ASIFT: A New Framework for Fully Affine Invariant Image Comparison

By Jean-Michel Morel, G. Yu. SIAM Journal of Imaging Sciences, 2009

1,651 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/20fb7391be73988852171e1fd096cb8a46c33a26

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

DeepMon: Mobile GPU-based Deep Learning Framework for Continuous Vision Applications (2017, methodology): “Although SIFT-based algorithm provides the lower output difference ratio (3.875”
A review of 3D reconstruction techniques in civil engineering and their applications (2018, methodology): “ASIFT [33] is the variation of SIFT detector.”
Image Matching from Handcrafted to Deep Features: A Survey (2020, methodology): “Such improvements include a fully affine invariant SIFT detector (ASIFT) (Morel and Yu 2009), a center- surround extremum (Agrawal et al. 2008) strategy feature detector with the Laplace calculation approximated by the proposed bilateral filtering to enhance the efficiency, and the efficient…”
GMS: Grid-Based Motion Statistics for Fast, Ultra-Robust Feature Correspondence (2017, neutral): “Examples includes classic works like SIFT [22], ORB [35], SURF [2], A-SIFT [26], Harris Corners [9] and affine covariant region detectors [25].”
RIFT: Multi-Modal Image Matching Based on Radiation-Variation Insensitive Feature Transform (2019, methodology): “Affine-SIFT (ASIFT) [26] extends SIFT to be invariant to affine transformations by simulating two camera axis direction parameters.”
Structure from motion photogrammetry in physical geography (2016, methodology): “While there are several alternative methods for identifying features (including SURF (Bay et al., 2008), ASIFT (Morel and Yu, 2009), BRIEF (Calonder et al., 2010) and LDAHash (Strecha et al., 2012)), the Scale Invariant Feature Transform (SIFT) object recognition system is used most widely in SfM (L”
Visual simultaneous localization and mapping: a survey (2012, methodology): “Currently there are many variants that improve the performance of the SIFT algorithm, for example: ASIFT, which incorporates invariance to affine transformations (Morel and Yu 2009), BRIEF (Binary Robust Independent Elementary Features) (Calonder et al.”
Accurate Geo-Registration by Ground-to-Aerial Image Matching (2014, methodology): “structures [19][29], but the actual 3D geometry can be more complicated, an assumption that fails for many landmarks (e.”
Generic 3D Representation via Pose Estimation and Matching (2016, background): “The matching score of ASIFT was the number of found correspondences in the test pair given the full images.”
Unmanned aerial systems for photogrammetry and remote sensing: A review (2014, methodology): “Computer vision techniques for automatic tie point generation are based on point detectors and descriptors of the SIFT type (Lowe, 2004) and its many variations or redesigns like SURF (Bay et al., 2008), ASIFT (Morel and Yu, 2009), BRIEF (Calonder et al., 2010) and LDAHash (Strecha et al., 2012).”
Deformable Medical Image Registration: A Survey (2013, background): “Morel and Guoshen proposed an afﬁne invariant version of SIFT [259].”
A review of multimodal image matching: Methods and applications (2021, methodology): “Many variants of SIFT or SURF tend to obtain better performance in computation efficiency, robustness, and discrimination, such as using color or affine information [161,199] or a square root kernel measurement [200] and gradient statistic in different domain sizes [201], just to name a few.”
SAR-SIFT: A SIFT-Like Algorithm for SAR Images (2015, methodology): “To achieve affine invariance, ASIFT (Affine-SIFT) [31] simulates different viewpoints to evaluate two camera axis orientation parameters.”
InLoc: Indoor Visual Localization with Dense Matching and View Synthesis (2019, background): “Consequently, there are works on learning keypoint descriptors [53], [54], [80] or to directly establish matches between images [51], [62], [63].”