Open AccessFramework for progressive segmentation of chest radiograph for efficient diagnosis of inert regions
Author(s) -
S. Savitha,
N C Naveen
Publication year - 2019
Publication title -
international journal of electrical and computer engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.277
H-Index - 22
ISSN - 2088-8708
DOI - 10.11591/ijece.v9i2.pp982-991
Subject(s) - segmentation , chest radiograph , computer science , artificial intelligence , feature (linguistics) , pattern recognition (psychology) , similarity (geometry) , image segmentation , scale space segmentation , identification (biology) , object (grammar) , computer vision , contrast (vision) , radiography , image (mathematics) , radiology , medicine , linguistics , philosophy , botany , biology
Segmentation is one of the most essential steps required to identify the inert object in the chest x-ray. A review with the existing segmentation techniques towards chest x-ray as well as other vital organs was performed. The main objective was to find whether existing system offers accuracy at the cost of recursive and complex operations. The proposed system contributes to introduce a framework that can offer a good balance between computational performance and segmentation performance. Given an input of chest x-ray, the system offers progressive search for similar image on the basis of similarity score with queried image. Region-based shape descriptor is applied for extracting the feature exclusively for identifying the lung region from the thoracic region followed by contour adjustment. The final segmentation outcome shows accurate identification followed by segmentation of apical and costophrenic region of lung. Comparative analysis proved that proposed system offers better segmentation performance in contrast to existing system.