Advances in Feed-Forward 3D Reconstruction and View Synthesis: A Survey

TL;DR

This survey reviews recent feed-forward deep learning methods for 3D reconstruction and view synthesis, highlighting their architectures, applications, datasets, and future challenges in computer vision and immersive tech.

Contribution

It provides a comprehensive taxonomy and analysis of feed-forward 3D reconstruction and view synthesis techniques, covering various representations and applications.

Findings

Feed-forward methods enable fast, real-time 3D reconstruction.

Deep learning approaches outperform traditional iterative methods.

Open challenges include generalization and dataset limitations.

Abstract

3D reconstruction and view synthesis are foundational problems in computer vision, graphics, and immersive technologies such as augmented reality (AR), virtual reality (VR), and digital twins. Traditional methods rely on computationally intensive iterative optimization in a complex chain, limiting their applicability in real-world scenarios. Recent advances in feed-forward approaches, driven by deep learning, have revolutionized this field by enabling fast and generalizable 3D reconstruction and view synthesis. This survey offers a comprehensive review of feed-forward techniques for 3D reconstruction and view synthesis, with a taxonomy according to the underlying representation architectures including point cloud, 3D Gaussian Splatting (3DGS), Neural Radiance Fields (NeRF), etc. We examine key tasks such as pose-free reconstruction, dynamic 3D reconstruction, and 3D-aware image and video synthesis, highlighting their applications in digital humans, SLAM, robotics, and beyond. In addition, we review commonly used datasets with detailed statistics, along with evaluation protocols for various downstream tasks. We conclude by discussing open research challenges and promising directions for future work, emphasizing the potential of feed-forward approaches to advance the state of the art in 3D vision.