Fine-grained vehicle recognition (FGVR) is an essential fundamental technology for intelligent transportation systems. However, it is challenging because of its inherent intra-class variation. Most previous FGVR studies only focus on the intra-class variation caused by spatial factors like shooting angles and positions, while the intra-class variation caused by image noise has received little attention. This paper proposes a progressive multi-task anti-noise learning (PMAL) framework and a progressive multi-task distilling (PMD) framework to address the intra-class variation problem in FGVR caused by image noise. The PMAL framework achieves high recognition accuracy by treating image denoising as an additional task in image recognition. It progressively trains the model to learn noise invariance. The PMD framework transfers the knowledge learned by PMAL-trained model into the original backbone network. This results in a model with similar recognition accuracy to the PMAL-trained model but without any additional overhead on the original backbone network. Combining the two frameworks, we obtain models that significantly exceed previous state-of-the-art methods in recognition accuracy on two widely-used, standard FGVR datasets, namely Stanford Cars and CompCars. In addition, we observed improvements on three other surveillance image-based vehicle-type classification datasets, namely Beijing Institute of Technology (BIT)-Vehicle, Vehicle Type Image Data 2 (VTID2), and Vehicle Images Dataset for Make & Model Recognition (VIDMMR), without any additional overhead on the original backbone networks. The source code is available at https://github.com/Dichao-Liu/Anti-noise_FGVR.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Progressive Multi-Task Anti-Noise Learning and Distilling Frameworks for Fine-Grained Vehicle Recognition


    Contributors:
    Liu, Dichao (author)


    Publication date :

    2024-09-01


    Size :

    1962303 byte




    Type of media :

    Article (Journal)


    Type of material :

    Electronic Resource


    Language :

    English