IEEE Transactions on Pattern Analysis and Machine Intelligence
New feature points based on geometric invariants for 3D image registration
International Journal of Computer Vision
IEEE Transactions on Pattern Analysis and Machine Intelligence
A Graduated Assignment Algorithm for Graph Matching
IEEE Transactions on Pattern Analysis and Machine Intelligence
Efficient Algorithms to Implement the Confinement Tree
DGCI '00 Proceedings of the 9th International Conference on Discrete Geometry for Computer Imagery
Efficient computation of new extinction values from extended component tree
Pattern Recognition Letters
Selection of relevant nodes from component-trees in linear time
DGCI'11 Proceedings of the 16th IAPR international conference on Discrete geometry for computer imagery
Interactive segmentation based on component-trees
Pattern Recognition
Dynamic programming algorithms for efficiently computing cosegmentations between biological images
WABI'11 Proceedings of the 11th international conference on Algorithms in bioinformatics
Component-Trees and Multivalued Images: Structural Properties
Journal of Mathematical Imaging and Vision
| Hi-index | 0.00 |
The problem of matching two images of the same objects but after movements or slight deformations arises in medical imaging, but also in the microscopic analysis of physical or biological structures. We present a new matching strategy consisting of two steps. We consider the grey level function (modulo a normalization) as a probability density function. First, we apply a density based clustering method in order to obtain a tree which classifies the points on which the grey level function is defined. Secondly, we use the identification of the hierarchical representations of the two images to guide the image matching or to define a distance between the images for object recognition. The transformation invariance properties of the representations allow to extract invariant image points. Using the identification of the trees, they allow, in addition, to find the correspondence between invariant points even if these have moved locally. Then, we obtain the transformation function as the thin plate interpolation of the corresponding point pairs. On the other hand, if we use tree identification, this enables us to propose several criterias to distinguish between real deformations and noise effects. In practice, we treat, for instance, first coarse trees (with few leaves) and pass to ever refining trees, after. The method's results on real images will be discussed.