Dynamic Grouping With Multi-Manifold Attention for Multi-View 3D Object Reconstruction

In a multi-view 3D reconstruction problem, the task is to infer the 3D shape of an object from various images taken from different viewpoints. Transformer-based networks have demonstrated their ability to achieve high performance in such problems, but they face challenges in identifying the optimal...

詳細記述

書誌詳細

出版年:	IEEE Access
主要な著者:	Georgios Kalitsios, Dimitrios Konstantinidis, Petros Daras, Kosmas Dimitropoulos
フォーマット:	論文
言語:	英語
出版事項:	IEEE 2024-01-01
主題:	Dynamic grouping multi-manifold attention multi-view 3D reconstruction transformer voxel representation
オンライン･アクセス:	https://ieeexplore.ieee.org/document/10721458/