In this project, a software solution in the form of an app was developed, hereinafter also referred to as “Keras-RetinaNet”. With the app it is possible to classify license plates. The most accurate object detectors to date are based on a two-step approach popular with R-CNN. In contrast, single-stage detectors that are applied through a regular, dense scan of possible object positions may be faster and easier, but have so far compromised the accuracy of two-stage detectors.
Connect to Google
Here you can connect your file to Google
Check access to graphics card
With! Nvidia-smi you can check the access to the graphics card
Now we come to the installation of the app.
Here you can install git clone
Here you see what needs to be imported and configured
- gdwon –id install
Here you can add your json file with pd.read_json() Method
Here you add your dataset and x-y min and max
We use it to show our license plates found on the pictures
Here we can get the pictures displayed and the indicator is marked
About Fellow Consulting
Fellow Consulting AG, based in Poing near Munich, has been advising clients on digitizing their business processes for more than 10 years.
We know how our customers can implement their digital vision. Together with our partners infor_ Amazon AWS, Microsoft Azure, SugarCRM, and Ephesoft, we bring the mission of digitalization to life.
We act as an operational and strategic specialist for our clients as well as a sparring partner for the everyday challenges of digital transformation.