Appearance-based Keypoint Clustering

We present an algorithm for clustering sets of detected interest points into groups that correspond to visually dis- tinct structure. Through the use of a suitable colour and tex- ture representation, our clustering method is able to identify keypoints that belong to separate objects or background re- gions. These clusters are then used to constrain the match- ing of keypoints over pairs of images, resulting in greatly improved matching under difficult conditions. We present a thorough evaluation of each component of the algorithm, and show its usefulness on difficult matching problems.

• Type: Conference
• Year: 2009
• Where: CVPR
• Authors: Francisco J. Estrada, Pascal Fua, Sabine Süsstrunk, Vincent Lepetit
• Keywords: Appearance Models, Computer Vision, CVPR 2009, Keypoint Clustering