NEW APPROACH FOR EMPHYSEMA PATTERN DETECTION IN COMPUTED TOMOGRAPHY IMAGES

Texture experimental analysis is one of the biggest challenges in the field of computer vision and pattern recognition.  Fractal dimension (FD) has proven to be a useful technique in this field to analyze texture patterns. On the other hand, it could be difficult to use this approach to construct descriptors since the fractal magnitude space usually fails in representing adequately the richness of detail present in a unique feature vector. This paper therefore proposes two new algorithms for pattern analysis and detection. The first method applies different Higuchi dimension methods for identifying regions or sections with emphysema pattern problem. The second approach investigates the importance of multi-fractal spectrum for efficient identification and analysis of emphysema patterns in high resolution computed tomography (HRCT) images. The proposed algorithms have been applied to identify the locations and corresponding quantity of the emphysema patterns in HRCT images. Experimental analyses of the results were investigated using statistical techniques to establish the relationship between the ground truth patches and the HRCT image slices. Results revealed that regions with higher FD values contained the presence of emphysema due to the image tissue complexities within the regions, while detection of emphysema failed in those regions with lower or minimum FD values.This research work therefore recommended that both Higuchi’s dimension and multifractal techniques could be accurately used in detection and analysis of regions with emphysema disease