Traffic Sign Detection And Recognition Using TensorFlow’ s Object Detection API With A New Benchmark Dataset
İrfan Kılıç, Galip Aydın
Abstract
In this study TensorFlow's Object Detection API used in Turkey traffic sign detection and recognition of the aid intended traffic signs. A dataset suitable for TensorFlow's Object Detection API has been prepared for traffic signs by using photographs in different traffic and weather conditions. A web-based data tagging application has been developed to prepare this dataset. This dataset was trained by selecting the appropriate deep learning model and results were obtained on the test data. Although the training data are insufficient, the results on the test images are satisfactory.
Topics & Concepts
Benchmark (surveying)Computer scienceTraffic signObject detectionArtificial intelligenceDeep learningSign (mathematics)Test dataObject (grammar)Cognitive neuroscience of visual object recognitionPattern recognition (psychology)Traffic sign recognitionComputer visionMachine learningMathematicsProgramming languageGeodesyGeographyMathematical analysisAdvanced Neural Network ApplicationsVehicle License Plate RecognitionHand Gesture Recognition Systems