A project for the image processing course at BME.
Just run download_images.py, it will get all the images from the urls.
It will also create a new database file, with the file names.
Note: In the database there are duplicated rows.
TODO:
- need some way to see the accuracy of the model
Run license_plate_extraction.py
It is currently not the best solution, but it's acceptable.
TODO:
- transform license plate for the same horizontal orientation
- scale the images, so they are the same dimension
- segment license plate into the images of each character
TODO:
- create neural network
- train and test etc.