PrivPresLoc

Adam Taras 2022. Work done under Donald G. Dansereau as part of the Robotic Imaging Group, The University of Sydney

This git repo contains a MATLAB implementation of a privacy preserving pipeline.

Requirements

We use a modified imds to allow for cleaner dataset managment. Need to change line in bagOfFeatures to not check type on input parser.addRequired('imds', @(x) true);

Datasets

Datasets can be found on Google Drive

Structure

Key folder:

• localisationPipeline: which contains functions for different feature extractors and functions for testing/evaluating them on a given dataset.

Filename convention:

• m - main file
• x - experiment
• d - modify dataset
• t - function tester