In traffic sign recognition tasks, recognition of road signs by observing synthetic reference images is a human-like ability that can be performed by one-shot learning algorithms. One-shot object recognition is a challenging task for deep neural networks in which a deep model classifies query examples based on support images. It becomes more difficult when there is a domain shift between support and query samples. The generalization of a deep model on an unknown domain with different distributions is another problematic task in on-shot recognition. This work introduces a novel deep network named SeqNet to overcome the aforementioned problems. To the best of our knowledge, this work outperforms all state-of-the-art models in one-shot traffic sign recognition and one-shot logo identification by superior results. Our proposed SeqNet model generalizes to unseen domains without further model fine-tuning on the test data. Also, we show how using transferred knowledge from an irrelevant but large domain could reduce the network parameters that result in model size reduction. By utilizing the power of transferred knowledge from a large deep model the SeqNet becomes smaller and has about 6X fewer parameters than its competitors. The smaller size of the SeqNet architecture enables it to be used in resource-constrained devices in many applications such as smart vehicles. The experimental results depict that our proposed SeqNet performance is ameliorated by large margins, with up to 20% accuracy for one-shot classification and 30% area under the curve (AUC) for image retrieval tasks.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    SeqNet: Sequential Networks for One-Shot Traffic Sign Recognition With Transfer Learning


    Beteiligte:

    Erschienen in:

    Erscheinungsdatum :

    01.12.2024


    Format / Umfang :

    2051138 byte




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Traffic Sign Recognition Framework Using Zero-Shot Learning

    Shah, Prachi / Patel, Parmanand / Kothadiya, Deep | Springer Verlag | 2024


    Ethiopian Traffic Sign Recognition Using Customized Convolutional Neural Networks and Transfer Learning

    Amlakie Aschale Alemu / Misganaw Aguate Widneh | DOAJ | 2025

    Freier Zugriff

    TRAFFIC SIGN RECOGNITION DEVICE AND TRAFFIC SIGN RECOGNITION METHOD

    MIYASATO KAZUHIRO / KOYASU TOSHIYA | Europäisches Patentamt | 2023

    Freier Zugriff

    TRAFFIC SIGN RECOGNITION DEVICE AND TRAFFIC SIGN RECOGNITION METHOD

    SHINOMIYA TERUHIKO | Europäisches Patentamt | 2017

    Freier Zugriff

    Traffic sign recognition

    Howard, A.R. | Engineering Index Backfile | 1964