Akademik Veri Yönetim Sistemi
30th Signal Processing and Communications Applications Conference, SIU 2022, Safranbolu, Türkiye, 15 - 18 Mayıs 2022
© 2022 IEEE.It is crucial that end users must choose the appropriate tire according to vehicle's requirements and standards for safety. With the increase in demand for e-commerce channels due to the change in consumer habits, users who want to buy tires online have great difficulty in distinguishing the necessary brand and size information from the surface of the tire and their customer experiences are interrupted without purchasing. To sort a solution out for this pain point, a model based on image processing and supervised deep learning algorithms has been developed to find required information for purchasing a tire such as brand and size from a single image. Proposed model detects the vehicle tire and makes the input image suitable for reading the information on it. Then the brand and size are determined with the segmentation deep learning models which are trained on the problem specific dataset prepared by the authors. The proposed model is one of the working examples that offers an end-to-end solution for online tire purchasing process of customers, with a success rate of 97.3% in brand detection and 95.1% in size detection.