RIC: Rotate-Inpaint-Complete for Generalizable Scene Reconstruction

TL;DR

This paper introduces RIC, a scene reconstruction method that combines inpainting with 2D to 3D lifting, leveraging large language models for generalization to unseen objects and scenes.

Contribution

The method innovatively uses large visual language models for inpainting and normal prediction for robust 3D scene reconstruction from a single view.

Findings

Outperforms multiple baselines in quantitative evaluation.

Demonstrates strong generalization to novel objects and scenes.

Robust to variations in depth and scale.

Abstract

General scene reconstruction refers to the task of estimating the full 3D geometry and texture of a scene containing previously unseen objects. In many practical applications such as AR/VR, autonomous navigation, and robotics, only a single view of the scene may be available, making the scene reconstruction task challenging. In this paper, we present a method for scene reconstruction by structurally breaking the problem into two steps: rendering novel views via inpainting and 2D to 3D scene lifting. Specifically, we leverage the generalization capability of large visual language models (Dalle-2) to inpaint the missing areas of scene color images rendered from different views. Next, we lift these inpainted images to 3D by predicting normals of the inpainted image and solving for the missing depth values. By predicting for normals instead of depth directly, our method allows for robustness to changes in depth distributions and scale. With rigorous quantitative evaluation, we show that our method outperforms multiple baselines while providing generalization to novel objects and scenes.