Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass

By Jianing Yang, Alexander Sax, Kevin J. Liang, Mikael Henaff, Hao Tang, Ang Cao, Joyce Chai, F. Meier, Matt Feiszli. Computer Vision and Pattern Recognition, 2025

356 citations recorded by the source corpus.

Multi-view 3D reconstruction remains a core challenge in computer vision, particularly in applications requiring accurate and scalable representations across diverse perspectives. Current leading methods such as DUSt3R employ a fundamentally pairwise approach, processing images in pairs and necessitating costly global alignment procedures to reconstruct from multiple views. In this work, we propose Fast 3D Reconstruction (Fast3R), a novel multi-view generalization to DUSt3R that achieves efficient and scalable 3D reconstruction by processing many views in parallel. Fast3R’s Transformer-based architecture forwards N images in a single forward pass, bypassing the need for iterative alignment. Through extensive experiments on camera pose estimation and 3D reconstruction, Fast3R demonstrates state-of-the-art performance, with significant improvements in inference speed and reduced error accumulation. These results establish Fast3R as a robust alternative for multi-view applications, offering enhanced scalability without compromising reconstruction accuracy.

Subfields: Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/62d564bc589cf6c37139144d0d6984d4aff1cc4a

Citations in the corpus

• VGGT: Visual Geometry Grounded Transformer (2025, neutral): “Some concurrent works [73, 85, 96, 105] explore replacing DUSt3R’s test-time optimization with neural networks, though these attempts achieve only suboptimal or comparable performance to DUSt3R.”
• Depth Anything 3: Recovering the Visual Space from Any Views (2025, neutral): “More recently, multi-view geometry foundation models [85, 91, 96, 110] have been integrated to improve modeling capacity, particularly in pose-free settings, yet methods built upon such models are often evaluated by relying on a single chosen foundation model [41, 79, 116].”