Matching Widely Separated Views Based on Affine Invariant Regions
Citations
7,057 citations
Cites background or methods from "Matching Widely Separated Views Bas..."
...by Baumberg [2] as well as Schaffalitzky and Zisserman [37]. Tuytelaars and Van Gool [ 42 ]...
[...]
...in applications such as wide baseline matching [37, 42 ], object recognition [10, 25], texture...
[...]
4,024 citations
Cites background from "Matching Widely Separated Views Bas..."
...The recognition quality is evaluated through retrieval on a database with ground truth, showing the power of the vocabulary tree approach, going as high as 1 million images....
[...]
3,362 citations
Cites background from "Matching Widely Separated Views Bas..."
...Local photometric descriptors computed for interest regions have proved to be very successful in applications such as wide baseline matching [37, 42], object recognition [10, 25], texture...
[...]
...Tuytelaars and Van Gool [42] construct two types of affine-invariant regions, one based on a combination of interest points and edges and the other one based on image intensities....
[...]
3,359 citations
Cites methods from "Matching Widely Separated Views Bas..."
...More details about this method can be found in [47, 48]....
[...]
...A more detailed explanation of this method can be found in [45, 48]....
[...]
...The detectors are: (i) the ‘Harris-Affine’ detector [24, 27, 34]; (ii) the ‘Hessian-Affine’ detector [24, 27]; (iii) the ‘maximally stable extremal region’ detector (or MSER, for short) [21, 22]; (iv) an edge-based region detector [45, 48] (referred to as EBR); (v) an intensity extrema-based region detector [47, 48] (referred to as IBR); and (vi) an entropy-based region detector [12] (referred to as salient regions)....
[...]
2,207 citations
Cites background from "Matching Widely Separated Views Bas..."
...Furthermore, Internet imagery provides an ideal test bed for developing robust and general computer vision algorithms that can work effectively “in the wild.”...
[...]
References
28,073 citations
"Matching Widely Separated Views Bas..." refers methods in this paper
...The first method for affine invariant region extraction starts from Harris corner points (Harris and Stephens, 1983) and the edges that can often be found close to such a point (extracted using the Canny edge detector (Canny, 1986))....
[...]
23,396 citations
"Matching Widely Separated Views Bas..." refers methods in this paper
...…rejected most false matches among the region correspondences using the geometric and photometric constraints described above, we apply RANSAC (Fischler and Bolles, 1981) (a robust method based on random sampling) to find a consistent epipolar geometry and to reject the remaining false…...
[...]
...After having rejected most false matches among the region correspondences using the geometric and photometric constraints described above, we apply RANSAC (Fischler and Bolles, 1981) (a robust method based on random sampling) to find a consistent epipolar geometry and to reject the remaining false correspondences....
[...]
...The best known constraint is checking for a consistent epipolar geometry in a robust way, e.g. using RANSAC (Fischler and Bolles, 1981), and rejecting all correspondences not conform with the epipolar geometry found....
[...]
...using RANSAC (Fischler and Bolles, 1981), and rejecting all correspondences not conform with the epipolar geometry found....
[...]
...second image (Gruen, 1985; Super and Klarquist, 1997). However, the search that is involved reduces the practicality of this approach. In contrast, our method is based on the extraction and matching of invariant regions, and hence works on the two images separately, without searching over the entire image or applying combinatorics. This is akin to the approach of Pritchett and Zisserman (1998) who start their wide baseline stereo algorithm by extracting quadrangles present in the image and match these based on normalized crosscorrelation to find local homographies, which are then exploited in a search for additional correspondences....
[...]
16,989 citations
"Matching Widely Separated Views Bas..." refers background or methods in this paper
...In summary, our system differs from other wide baseline stereo methods in that we do not apply a search between images but process each image and each local feature individually (Gruen, 1985; Super and Klarquist, 1997; Schaffalitzky and Zisserman, 2001), in that we fully take into account the affine deformations caused by the change in viewpoint (Lowe, 1999; Montesinos et al., 2000; Schmid and Mohr, 1997; Dufournaud et al., 2000) and in that we can deal with general 3D objects without assuming specific structures to be present in the image (Pritchett and Zisserman, 1998; Tell and Carlsson, 2000)....
[...]
...For instance, Lowe (1999) uses extrema of a difference of Gaussians filter....
[...]
...The consistency of the matches found is tested using semi-local constraints, followed by a test on the epipolar geometry using RANSAC. As shown in the experimental results, the feasibility of affine invariance even on a local scale has been demonstrated. Robust matching is quite a generic problem in vision and several other applications can be considered. Object recognition is one, where images of an object can be matched against a small set of reference images of the same object. The sample set can be kept small because of the invariance. Moreover, as the features are local, recognition against variable backgrounds and under occlusion is supported by this method. Another application is grouping, where symmetries can be found as repeated structures. Image database retrieval can also benefit from these regions, where other pictures of the same scene or object can be found. Here, the viewpoint and illumination invariance gives the system the capacity to generalize to a great extent from a single query image. Finally, being able to match a current view against learned views can allow robots to roam extended spaces, without the need for a 3D model. Initial results for such applications can be found in Tuytelaars and Van Gool (1999), Tuytelaars et al....
[...]
...…Klarquist, 1997; Schaffalitzky and Zisserman, 2001), in that we fully take into account the affine deformations caused by the change in viewpoint (Lowe, 1999; Montesinos et al., 2000; Schmid and Mohr, 1997; Dufournaud et al., 2000) and in that we can deal with general 3D objects without assuming…...
[...]
...Lowe (1999) has extended these ideas to real scale-invariance, using circular regions that maximize the output of a difference of gaussian filters in scale space, while Hall et al. (1999) not only applied automatic scale selection (based on Lindeberg (1998)), but also retrieved the orientation of…...
[...]
13,993 citations
"Matching Widely Separated Views Bas..." refers methods in this paper
...Harris corner points (Harris and Stephens, 1983) are good candidates....
[...]
...Harris corner points (Harris and Stephens, 1983) are good candidates. Apart from the necessary properties of good anchor points mentioned above, they typically contain a large amount of information (Schmid and Mohr, 1998), resulting in a high distinctive power, and they are well localized, i.e. the position of the corner point is accurately defined (even up to sub-pixel accuracy) (Shi and Tomasi, 1994). Instead of using corners, local extrema of image intensity can serve as anchor points as well. To this end, we first apply some smoothing to the image to reduce the effect of noise, causing too many unstable local extrema. Then, the local extrema are extracted with a non-maximum suppression algorithm. These points cannot be localized as accurately as corner points, since the local extrema in intensity are often rather smooth. However, they can withstand any monotonic intensity transformation and they are less likely to lie close to the border of an object resulting in a non-planar region. This last property is a major drawback when working with corner points. Of course, which kind of anchor points perform best also depends on the method used for the region extraction, and how good this method deals with the shortcomings of the anchor points. For instance, for the corner points, the high chance of a non-planar region can be alleviated by constructing a region that is not centered around the corner point. Similarly, regions starting from local intensity extrema should not depend too much on the exact position of the extremum, to overcome the inaccurate localization of these points. Other types of anchor points could be used as well. For instance, Lowe (1999) uses extrema of a difference of Gaussians filter....
[...]
...The first method for affine invariant region extraction starts from Harris corner points (Harris and Stephens, 1983) and the edges that can often be found close to such a point (extracted using the Canny edge detector (Canny, 1986))....
[...]
8,432 citations
"Matching Widely Separated Views Bas..." refers background in this paper
...the position of the corner point is accurately defined (even up to sub-pixel accuracy) (Shi and Tomasi, 1994)....
[...]
...…good anchor points mentioned above, they typically contain a large amount of information (Schmid and Mohr, 1998), resulting in a high distinctive power, and they are well localized, i.e. the position of the corner point is accurately defined (even up to sub-pixel accuracy) (Shi and Tomasi, 1994)....
[...]