Automatic marker detection from X-ray images

In this paper, we present a novel automatic marker detection method for X-ray images in the framework of machine learning, which is different from those approaches using traditional template matching or fitting algorithms based on prior knowledge. First we propose to use the covariance-based descriptors to effectively represent the marker features in X-ray images. Then we utilize the cascade of LogitBoost classifiers based on covariance features to learn the marker detector, which automatically locates the markers on a X-ray image with a high detection rate and a low false alarm rate. Finally a large amount of experimental results demonstrate that our proposed approach is quite suitable and effective for the marker detection in X-ray images.