Content based image retrieval from chest radiography databases

Content-based image retrieval (CBIR) has the potential to provide medical doctors with a powerful resource to help make accurate diagnosis. The main goal of CBIR is to efficiently retrieve images that are visually similar to a query image. In this paper we focus on CBIR from chest x-ray image databases. We describe a novel approach for extraction of region of interest (ROI), which is a bounding box of lung fields. The region is achieved by 3×3 key points based on density distribution of the chest x-ray image. Then we propose an elastic reformation registration algorithm based on the control points. A multi-dimensional feature space is used to represent the image content, including texture, edge, and descriptors based on filtering. Experiments results demonstrate that our approach can retrieve visually similar images effectively and increase the diagnostic accuracy.