Improving point matching on multimodal images using distance and orientation automatic filtering

Speed Up Robust Features SURF is one of the most popular and efficient methods used for image registration task. In order to achieve a correct registration, a good matching of feature point is required. However in the case of multimodal images, the high and non-linear intensity changes between different modalities led to many outliers (mismatching of detected points) and consequently a fail in the registration. Therefore, in this paper we introduce an efficient method devoted to the detection and removal of such outlier. It's based on an automatic filtering of outliers on both distance and orientation between features points. We tested our proposed method on a set of real multimodal images (4 modalities covering UV, IR Visible and fluorescence images) and compared it to classical SURF as well as SURF followed by RANSAC filtering. The results show that our method outperforms the others regarding all assessment criteria.