The authors present a new approach for affine distorted planar curve matching and exploit it for occluded object recognition. There are two main contributions in the study: First, a novel affine‐invariant curve descriptor (AICD) based on a new‐defined affine‐invariant signature and its unsigned sum is proposed to represent the local shape of a curve with high distinctiveness. Second, a part‐to‐part curve matching algorithm is developed by combining AICD with a curve segmentation strategy based on inflexion points, which can be applied to object recognition under affine distortions and partial occlusions. Experimental results demonstrate that the proposed method exhibits effectiveness in occluded object recognition better than the state‐of‐the‐art partial curve matching methods.
To solve the mis-clusters caused by the traditional information cut algorithm when it is applied to segment images with gray changes, modified information cut in wavelet domain (W-MIC) algorithm is proposed. First, using the gray relevance and space relevance between image pixels, a modified information cut (MIC) is presented, which utilizes a new Parzen windowing function to evaluate probability density functions, and reduces the effect of gray changes to image segmentation; further, considering the difficulties of selecting the optimal parameter in MIC, the proposed W-MIC can reduce the complexity of parameter selection via the smoothing role of wavelet, and improve segmentation results by fusing low frequency information and high frequency information derived from undecimated wavelet decomposition. Segmentation experiments demonstrate that it effectively decreases the influence of parameters selection, and it can not only avoid the mis-clusters caused by gray changes, but also keep image edges.
Contour-based registration provides a feasible approach to object-based change detection with the development of segmentation techniques in remote sensing. In this paper, an affine-invariant registration algorithm based on orthogonal projection matrices is proposed for object-based change detection. First, we extract the objects of interest using segmentation technique and detect the curvature extreme points as feature points in the contour of each object. Then, for each feature point, we construct its descriptor using the orthogonal project matrix of its affine-invariant neighborhood. Finally, object registration is derived through feature point matching based on the descriptor. Experiments of reservoir change detection demonstrate the proposed algorithm is effective in change detection of remote sensing images.Index Terms-Object-based change detection, affine-invariant registration, orthogonal projection matrix, remote sensing images.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.