Distinctive Image Features from Scale-Invariant Keypoints
Citations
108 citations
Cites methods from "Distinctive Image Features from Sca..."
...Then, similarly to [17, 14], we reduce this set even further, by deploying Hough voting in the scaling and rotation transformation space....
[...]
...For instance, in the SIFT [17] scheme, which is widely deployed for this purpose, salient points are detected by a Difference of Gaussians (DOG) function applied in the scale space....
[...]
...Different from RANSAC-based methods, Lowe [17] applied Hough transform to the geometric transformation space to find groups of consistently transformed correspondences prior to estimating the transformation model....
[...]
...We now depart from the set C1vs1 and follow the strategy from [17, 14] to apply a Hough voting scheme in search for dominant ranges of the target transformation parameters, specifically for the rotation and scaling, in the transformation space....
[...]
108 citations
Cites background from "Distinctive Image Features from Sca..."
...As far as human tracking is concerned, tracking-bydetection framework [7], [9], [30] is becoming increasingly popular that uses locally invariant features like SURF [8] or scale invariant feature transformation (SIFT) [31], [32] to recognize target in each frame....
[...]
...Since SURF is known to be computationally more efficient compared to SIFT features, the current review is limited only to SURF-based methods....
[...]
108 citations
Cites methods from "Distinctive Image Features from Sca..."
...The contrast mismatch has a significant effect in SIFT matching....
[...]
...[21] obtain best results to date using SIFT feature matching [18]....
[...]
...Invariant features (e.g., SIFT) are typically used to tackle viewpoint changes....
[...]
...As expected, SIFT and ASIFT are not capable of handling the drastic viewpoint changes (See Fig....
[...]
...Shan et al. [21] obtain best results to date using SIFT feature matching [18]....
[...]
107 citations
Cites background from "Distinctive Image Features from Sca..."
...Indeed such points (frequently indicated as corner points or junctions) usually provide the most conspicuous and useful features for many algorithms of edge analysis and feature extraction [17, 12]....
[...]
107 citations
References
46,906 citations
16,989 citations
13,993 citations
7,057 citations
3,422 citations