This repository contains a few test projects to test how to integrate an ML model for anonymizing images into an Android App. This projects serves as a template for the integration within the Cyface App itself and was initially developed to integrate the Anonymization developed on University of Leipzig in the DiGuRaL research project into Cyface Image Capturing Technology.
The first try of getting an ML model to run within an Android App. This project has rather wild code and several trys using different technologies.
This is a test project to get a quick solution using an exisitng Tensorflow lite model, bot not the one from Leipzig. It was developed to get a fast solution for the first test actual test run.
A second try to integrate the model from the University of Leipzig with an Android App. This version tries to reach a much cleaner implementation based on the lessons learned from Test1 and Test2.
Test3 contains a few anonymization models in its asset folder.
All those models have been create by University Leipzig and are based on
YOLO v8.
All of these models may be used by changing the modelPath in MainActivity.