Human-in-the-Loop Local Corrections of 3D Scene Layouts via Infilling
This addresses the challenge of accurately modeling complex 3D scene layouts for applications like robotics or AR/VR, though it is incremental as it builds on an existing framework.
The paper tackles the problem of improving 3D scene layout estimation by introducing a human-in-the-loop approach where users identify local errors and prompt a model to correct them via infilling, resulting in a system that maintains global prediction performance while significantly enhancing local correction ability.
We present a novel human-in-the-loop approach to estimate 3D scene layout that uses human feedback from an egocentric standpoint. We study this approach through introduction of a novel local correction task, where users identify local errors and prompt a model to automatically correct them. Building on SceneScript, a state-of-the-art framework for 3D scene layout estimation that leverages structured language, we propose a solution that structures this problem as "infilling", a task studied in natural language processing. We train a multi-task version of SceneScript that maintains performance on global predictions while significantly improving its local correction ability. We integrate this into a human-in-the-loop system, enabling a user to iteratively refine scene layout estimates via a low-friction "one-click fix'' workflow. Our system enables the final refined layout to diverge from the training distribution, allowing for more accurate modelling of complex layouts.