview of the school training program, instructions about lab works, overview of social program, etc.
9:00 – 10:00 Lecture 1A (Alessandro Foi – Tampere University): Basics of vector-space formalism for linear approximations: from orthonormal bases to redundant frames for global and local weighted least-squares approximations. Coupled frames and principled design of operators and (convolution) kernels for the interpolation, resampling, smoothing, and differentiation of regularly and irregularly sampled data. Matlab assignment: Convolution kernels for multivariate differentiation.
10:00 – 10:20 Coffee Break
10:20 – 11:20 Lecture 1B (Alessandro Foi – Tampere University): continuation of Lecture 1A
11:20 – 11:30 Break
11:30 – 12:40 Lecture 2 (Giacomo Boracchi – Politecnico di Milano): Redundant representations and Sparse Coding (Minimum l0 norm). Sparse Coding w.r.t redundant dictionaries. Greedy Algorithms: Matching Pursuit, Orthogonal Matching Pursuit. Matlab assignment: Minimum l0 norm sparse coding
12:40 – 13:40 Lunch Break
13:40 – 14:50 Lecture 3 (Giacomo Boracchi – Politecnico di Milano): Dictionary Learning. Dictionary Learning Algorithms: Gradient Descent, MOD, KSVD. Matlab assignment: Dictionary learning
14:50 – 15:10 Coffee Break
15:10 – 16:20 Lecture 4 (Giacomo Boracchi – Politecnico di Milano): Convex Relaxation of the Sparse Coding Problem (Minimum l1 norm). Sparse coding as a convex optimization problem, connections with minimum l0 solutions. BPDN and the LASSO, the two formulation and promoting l1 sparsity as a regularization prior in linear regression. Other norms promoting sparsity: visual intuition.
Social programme: 18:00 Dinner in Finlayson Palace
Day 2 – Tuesday, 22nd of March
Lectures to be held in Tampere University, Hervanta campus, lecture room TB109
Demos / Exercises to be held in Hervanta campus, meeting room SL401
Physics and Optics
9:00 – 9:40 Lecture 1 (Giacomo Boracchi – Politecnico di Milano): Minimum l1 Sparse Coding Algorithms: Iterative Soft Thresholding.
9:40 – 9:45 Break
9:45 – 10:30 Lecture 2 (Giacomo Boracchi – Politecnico di Milano): Minimum l1 Sparse Coding Algorithms: Iterative Reweighted Least Squares. Matlab assignment: Minimum l1 norm sparse coding
10:30 – 10:50 Coffee Break
10:50 – 11:30 Lecture 3 (Humeyra Caglayan – Tampere University): Fourier Approach to Wave Propagation
11:30 – 11:35 Break
11:35 – 12:20 Lecture 4 (Humeyra Caglayan – Tampere University): Approximations: Ray optics, Fresnel and Fraunhofer Diffraction
12:20 – 13:20 Lunch Break
13:20 – 14:05 Lecture 5 (Humeyra Caglayan – Tampere University): Imaging: Coherent, incoherent imaging, PSF, MTF
14:05 – 14:35 Coffee Break
14:35 – 16:15 Demo/Exercise (Humeyra Caglayan / Erdem Sahin – Tampere University) combined with Lab visit
Social programme: ~18:00 Finlayson Escape Room + ~19:45 Dinner in pizzeria Luca
Day 3 – Wednesday, 23rd of March
Lectures to be held in Scandic City, lecture hall Ratina 1–2
Multi-view Geometry and Light Field Imaging
9:00 – 09:40 Lecture 1 (Sebastian Knorr – Ernst Abbe University of Applied Sciences Jena): Multi-view geometry. Transformations. Homography. Camera motions. Epipolar geometry.
09:40 – 09:50 Break
09:50 – 10:30 Lecture 2 (Sebastian Knorr – Ernst Abbe University of Applied Sciences Jena): 3D reconstruction from two views. Features. F-matrix. Rectification. Structure from Motion.
10:30 – 11:00 Coffee Break
11:00 – 11:40 Exercise 1A (Sebastian Knorr – Ernst Abbe University of Applied Sciences Jena): Matlab exercise on multi-view geometry
11:40 – 11:50 Break
11:50 – 12:30 Exercise 1B (Sebastian Knorr – Ernst Abbe University of Applied Sciences Jena): Matlab exercise on multi-view geometry
12:30 – 13:30 Lunch Break
13:30 – 14:30 Lecture 3 (Erdem Sahin – Tampere University): Physically realizable light field; how to relate rays to waves
14:30 – 14:40 Break
14:40 – 15:25 Exercise 2 (Alessandro Foi and Giacomo Boracchi): Individual work on Matlab assigments
15:25 – 15:40 Coffee Break
Social programme: Outdoor activity (depending on the weather)
Day 4 – Thursday, 24th of March
Lectures to be held in Maisansalo
Signal-Dependent and Correlated Noise in Imaging
8:15 – 9:00 Joint transportation from Scandic City to Maisansalo
9:00 – 09:45 Lecture 1 (Alessandro Foi – Tampere University): Signal-dependent noise models; one-parameter families of distributions; heteroskedasticity; white vs colored noise; variance function, response function
09:45 – 09:55 Break
09:55 – 10:40 Lecture 2 (Alessandro Foi – Tampere University): Noise models for raw data.
10:40 – 11:00 Coffee Break
11:00 – 11:45 Lecture 3 (Alessandro Foi – Tampere University): Estimation of noise-model parameters; theoretical aspects; methods for noise variance-mean curve estimation and fitting; estimation under saturation and clipping. Models for correlated signal-dependent noise.
11:45 – 11:55 Break
11:55 – 12:40 Lecture 4 (Alessandro Foi – Tampere University): Variance stabilization; variance stabilizing transformations (VST); inverse VSTs; application to image restoration.
12:40 – 14:00 Lunch Break
14:00 – 15:30 Discussion and feedback on the Matlab exercises (Alessandro Foi, Giacomo Boracchi)
15:30 – 16:00 Panel Discussion on School Outcomes (Atanas Gotchev)
Social programme: Sauna and dinner in Maisansalo
21:15 – 22:00 Joint transportation from Maisansalo to Scandic City
Day 5 (ONLINE) – Wednesday, 30th of March
09:00 – 09:55 (CET) Lecture 1 (Donald Dansereau, University of Sydney): Introduction to Light Fields (LF): Light as a field, parameterizations, visualizations, sampling & calibration. A simple interpolating renderer.
10:00 – 10:30 (CET) Exercise 1 (Donald Dansereau, University of Sydney): Guided hands-on handling LFs in MatLab: Live demonstration of loading and visualising a LF in Matlab. Distribution of take-home exercises (programming own interpolating renderer).
10:15 – 11:15 (CET) (optional) Assistance for getting started on the assignment. (Donald Dansereau, University of Sydney)
Day 6 (ONLINE) – Thursday, 31th of March
09:00 – 9:55 (CET) Lecture 1 (Donald Dansereau, University of Sydney): Key characteristics / capabilities: Parallax, LFs in the frequency domain, basic EPI image analysis. (Donald Dansereau, University of Sydney)
10:00 – 10:30 (CET) Exercise 1 (Donald Dansereau, University of Sydney): Guided exercise review: Interpolating renderer solution. (Donald Dansereau, University of Sydney)
End of the Training School