In this paper, we present a novel architecture and training methodology for learning monocular depth prediction, camera pose estimation, optical flow, and moving object segmentation using a common encoder in an unsupervised fashion. We demonstrate that the geometrical relationships between these tasks not only support joint unsupervised learning as shown in previous works but also allow them to share common features. We also show the advantage of using a two-stage learning approach to improve the performance of the base network.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Unsupervised Joint Multi-Task Learning of Vision Geometry Tasks


    Beteiligte:
    Jha, Prabhash Kumar (Autor:in) / Tsanev, Doychin (Autor:in) / Lukic, Luka (Autor:in)


    Erscheinungsdatum :

    11.07.2021


    Format / Umfang :

    1568700 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Multi-task Learning of Semantics,Geometry and Motion for Vision-based End-to-End Self-Driving

    Ni, Hailong / Wu, Jian / Zhang, Dong et al. | British Library Conference Proceedings | 2021


    Multi-task Learning of Semantics, Geometry and Motion for Vision-based End-to-End Self-Driving

    Zhang, Dong / Wang, Guojun / Ni, Hailong et al. | SAE Technical Papers | 2021



    Unsupervised Multi-Spectrum Stereo Depth Estimation for All-Day Vision

    Guo, Yubin / Kong, Hui / Gu, Shuo | IEEE | 2024


    Unsupervised Learning of Optical Flow With Non-Occlusion From Geometry

    Wang, Guangming / Ren, Shuaiqi / Wang, Hesheng | IEEE | 2022