Untitled

We present a technique for local image representation that is invariant to viewpoint for scenes with arbitrary non-planar shape. We show that generic viewpoint invariance can be achieved, under suitable conditions, although the resulting invariant is not shape-discriminative. Our results serve both to validate existing approaches to local feature detection and description, as well as to complement them where they are not applicable. We illustrate our approach on images of 3-D corners where existing approaches fail.

show abstract

“…reminds what done in [1]. The template is matched 7 to the image by normalized cross correlation (NCC).…”

Section: Feature Detectionmentioning

confidence: 99%

Features for recognition: viewpoint invariance for non-planar scenes

Vedaldi

Soatto

2005

Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1

View full text Add to dashboard Cite

show abstract

“…The survey ... demonstrated that newspapers were not exploited to their full potential as an information resource. 20 Looking back through a different set of lenses, a case can be made that historians may not have been interested in reading them at different times over the past hundred years. 19 At one point Baker claims that "real historians of the late nineteenth and early twentieth century aren't reading the old newspapers very much any more."…”

Section: The Uphill Battle To Protect the Daily Fishwrappermentioning

confidence: 99%

Nicholson Baker, Double Fold: Libraries and the Assault on Paper

Cybulski¹

2001

Microform &Amp; Imaging Review

View full text Add to dashboard Cite

show abstract

“…Detecting objects include invariant corners at the maximum curvature of the boundary chain code and intersection points after polygon approximation [5]. The third category is based on the parameter model that models each pixel of an image [6]. Since the same target in the distorted image often exhibits local distortions and gray differences compared to in the reference image, the second category method that depends on the edge quality, and the third category method that is restricted to the parametric model are unsuitable.…”

Section: Introductionmentioning

confidence: 99%

An Effective Correction Method for Seriously Oblique Remote Sensing Images Based on Multi-View Simulation and a Piecewise Model

Wang

Liu

Zhao

et al. 2016

Sensors

View full text Add to dashboard Cite

Conventional correction approaches are unsuitable for effectively correcting remote sensing images acquired in the seriously oblique condition which has severe distortions and resolution disparity. Considering that the extraction of control points (CPs) and the parameter estimation of the correction model play important roles in correction accuracy, this paper introduces an effective correction method for large angle (LA) images. Firstly, a new CP extraction algorithm is proposed based on multi-view simulation (MVS) to ensure the effective matching of CP pairs between the reference image and the LA image. Then, a new piecewise correction algorithm is advanced with the optimized CPs, where a concept of distribution measurement (DM) is introduced to quantify the CPs distribution. The whole image is partitioned into contiguous subparts which are corrected by different correction formulae to guarantee the accuracy of each subpart. The extensive experimental results demonstrate that the proposed method significantly outperforms conventional approaches.

show abstract

Untitled

Cited by 96 publications

References 26 publications

Features for recognition: viewpoint invariance for non-planar scenes

Features for recognition: viewpoint invariance for non-planar scenes

Nicholson Baker, Double Fold: Libraries and the Assault on Paper

An Effective Correction Method for Seriously Oblique Remote Sensing Images Based on Multi-View Simulation and a Piecewise Model

Contact Info

Product

Resources

About