In order to install the system is sufficient to clone the repository to a local folder.

mkdir mmr_project
cd mmr_project
git clone git@github.com:Olu93/project_multimedia_retrieval.git

When downloaded the structure of the project should look like indicated in structure.txt.

The project requires different packages to be installed, this can be done either using Anaconda or pip package managers. From the project directory run the following.

Anaconda:

conda env create -f environment.yml
conda activate mmr_system_env

Pip:

pip install -r requirements

The databases needed for the system to run on the Princeton shape dataset are included in the repository. To initialise the correct paths to each just run:

python initialise.py

This should create defaults value for every parameter and prompt for further action or exit. This script also aims to ask for classification files in case the user has not specified any yet. Classification files should be of type "*.cla" like the ones present in the psb dataset, if no file is specified and the normalisation is run again, the system will label every mesh as "no_class". If the aim is to run the system, the user should now be able to run it with the defaults values set. Otherwise is possible to further change the settings as described in Setting new paths section below. Once the defaults are set, is possible to finally run the system itself. To do so, type:

python gui_maker.py

This should open the upload mesh windows, here is possible to select a mesh from the explorer or drop it onto the window. This will activate the query interface. In this second window is possible to set some parameters before querying. The best evaluated parameters and distance function combinations are the default values.

Once the initialise.py script is run, is also possible to set new values (for instance this could be setting a new database and compute normalisation and feature extraction). To do so, when prompted with exit or continue, press 2 to continue and visualize the option available. This gives the option to further change paths to files or even run a normalisation and feature extraction pipeline on a another shape dataset. If the latter action wants to be performed, the user should make sure to have set a new database path. The database path can be the root folder of the database (in this case if classification files are present the system will pick them up). The system will recursively search all subfolders starting from the path specified, looking for any *.off or *.ply file.