TY - JOUR
AU - Maciej Philipp
AU - Agnieszka Anna Tomaka
AB - The article presents a modified approach to cluster based matching. Each cluster from one image is matched with the one taken from the other image. The clusters are described by feature vectors which are used to guide matching process based on optimization. As a result of matching a motion estimation in the image scene is obtained. The proposed modification of the feature vector consists in the introduction of shape descriptors.
BT - Archiwum Informatyki Teoretycznej i Stosowanej
LA - eng
N2 - The article presents a modified approach to cluster based matching. Each cluster from one image is matched with the one taken from the other image. The clusters are described by feature vectors which are used to guide matching process based on optimization. As a result of matching a motion estimation in the image scene is obtained. The proposed modification of the feature vector consists in the introduction of shape descriptors.
PY - 1998
T2 - Archiwum Informatyki Teoretycznej i Stosowanej
TI - Displacement and rotation estimation by cluster matching
ER -