Assignments

We will post the different assignments taken from Numerical Tours of Data Sciences here.

We will offer support for Python. You are free to work on your Assignments in Matlab, Julia, or R but you are more on your own.

To get the the recently tested Jupyter notebooks in Python, please clone the Numerical Tours from this fork. Follow the instructions in the README.md file to install required libraries.

Posted assignments:

Note: The homework notebooks come with a runnable script for the solution of each exercise. That is not taken away in order for you to get an idea of what the output of the solution should look like. The source code for the solution is also inside the repo. We don’t attempt to hide this as these numerical tours are extensively used and it would be easy for you to find these solutions regardless. For optimal learning, do not look at the source code of the solution and give it an honest attempt in writing your own source code from scratch. A large portion of the grade comes from responding the prompts that the exercises ask so make sure to respond them as clearly as you can.

Reminder: Make sure you pull the latest commits in the fork after each assignment is posted.

  1. Exercises 1 and 2 from Introduction to Image Processing, exercises 1-4 from Image Approximation with Fourier and Wavelets, and exercises 1-3 from 2-D Daubechies Wavelets.

    • Due date: Thursday, January 20, at 2:00 PM.

  2. Exercises 1-5 from Volumetric wavelet Data Processing, exercises 1 and 2 from Linear Image Denoising, exercises 1-3 from Wavelet Denoising, exercises 1-4 from Wavelet Block Thresholding, and exercises 1-6 from Stein Unbiased Risk Estimator.

    • Due date: Thursday, February 10, at 2:00 PM.

  3. Exercises 1-4 from Image Deconvolution using Variational Method, exercises 1-5 from Inpainting using Sparse Regularization, and exercises 1-6 from Performance of Sparse Recovery Using L1 Minimization.

    • Due date: Thursday, February 24, at 3:00 PM.

Handins

Make a script that produces the plots and carefully describes your findings. The code itself is part of the assignment; make sure that it is readable and carefully documented.

Generate a pdf or HTML file from your script. You may do this by exporting the Jupyter notebook that contains your answers/codes. Make sure that the lay-out is readable. Look for documentation on making up your code for publication.

The naming convention for this file is GTID_assignmentnumber.[pdf/html]. Send the file to Rafael Orozco with subject CSE 8803. For some exercises you will be asked to write separate functions. In this case include ALL the files in a zip-file named GTID_assignmentnumber.zip.