Inspired by the research of human visual system in neuroanatomy and psychology, the paper proposes an road traffic sign identification model based on vision bionics.The model combines data-driven and task-driven visual attention mechanism to focus on traffic sign target rapidly and accuractly.Firstly,It simulates the Itti attention model to obtain “what” information and uses the priori knowledge of positional distribution of traffic sign as “where” information.Then,it adjusts saliency map according to “what” and “where” stream so as to select traffic sign focus preferentially.Secondly,it measures the similarity of shape and color features between traffic sign and attention region to get interested region.Finally, it segments traffic sign based on color characteristics and classify shape of traffic sign based on the Support Vector Machine method. The experimental results demonstrate that the feasibility and effectiveness of the proposed model; Furthermore, the average accuracy rate of shape classification on DtBs matrix reaches 98%.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Traffic Sign Recognition Based on Vision Bionics



    Erschienen in:

    Erscheinungsdatum :

    26.07.2012


    Format / Umfang :

    6 pages




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    Road traffic sign recognition system based on machine vision

    XU XU | Europäisches Patentamt | 2021

    Freier Zugriff

    Vehicle vision system with enhanced traffic sign recognition

    BIEMER MICHAEL / BOEGEL RUEDIGER | Europäisches Patentamt | 2019

    Freier Zugriff

    VEHICLE VISION SYSTEM WITH ENHANCED TRAFFIC SIGN RECOGNITION

    BIEMER MICHAEL / BOEGEL RUEDIGER | Europäisches Patentamt | 2018

    Freier Zugriff

    A new vision system for traffic sign recognition

    Yanlei Gu, / Yendo, Tomohiro / Tehrani, Mehrdad Panahpour et al. | IEEE | 2010