We successfully implemented the LPD/LPR sample app and the results are amazing. Thank you for putting together this example.
For our use case though, we have changed the secondary detection (LPD) to have multiple classes. E.g.: let’s say, licence plate, wheels, roof rack etc. Our custom detectnet_v2 model is working fine and it is showing the multiple classes detected by this LPD model.
We still prefer to run LPR on license plates. For this, we changed the LPR config file to have the following parameter “operate-on-class-ids” and provided it with the class id of licence plates. However, it still runs the LPR process on all classes. Could you please suggest any other changes that I should be doing?