Distinctive Image Features from Scale-Invariant Keypoints
Citations
450 citations
Cites methods from "Distinctive Image Features from Sca..."
...This method adopts the conventional SIFT as the local descriptor....
[...]
...The texture feature is represented by the conventional descriptors such as Scale-Invariant Feature Transform (SIFT) [21]....
[...]
...Table 1 shows the search results which demonstrate that the deep learned model is much better than the SIFT feature....
[...]
...The settings of the two models are as follows: (1) FACT + Plate-SIFT....
[...]
...To evaluate the Siamese neural network-based plate verification, we compare it with the conventional handcraft features, SIFT [21]....
[...]
449 citations
Cites methods from "Distinctive Image Features from Sca..."
...SIFT [17] were used to generate tentative corresponding points....
[...]
445 citations
441 citations
Cites background or methods from "Distinctive Image Features from Sca..."
...Lowe (1999, 2004) proposed to select the local extrema of an image filtered with differences of Gaussians, which are separable and hence faster to compute than the Laplacian....
[...]
...Affine-invariant detectors provide higher repeatability for large affine distortions (Lowe 2004; Mikolajczyk and Schmid 2002), but are typically expensive to compute (Mikolajczyk et al. 2005; Moreels and Perona 2007)....
[...]
...Like Lowe (2004)’s DoGs, the filters designed by Agrawal et al. (2008) aim at approximating a Laplacian of a Gaussian filter, though simplified further: In the first step, the filter is reduced to a bi-level filter, i.e., with filter values −1 and 1....
[...]
.../known targets not specified Ferns RANSAC, P-n-P Se et al. (2002) SLAM/trinocular camera DoG [scale, orientation] LSE, Kalman filter Skrypnyk and Lowe (2004) tracking/known scene DoG SIFT RANSAC, non-lin....
[...]
...For each keypoint p, the SIFT algorithm (Lowe 1999; Lowe 2004) first assigns an orientation αp in order to make the descriptor invariant to image rotation: The gradient magnitude m and orientation α are computed for each pixel around p, and a histogram of these orientations, weighted by m and a Gaussian window around p, is computed....
[...]
437 citations
Cites background from "Distinctive Image Features from Sca..."
...The first dataset is SIFT1M from [10], containing 1 million 128-d SIFT features [17] and 10,000 independent queries....
[...]
References
46,906 citations
16,989 citations
13,993 citations
7,057 citations
3,422 citations