SCFusion: Real-time Incremental Scene Reconstruction with Semantic Completion

TL;DR

This paper introduces SCFusion, a real-time framework that jointly performs scene reconstruction and semantic completion from depth data, addressing occlusion issues for applications like AR and robotics.

Contribution

It presents a novel neural architecture that fuses semantic completion with 3D reconstruction incrementally and efficiently in real-time.

Findings

Achieves accurate 3D semantic scene completion in real-time

Demonstrates superior performance over offline methods

Effectively handles occlusions in dynamic scenes

Abstract

Real-time scene reconstruction from depth data inevitably suffers from occlusion, thus leading to incomplete 3D models. Partial reconstructions, in turn, limit the performance of algorithms that leverage them for applications in the context of, e.g., augmented reality, robotic navigation, and 3D mapping. Most methods address this issue by predicting the missing geometry as an offline optimization, thus being incompatible with real-time applications. We propose a framework that ameliorates this issue by performing scene reconstruction and semantic scene completion jointly in an incremental and real-time manner, based on an input sequence of depth maps. Our framework relies on a novel neural architecture designed to process occupancy maps and leverages voxel states to accurately and efficiently fuse semantic completion with the 3D global model. We evaluate the proposed approach quantitatively and qualitatively, demonstrating that our method can obtain accurate 3D semantic scene completion in real-time.