Local-Global Description of Digital Images for Detection and Analysis of Medical Features

Abstract

This paper presents two descriptors to tackle the existing problems in medical imaging by providing more information to describe different textural structures of digital images. The proposed global and local descriptors can provide more accurate analysis of medical features by using hybrid concatenation approach. Several mathematical models in the form of local and global descriptors have been developed and used in the computation and analysis of medical problems. The experimental results showed that both local and global features are very useful in detection and analysis of biomedical features. The results also indicate that the global descriptor outperforms the earlier approaches and demonstrates high discriminating power and robustness of combined features for accurate classification of CT images.