Single View Metrology

By Antonio Criminisi, Ian David Reid, Andrew Zisserman. 2000

850 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Environmental Engineering

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Environmental Engineering

Citations in the corpus

Synthetic Data for Text Localisation in Natural Images (2016, methodology): “[37] also start from a ﬁxed set of proposal, but reﬁned them prior to detection by using a Region Proposal Network which shares weights with the later detection network and streamlines the multi-stage R-CNN framework.”
Deep Level Sets: Implicit Surface Representations for 3D Shape Inference (2019, background): “Existing 3D reconstruction methods can be divided into two broad categories: reconstruction from a single image [5], and from multiple images ( e.g . structure from motion [14]).”
Geometric context from a single image (2005, background): “[23]), manual labeling [4, 17], or very specific scenes [9], we want to automatically estimate the 3D geometric properties of general outdoor scenes from a single image.”
Example Based 3D Reconstruction from Single 2D Images (2006, background): “Methods for single image reconstruction commonly use cues such as shading, silhouette shapes, texture, and vanishing points [5, 6, 12, 16, 28].”
LabelMe: Online Image Annotation and Applications (2010, background): “An alternative approach is to ask humans to explicitly label 3-D information [7], [26], [39].”
High speed obstacle avoidance using monocular vision and reinforcement learning (2005, methodology): “Additionally, we experimented with outlier rejection methods such as robust regression (iteratively reweighted using the " fair response " function, Huber, 1981) and support vector regression (Criminisi et al., 2000).”
Single View Reconstruction of Curved Surfaces (2006, background): “Subsequent systems [1, 13] improved the geometric accuracy of SVR— particularly for scenes with multiple vanishing points or planes—as well as demonstrating the extraction of (relative) metric information from uncalibrated pictures, or even reFigure 1.”
3-D Depth Reconstruction from a Single Still Image (2007, background): “Single view metrology [23] assumes that vanishing lines and points are known in a scene, and calculates angles between parallel lines to infer 3-d structure from Manhattan images.”
Efficient Edge-Based Methods for Estimating Manhattan Frames in Urban Imagery (2008, background): “Single-view reconstruction has already been demonstrated (Heuvel, 1998; Criminisi et al., 1999; Denis et al., 2006), but requires human interaction.”
Understanding images of groups of people (2009, neutral): “[8] and Hoiem et al.”
Building a database of 3D scenes from user annotations (2009, background): “More recently, there has been work to develop systems for humans to explicitly label 3D information [15, 6, 22].”
Automatic photo pop-up (2005, background): “[Liebowitz et al. 1999; Criminisi et al. 2000] offer the most accurate (but also the most labor-intensive) approach, recovering a metric reconstruction of an architectural scene by using projective geometry constraints [Hartley and Zisserman 2004] to compute 3D locations of user-specified points giv”
Tracking Multiple Occluding People by Localizing on Multiple Scene Planes (2009, neutral): “[8], we have”
Autonomous MAV flight in indoor environments using single image perspective cues (2011, background): “estimating 3D structure from a single image [34].”
Scene Parsing by Integrating Function, Geometry and Appearance Models (2013, methodology): “The vanishing points are then used to determine the intrinsic and extrinsic calibration parameters [2, 11].”