Comparison of Faster-RCNN, YOLO, and SSD for Real-Time Vehicle Type Recognition
Jeong-Ah Kim, Ju-Yeong Sung, Se-Ho Park
Abstract
This paper studies a method to recognize vehicle types based on deep learning model. Faster-RCNN, YOLO, and SSD, which can be processed in real-time and have relatively high accuracy, are presented in this paper. We trained each algorithm through an automobile training dataset and analyzed the performance to determine what is the optimized model for vehicle type recognition. The Yolov4 model outperforms other methods, showing 93% accuracy in recognizing the vehicle model.
Topics & Concepts
Computer scienceArtificial intelligenceVehicle typeDeep learningPattern recognition (psychology)Computer visionEngineeringTransport engineeringAdvanced Neural Network ApplicationsVehicle License Plate RecognitionHandwritten Text Recognition Techniques