3D Scene Reconstruction

37 papers with code • 0 benchmarks • 3 datasets

Creating 3D scene either using conventional SFM pipelines or latest deep learning approaches.

Most implemented papers

NeuralRecon: Real-Time Coherent 3D Reconstruction from Monocular Video

zju3dv/NeuralRecon CVPR 2021

We present a novel framework named NeuralRecon for real-time 3D scene reconstruction from a monocular video.

The Replica Dataset: A Digital Replica of Indoor Spaces

facebookresearch/Replica-Dataset 13 Jun 2019

We introduce Replica, a dataset of 18 highly photo-realistic 3D indoor scene reconstructions at room and building scale.

CoReNet: Coherent 3D scene reconstruction from a single RGB image

google-research/corenet ECCV 2020

Furthermore, we adapt our model to address the harder task of reconstructing multiple objects from a single image.

UniScene: Multi-Camera Unified Pre-training via 3D Scene Reconstruction

chaytonmin/uniscene 30 May 2023

When compared to monocular pre-training methods on the nuScenes dataset, UniScene shows a significant improvement of about 2. 0% in mAP and 2. 0% in NDS for multi-camera 3D object detection, as well as a 3% increase in mIoU for surrounding semantic scene completion.

HOC-Search: Efficient CAD Model and Pose Retrieval from RGB-D Scans

stefan-ainetter/SCANnotateDataset 12 Sep 2023

We present an automated and efficient approach for retrieving high-quality CAD models of objects and their poses in a scene captured by a moving RGB-D camera.

DeepV2D: Video to Depth with Differentiable Structure from Motion

princeton-vl/DeepV2D ICLR 2020

We propose DeepV2D, an end-to-end deep learning architecture for predicting depth from video.

Neural RGB->D Sensing: Depth and Uncertainty from a Video Camera

NVlabs/neuralrgbd 9 Jan 2019

Depth sensing is crucial for 3D reconstruction and scene understanding.

Atlas: End-to-End 3D Scene Reconstruction from Posed Images

magicleap/Atlas ECCV 2020

Traditional approaches to 3D reconstruction rely on an intermediate representation of depth maps prior to estimating a full 3D model of a scene.

GRF: Learning a General Radiance Field for 3D Representation and Rendering

alextrevithick/GRF ICCV 2021

We present a simple yet powerful neural network that implicitly represents and renders 3D objects and scenes only from 2D observations.

Learning to Recover 3D Scene Shape from a Single Image

aim-uofa/AdelaiDepth CVPR 2021

Despite significant progress in monocular depth estimation in the wild, recent state-of-the-art methods cannot be used to recover accurate 3D scene shape due to an unknown depth shift induced by shift-invariant reconstruction losses used in mixed-data depth prediction training, and possible unknown camera focal length.