Lab 3 project on DTU by Benjamin Banks(s234802), Bjørn Hagbarth(s234870) and Tobias Rodrigues Bjerre(s234823)
You need to have Python3 installed.
When installed, run the following to install the required python packages:
pip install -r requirements.txt
Run the acutal script by running:
python main.py
- Do K-nearest once on the image
- Use the edge_score to calculate the inital backgroundcentroid
- Find the oposite, and make it a foregroudcentroid
- Run K-nearest n number of times.
- The Python Image Preview VSCode extention can be used.
- The tests have been written using python unittest. It is helpfull to run the tests in debug mode.
- The modules are code written to do the unsupervised learning.
- The services are meant support thing, handeling stuff like loading and converting.
- If no dimension is writen, 1d is implied.