Implicit Scene
Implicit scene representation focuses on encoding 3D scenes using neural networks, aiming for efficient and accurate scene understanding and novel view synthesis. Current research emphasizes developing robust and generalizable models, often leveraging neural radiance fields (NeRFs), Gaussian splatting, and signed distance functions (SDFs), with a growing interest in incorporating semantic information and handling dynamic scenes. This field is significant for advancing robotics, augmented/virtual reality, and autonomous driving, by enabling efficient 3D scene reconstruction, object manipulation, and navigation based on compact, learned representations.
Papers
GOAT-Bench: A Benchmark for Multi-Modal Lifelong Navigation
Mukul Khanna, Ram Ramrakhya, Gunjan Chhablani, Sriram Yenamandra, Theophile Gervet, Matthew Chang, Zsolt Kira, Devendra Singh Chaplot, Dhruv Batra, Roozbeh Mottaghi
Incremental Joint Learning of Depth, Pose and Implicit Scene Representation on Monocular Camera in Large-scale Scenes
Tianchen Deng, Nailin Wang, Chongdi Wang, Shenghai Yuan, Jingchuan Wang, Danwei Wang, Weidong Chen