Multi-View Fine-Grained Vehicle Classification with Multi-Loss Learning

摘要: A computer vision solution applied to an automatic toll collection (ATC) with a subscription/membership is proposed in this paper. In application, unique identifier (ID) related concrete vehicle and membership. camera system put place verify that for each transaction the ID correspond actual membership data. The visual extracts different characteristics including license plate number, make, model, color, number of axles, etc. then compares extracted those found We focus on solving make classification that, we propose fine-grained exploits multi-camera composition by feeding convolutional neural network (CNN) multiple views vehicle. Our multi-view (MV) features from vehicle’s classification. presented evaluations show using information improves performance particular more challenging scenarios.