A space-sweep approach to true multi-image matching

By Robert T. Collins. 1996

How researchers see this paper

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

Citations in the corpus

Learning a Multi-View Stereo Machine (2017, methodology): “Another approach to MVS involves representing the reconstruction as a collection of depth maps [6, 57, 41, 13, 40].”
Deep Stereo: Learning to Predict New Views from the World's Imagery (2015, methodology): “We select one of the images in the center rosette as the target and train to produce it from the others.”
Just Look at the Image: Viewpoint-Specific Surface Normal Prediction for Improved Multi-View Reconstruction (2016, background): “Many multi-view stereo methods only estimate depth, e.g. [7] [30] [36].”
Semi-global Stereo Matching with Surface Orientation Priors (2017, unknown): “Plane-sweep stereo with a preferred plane orientation was proposed by Collins [11], and subsequently extended to multiple plane orientations, Manhattan-world scenes, and piece-wise planar scenes [18, 17, 43, 30].”
DeMoN: Depth and Motion Network for Learning Monocular Stereo (2016, methodology): “Only after estimation of the camera motion and a sparse 3D point cloud, dense depth maps are computed by exploiting the epipolar geometry [4].”
Learned Multi-patch Similarity (2017, methodology): “One exhaustively tests for all possible depths, often by “sweeping” a fronto-parallel plane through the 3D object space along the depth axis [4] or in multiple directions [10] to sample different depth values efficiently.”
Event-Based Vision: A Survey (2019, background): “It back-proj-ects events into space, creating a ray density volume [194], and then ﬁnds scene structure as local maxima of ray density.”
Manhattan-world stereo (2009, background): “They then sweep families of planes through the scene [6, 16] and measure the photoconsistency or correlation at each pixel in order to estimate depth maps.”
Handling occlusions in dense multi-view stereo (2001, methodology): “In a more general (plane sweep) multi-view setting [9, 25], each disparity corresponds to some plane equation in3-D.”
Multi-view reconstruction preserving weakly-supported surfaces (2011, methodology): “There are several approaches to computing the point cloud, such as plane-sweeping based [6], stereo based [19], or growing based [8] methods.”
Consistent Depth Maps Recovery from a Video Sequence (2009, background): “STEREO reconstruction of dense depth maps from naturalvideo sequences is a fundamentally important and challenging problem in computer vision.”
Free View Synthesis (2020, methodology): “[13] used a plane-sweep volume [9] within a network architecture for image-based rendering.”
Track to Detect and Segment: An Online Multi-Object Tracker (2021, methodology): “The cost volume technique has been successfully applied in depth estimation [11, 55, 18] and optical flow estimation [35, 10, 53] for associating pixels between two frames.”
MVSNet: Depth Inference for Unstructured Multi-view Stereo (2018, methodology): “Initial Estimation The simplest way to retrieve depth map D from the probability volume P is the pixel-wise winner-take-all [30] (i.”
A Theory of Shape by Space Carving (1999, unknown): “, they r el on the presence of specific image features such as edges and hence generate only sparse reconstructions [28], or they place strong constraints on the input viewpoints relative to the scene [29, 30].”
Massively Parallel Multiview Stereopsis by Surface Normal Diffusion (2015, methodology): “The plane-sweeping method [8] is an early example of true multiview matching.”