Building Rome in a day

By Sameer Agarwal, Yasutaka Furukawa, Noah Snavely, Ian Simon, Brian Curless, S. M. Seitz, Richard Szeliski. 2009

2,551 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

3D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstruction (2016, neutral): “Apr 2016 2 C. B. Choy, D. Xu, J. Gwak, K. Chen, and S. Savarese These restrictions stem from a number of key technical assumptions. One typical assumption is that features can be matched across views [3,4,5,6] as hypothesized by the majority of the methods based on SFM or SLAM [7,8]. It has been dem”
Generic 3D Representation via Pose Estimation and Matching (2016, methodology): “This is diﬀerent from many previous works [3,26,59,52,8,14,20] from both visual odometery and SfM literature that perform the estimation through a two step process consisting of ﬁnding point correspondences between images followed by pose estimation.”
Shading-based refinement on volumetric signed distance functions (2015, methodology): “Sparse Bundle Adjustment Step First, we perform a sparse bundle adjustment step, similar to traditional bundle adjustment of RGB images [Triggs et al. 2000; Snavely et al. 2006; Agarwal et al. 2011]; however, we formulate our pose optimization to take advantage of the depth channel as well.”
Cross-View Image Geolocalization (2013, background): “[1] developed an approach to build 3D point cloud models of famous buildings automatically from publicly available images.”
StereoScan: Dense 3d reconstruction in real-time (2011, background): “More recently, the availability of photo sharing platforms like Flickr led to efforts of modeling cities as large as Rome [1], [8].”
Fine-Tuning CNN Image Retrieval with No Human Annotation (2017, methodology): “ing SfM reconstruction, which is performed using the stateof-the-art SfM pipeline [60], [61], [62].”
Emergent Correspondence from Image Diffusion (2023, background): “Establishing visual correspondences between different images is crucial for various computer vision tasks such as Structure-from-Motion / 3D reconstruction [2, 74, 61, 75], object tracking [23, 93], image recognition [64, 81, 9] and segmentation [51, 48, 73, 30].”
DeepV2D: Video to Depth with Differentiable Structure from Motion (2018, methodology): “This pipeline has shown impressive results on a number of tasks [1, 15] and is often paired with Multi-View Stereo to build a more complete 3D representation [8, 9].”
Instant3D: Fast Text-to-3D with Sparse-View Generation and Large Reconstruction Model (2023, background): “Traditional 3D reconstruction with multi-view stereo (Agarwal et al., 2011; Sch ¨ onberger et al., 2016; Furukawa et al., 2015) typically require a dense set of input images that have significant overlaps to find correspondence across views and infer the geometry correctly.”
Robust Global Translations with 1DSfM (2014, methodology): “While some earlier SfM methods were global, such as factorization [20], most current large-scale SfM systems involve sequential reconstruction [19,2,9].”
AutoScaler: Scale-Attention Networks for Visual Correspondence (2016, background): “It is crucial for a plethora of applications, including optical flow [32, 37, 46], structure-from-motion [1], visual SLAM [26, 35, 36], stereo matching [31, 53], non-rigid 3D reconstruction [14] as well as video segmentation [18].”
Let There Be Color! Large-Scale Texturing of 3D Reconstructions (2014, methodology): “Our method takes as input a set of (typically several hundred) images of a scene that were registered using structure-from-motion [1,20].”
Large-Scale Multi-resolution Surface Reconstruction from RGB-D Sequences (2013, background): “While impressive progress in this domain has been achieved over the last decade [2, 1, 9, 5], many of these approaches are based on visual keypoints that are reconstructed in 3D, which typically leads to sparse reconstructions in form of 3D point clouds.”
Tanks and temples (2017, methodology): “In the past decade, structure from motion (SfM) and multi-view stereo (MVS) techniques have advanced to enable remarkable reconstructions of landmark scenes from community photo collections [Agarwal et al. 2011; Shan et al. 2013; Snavely et al. 2008].”
Non-Rigid Neural Radiance Fields: Reconstruction and Novel View Synthesis of a Dynamic Scene From Monocular Video (2020, methodology): “These methods are based on light fields [34, 19, 4], multi-view stereo to capture dense depth maps [94], layered depth images [66], or representations using 3D point clouds [1, 39, 65], meshes [43, 82] or surfels [60, 7, 84] for dynamic scenes.”
Towards Internet-scale multi-view stereo (2010, background): “’s Rome-in-a-day project [1]), we focus our efforts in this paper on the last stage of the pipeline, i.”
KAZE Features (2012, neutral): “These multiscale feature algorithms are a key component in modern computer vision frameworks, such as scene understanding [1], visual categorization [2] and large scale 3D Structure from Motion (SfM) [3].”
Learning and Calibrating Per-Location Classifiers for Visual Place Recognition (2013, methodology): “Experiments are performed on a database of 25,000 geotagged street view images of Pittsburgh and demonstrate improved place recognition accuracy of the proposed approach over the previous work.”
VGGT: Visual Geometry Grounded Transformer (2025, neutral): “The traditional SfM pipeline [2, 24, 48, 62, 68, 92] consists of multiple stages, including image matching, triangulation, and bundle adjustment.”