ULTRASOUND LUNG IMAGES AUTOMATIC ANNOTATION

Abstract

This paper aims at an elegant mixture of methods for automatic annotation, detection, clustering, segmentation and retrieval of ultrasound lung images. The annotation of lung images done by using a method called Speeded up Robust Features that is based on the Support Vector Machine classifier. For the features extraction a Fast-Hessian detector was used. The feature matching was performed with SVM. The featured images were clustered using Independent Component Analysis. Micro structure descriptor was used for segmentation of these images while extracting the features. The testing of the developed system was performed using a subset of the IRMA radiographic images. The results provided with the propsed methods were compared with independent methods. Altogether it prospectively constructed an efficient system for automatic medical image retrieval and annotation.