日本フィジカルAI新聞

世界のフィジカルAIを、日本語で。

週刊ニュースレター購読
arXiv:2301.13244

Mono-STAR: Mono-camera Scene-level Tracking and Reconstruction

Mono-STAR: Mono-camera Scene-level Tracking and Reconstruction

シェア:XThreadsFacebookLINEはてブBluesky

著者: Haonan Chang, Dhruv Metha Ramesh, Shijie Geng, Yuqiu Gan, Abdeslam Boularias

分類: cs.RO, cs.CV

原文アブストラクト

We present Mono-STAR, the first real-time 3D reconstruction system that simultaneously supports semantic fusion, fast motion tracking, non-rigid object deformation, and topological change under a unified framework. The proposed system solves a new optimization problem incorporating optical-flow-based 2D constraints to deal with fast motion and a novel semantic-aware deformation graph (SAD-graph) for handling topology change. We test the proposed system under various challenging scenes and demonstrate that it significantly outperforms existing state-of-the-art methods.