Distinctive Image Features from Scale-Invariant Keypoints
Citations
136 citations
Cites methods from "Distinctive Image Features from Sca..."
...Our target detection method is based on a modified SIFT [15] implementation that replaces the slow parts of the original SIFT with simpler variants, yet keeping many of the attractive properties of the original approach....
[...]
...For instance, Skrypnyk and Lowe [21] use SIFT descriptors [15] for object localization in AR....
[...]
...The work in this paper builds upon our previous publication [24], where we described modified SIFT [15] and Ferns [17] approaches and created the first real-time 6DOF natural feature tracking system running on mobile phones....
[...]
136 citations
Cites background or methods from "Distinctive Image Features from Sca..."
...the Scale Invariant Feature Transform (SIFT) feature [23], their method was shown to be very robust against...
[...]
...For more details about the SIFT, please refer to [23]....
[...]
...It is a well-known fact that we cannot fully trust the result of RANSAC especially when the number of inliers is limited [23]....
[...]
...As one of the most popular algorithms in computer vision to extract and describe image local features, the SIFT [23] has been shown to be excellently robust against noise distortion and geometric transformations [26], [27]....
[...]
...1In the original implementation [23], C is set as 0....
[...]
136 citations
Cites background from "Distinctive Image Features from Sca..."
...During the past decades, the works for scene classification were mainly based on handcrafted features, such as GIST [1], scale-invariant feature transform [2], and histogram of oriented gradients [3]....
[...]
136 citations
Cites background or methods from "Distinctive Image Features from Sca..."
...2004; Winn and Jojic 2005; Chum and Zisserman 2007; Ling and Soatto 2007) and robust local feature representations (Lowe 2004; Agarwal and Triggs 2006; Lazebnik et al. 2004)....
[...]
...Then we construct a standard n-word visual vocabulary by clustering a random pool of descriptors (we use SIFT (Lowe 2004)) extracted from the unlabeled image dataset, U , and record each feature’s word type....
[...]
...A strength of the affinity propagation method is that non-metric affinities are allowed, and so the authors compare images with SIFT features and a voting-based match, which is insensitive to clutter (Lowe 2004)....
[...]
...…have shown encouraging progress, particularly in terms of generic visual category learning (Weber et al. 2000; Leibe et al. 2004; Winn and Jojic 2005; Chum and Zisserman 2007; Ling and Soatto 2007) and robust local feature representations (Lowe 2004; Agarwal and Triggs 2006; Lazebnik et al. 2004)....
[...]
[...]
136 citations
Cites methods from "Distinctive Image Features from Sca..."
...The most commonly used detectors are SIFT [7], SURF [8], STAR [9], GFTT [10], FAST [11], AGAST [12], and the relatively recently 3 proposed ORB [13], while among the most used descriptors we can mention SIFT, SURF, ORB, BRIEF [14], BRISK [15], and LATCH [16]....
[...]
...And the third one, a seminal work, estimates a sparse map of SIFT features....
[...]
...ORB – Oriented FAST and Rotated BRIEF [13] is another attempt to achieve a scale and rotation invariant BRIEF, as a computationally efficient alternative to SIFT and SURF....
[...]
...Given the high computational cost of SIFT and SURF feature extractors, they are not considered here, since the system is expected to run in real time....
[...]
...The most commonly used detectors are SIFT [7], SURF [8], STAR [9], GFTT [10], FAST [11], AGAST [12], and the relatively recently...
[...]
References
46,906 citations
16,989 citations
13,993 citations
7,057 citations
3,422 citations