Shape Determination of Aspired Foreign Body on Pediatric Radiography Images Using Rule-Based Approach

Abstract

Foreign body aspiration (FBA) is a common problem among pediatric population that requires early diagnosis and prompt successful management. Conventionally the radiography image inspection processes are carried out manually by the experts of medical field. Recently automated systems are developed to improve the quality of the radiography images but none of the work carried out especially to determine the characteristics of the foreign bodies in pediatric foreign body aspired radiography images. The proposed approach focuses on characterizing the foreign body shape using sixteen various geometric and edge features. The shapes are determined by using fuzzy logical connectives formed by logically connecting two or more extracted information and a binary decision tree. More than 100 X-Ray radiography images are used to obtained the experimental research. The method proved that the results are more accurate in determining the foreign body shapes as circle, polygon, sharp and irregular automatically with less time.