EscherNet: A Generative Model for Scalable View Synthesis

doi:doi:10.57702/qhbmtyy6

EscherNet: A Generative Model for Scalable View Synthesis

EscherNet is a multi-view conditioned diffusion model designed for scalable view synthesis. It leverages Stable Diffusion's 2D architecture empowered by the innovative Camera Positional Embedding (CaPE), EscherNet adeptly learns implicit 3D representations from varying number of reference views, achieving consistent 3D novel view synthesis.

BibTex:

@dataset{Xin_Kong_and_Shikun_Liu_and_Xiaoyang_Lyu_and_Marwan_Taher_and_Xiaojuan_Qi_and_Andrew_J_Davison_2024,
    abstract = {EscherNet is a multi-view conditioned diffusion model designed for scalable view synthesis. It leverages Stable Diffusion's 2D architecture empowered by the innovative Camera Positional Embedding (CaPE), EscherNet adeptly learns implicit 3D representations from varying number of reference views, achieving consistent 3D novel view synthesis.},
    author = {Xin Kong and Shikun Liu and Xiaoyang Lyu and Marwan Taher and Xiaojuan Qi and Andrew J. Davison},
    doi = {10.57702/qhbmtyy6},
    institution = {No Organization},
    keyword = {'3D reconstruction', 'camera positional encoding', 'diffusion model', 'view synthesis'},
    month = {dec},
    publisher = {TIB},
    title = {EscherNet: A Generative Model for Scalable View Synthesis},
    url = {https://service.tib.eu/ldmservice/dataset/eschernet--a-generative-model-for-scalable-view-synthesis},
    year = {2024}
}