SimRecon: SimReady Compositional Scene Reconstruction from Real Videos

TL;DR

SimRecon introduces a novel pipeline for compositional scene reconstruction from real videos, integrating semantic reconstruction, object generation, and assembly with modules enhancing visual fidelity and physical plausibility.

Contribution

It proposes a new perception-generation-simulation framework with bridging modules for improved realism and physical accuracy in scene reconstruction from videos.

Findings

Outperforms previous methods on ScanNet dataset

Achieves higher visual fidelity in reconstructed assets

Ensures physically plausible scene assembly

Abstract

Compositional scene reconstruction seeks to create object-centric representations rather than holistic scenes from real-world videos, which is natively applicable for simulation and interaction. Conventional compositional reconstruction approaches primarily emphasize on visual appearance and show limited generalization ability to real-world scenarios. In this paper, we propose SimRecon, a framework that realizes a "Perception-Generation-Simulation" pipeline towards cluttered scene reconstruction, which first conducts scene-level semantic reconstruction from video input, then performs single-object generation, and finally assembles these assets in the simulator. However, naively combining these three stages leads to visual infidelity of generated assets and physical implausibility of the final scene, a problem particularly severe for complex scenes. Thus, we further propose two bridging modules between the three stages to address this problem. To be specific, for the transition from Perception to Generation, critical for visual fidelity, we introduce Active Viewpoint Optimization, which actively searches in 3D space to acquire optimal projected images as conditions for single-object completion. Moreover, for the transition from Generation to Simulation, essential for physical plausibility, we propose a Scene Graph Synthesizer, which guides the construction from scratch in 3D simulators, mirroring the native, constructive principle of the real world. Extensive experiments on the ScanNet dataset validate our method's superior performance over previous state-of-the-art approaches.