Litcius/Paper detail

Classification and Prediction of License Plates Using Deeply Learned Convolutional Neural Networks

Manik Rakhra, Dalwinder Singh, Arun Singh, Ahmed Al Ahdal, Deepa Gupta

20222022 10th International Conference on Reliability, Infocom Technologies and Optimization (Trends and Future Directions) (ICRITO)26 citationsDOI

Abstract

Classifying and detecting of vehicle license plates using image processing techniques is an exciting research topic in IoT technology and the Internet of Things. Recognizing a vehicle's plate number is necessary since the number of vehicles on the streets is increasing and human capacity to perform this activity is limited. so it is simple for humans to read and recognize license plates that belonging to which country or state, but would the machines do the same when detecting or identifying plate numbers that correspond to different country or state?in this paper will discuss and implement a model that can read any License plates belong to any American states using Deep Learning Model, and (CNN) convolutional neural network in this paper we used DenseNet201 applied to Us license plate dataset which available on Kaggle scored 90.38%, then we used InceptionResNetV2 from keras for the model and it score 89% accuracy, EfflcientNetB5 and score 91.5% accuracy, and Xception it scored 81%accuracy.

Topics & Concepts

LicenseConvolutional neural networkComputer scienceArtificial intelligenceDeep learningState (computer science)Artificial neural networkMachine learningPattern recognition (psychology)The InternetComputer visionWorld Wide WebAlgorithmOperating systemVehicle License Plate RecognitionAdvanced Neural Network ApplicationsSmart Parking Systems Research