Welcome to the project for the practical course "Applied Optimization Method for Inverse Problems" offered during the summer term 2023 at TUM. This is the basis of the practical work and homework for the course.
The course website can be found here
The easiest and recommended way to install is using poetry (see
here). Once you've installed poetry,
run poetry install from the root directory.
Next, we miss the dependency on elsa (see
here), our tool for tomographic
reconstruction. First run poetry shell, which will activate the virtual
environment created by poetry, then clone elsa to a new directory,
move into the directory and run pip install . --verbose (the --verbose is
optional, but then you'll see the progress).
From now you can either active the virtual environment by running poetry shell,
or you run poetry run python myscript, which will active the environment for
that single command.
If you do not want to use poetry, you can use virtual environments.
From the root directory of this repository run python -m venv /path/to/venv, active
it using source /path/to/venv/bin/activate, and then install everything with
pip install --editable . (from the root directory).
Then again you need to install elsa. Follow the steps described above in
the poetry section.
If you have trouble installing elsa (how can that happen :D), see the README
of elsa. If you use an Ubuntu based distro and want to use CUDA, you might
need to set CUDA_HOME, to wherever CUDA is installed.
Please note, that you do not need CUDA, but it might speed up your reconstructions quite dramatically.
As generally, it might make certain things more pleasant, you can get access to one of our CUDA capable machines.
Please, send us a public key (see this for instructions to create one), and your RBG username. Then we will arange access to our machines. To access them you need to be either in a university WiFi (e.g. eduroam), or use a VPN (see here for configuration files).
There you need to clone your personal repository, do the same setup steps as described above.
Please note: You are sharing the resource with all of your students, so please
do not block all GPU's all the time! Check which GPU's are currently free
using nvidia-smi, and then use CUDA_VISIBLE_DEVICES=id to only
use one (and not the first).
To get the dataset for the challenge, head over to
Zenodo and download the
htc2022_test_data.zip. Extract it to a folder and you should be good to go.
In the folder you will see a couple of different files.
.mat files contain the actual measurements/sinogram, which will need for
reconstruction. There are the full measurements, one with limited number of
projections and example reconstructions of both. Further, there are segmented
files, which show the required binary thresholding segmentation done to
evaluate the score (again for full data and limited). Finally, there are a
couple of example images.
Then run main.py in the challenge subfolder, with the required argument
of the path to the folder you just extracted.