Location Recognition Using Prioritized Feature Matching

By Yunpeng Li, Noah Snavely, Dan Huttenlocher. 2010

515 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

CPlaNet: Enhancing Image Geolocalization by Combinatorial Partitioning of Maps (2018, methodology): “Pose estimation approaches [9,18,19,20,21,22,23] match query images against 3D models of an area, and employ 2D-3D feature correspondences to identify 6-DOF query poses.”
PlaNet - Photo Geolocation with Convolutional Neural Networks (2016, methodology): “A query image is localized by establishing correspondences between its interest points and the points in the 3D model and solving the resulting perspective-n-point (PnP) problem to obtain the camera parameters [30, 31, 40].”
Image-Based Localization Using LSTMs for Structured Feature Correlation (2016, methodology): “Similarly, descriptor matching can be accelerated by using only a subset of all 3D points [8, 31], which at the same time reduces the memory footprint of the 3D models.”
Learning and Calibrating Per-Location Classifiers for Visual Place Recognition (2013, background): “Examples include: (i) building an explicit 3D reconstruction of the scene [10, 15, 16]; (ii) constructing an image graph [24, 30], where images are nodes and edges connect close-by images on the map [29], or (iii) using the geotagged data as a form of supervision to select local features that…”
Efficient Global 2D-3D Matching for Camera Localization in a Large-Scale 3D Map (2017, methodology): “[13, 31, 32, 45]), though in a local heuristic manner (for example, to improve local search efficiency via query expansion, to prioritize candidate matches, or to filter out false 3D points via geometric validation [40, 43, 44]).”
Visual landmark recognition from Internet photo collections: A large-scale evaluation (2014, background): “Other approaches are based on 3D point clouds created by applying structure-frommotion on Internet photo collections or manually collected photos [36, 37, 21, 23].”
Efficient & Effective Prioritized Matching for Large-Scale Image-Based Localization (2017, methodology): “2D-to-3D matching approaches (tree-based and VPS) achieve a higher localization effectiveness than methods based on 3D-to-2D search such as Point-to-Feature (P2F) matching [2] on three standard datasets.”
Predicting Good Features for Image Geo-Localization Using Per-Bundle VLAD (2015, unknown): “Image retrieval techniques based on local image features [27] can achieve increased robustness against photometric and geometric changes [24, 42].”
Learning Less is More - 6D Camera Localization via 3D Surface Regression (2017, background): “Matching local image descriptors to 3D points of a Structure-from-Motion scene reconstruction yields a set of 2D-3D correspondences, from which an accurate pose estimate can be recovered [13].”
Leveraging Structure from Motion to Learn Discriminative Codebooks for Scalable Landmark Classification (2013, methodology): “Recently proposed alternatives include direct image matching to sparse SfM point clouds [27, 17, 18].”
Fine-Tuning CNN Image Retrieval with No Human Annotation (2017, methodology): “A 3D model is described as a bipartite visibility graph G 1⁄4 ðI [ P; EÞ [65], where images I and points P are the vertices of the graph.”
Image Geo-Localization Based on MultipleNearest Neighbor Feature Matching UsingGeneralized Graphs (2014, methodology): “We evaluated the proposed framework on a new data set of 102k street view images; the experiments show it outperforms the state of the art by 10 percent.”
Learned Contextual Feature Reweighting for Image Geo-Localization (2017, background): “Building image representations from local invariant features have been shown to be effective as they provide robustness to photometric and geometric changes [58, 4, 67, 31, 66].”
Recent developments in large-scale tie-point matching (2016, background): “The second dataset, ROME provided by [22], is not a raw search result from Flickr, but rather a collection of orientable images of famous Rome landmarks like the Colosseum, St.”
From single image query to detailed 3D reconstruction (2015, methodology): “To compare our system against full pairwise reconstruction, we injected the Dubrovnik6k dataset [14] with 6,036 images into the 7.”
VocMatch: Efficient Multiview Correspondence for Structure from Motion (2014, methodology): “The dataset mainly contains related images because they were connected to a 3D model in [10].”
Scene Coordinate Regression Forests for Camera Relocalization in RGB-D Images (2013, methodology): “Given sparse keypoint matches, pose estimation is achieved either via intermediate image retrieval [2, 8, 14] or by directly establishing 2D-3D correspondences for a robust pose optimization [19, 28].”