Autonomous vehicle driving systems have become one of the most important topics recently, some people assert that they can identify traffic signs automatically, which is a revolutionary improvement for transportation; however, the accuracy of autonomous vehicle driving systems still remains controversial. Therefore, this research analyzes the German Traffic Sign Benchmark dataset, in three different ways: AlexNet, VGG-16, and ResNet-50 of autonomous vehicle driving systems, to compare the best approach for identifying traffic signs. One can conclude that AlexNet, VGG-16 and ResNet-50 performed well, as they got an accuracy score of 95%, 95%, 89% respectively. To improve the classification accuracy to achieve nearly 100% for total security, Bottleneck Attention Module (BAM) is examined as a way of improving the classification accuracy of models and it is confirmed that BAM is able to boost the accuracy score of select models.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Attention-aware CNN model for Traffic Signs Classification


    Beteiligte:
    Gu, Xiangsheng (Autor:in) / Lang, Qiuyi (Autor:in) / Lin, Fu (Autor:in) / Wang, Pengfei (Autor:in)


    Erscheinungsdatum :

    01.09.2022


    Format / Umfang :

    1282196 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Classification of Traffic Signs

    Habibi Aghdam, Hamed / Jahani Heravi, Elnaz | Springer Verlag | 2017


    Analysis of Driver Gaze and Attention to Traffic Signs

    Shabnam Shabani / Steven Beauchemin / Michael Bauer | DOAJ | 2022

    Freier Zugriff

    Traffic signs

    Engineering Index Backfile | 1946


    Traffic signs

    Engineering Index Backfile | 1933


    Traffic signs

    Engineering Index Backfile | 1946