Active Object Reconstruction Using a Guided View Planner
This addresses the challenge of efficient 3D reconstruction for robotics or computer vision applications, but it is incremental as it builds on existing deep learning-based reconstruction methods.
The paper tackles the problem of 3D object reconstruction by planning a sequence of informative views, resulting in increased reconstruction accuracy with more views and more informative view sequences compared to alternative methods.
Inspired by the recent advance of image-based object reconstruction using deep learning, we present an active reconstruction model using a guided view planner. We aim to reconstruct a 3D model using images observed from a planned sequence of informative and discriminative views. But where are such informative and discriminative views around an object? To address this we propose a unified model for view planning and object reconstruction, which is utilized to learn a guided information acquisition model and to aggregate information from a sequence of images for reconstruction. Experiments show that our model (1) increases our reconstruction accuracy with an increasing number of views (2) and generally predicts a more informative sequence of views for object reconstruction compared to other alternative methods.