Content Based Medical Image Retrieval (CBMIR) systems are a new technique which researchers aim to integrate with Computer Aided Diagnosis systems. These systems usually find and retrieve images from a large image-database which have a similar conten
t to a query image. Retrieval is done by extracting the visual features from the query image, formulating them in a features vector, comparing features vector components with those of the images in the database, and then, similarity measures are computed. Based on the similarity measures, images which have a similar content to the query image are retrieved. The introduced analysis study surveys and analyzes the current status of the CBMIR systems, evaluates our findings from this survey, and concludes some specific research directions in this field.
Content based 2Dcerebral digital subtraction angiography(DSA) images retrieval
system has been built. The systemfinds and retrieves images fromcerebral DSA imagedatabase(
Cerebral Sacular Aneurysms) which have a similar content to a query image.
R
etrieval is done by extracting the visual shape features of cerebral saccular aneurysms
from a query image, formulating them in a feature vector, comparing feature vector
components with those of the cerebralDSA images in the database. Similarity measures
using Euclidian distanceare computed,based on the similarity measures, images which
have a similar content to the query image are retrieved. Resolution has been calculated by
finding the ratio between cerebral sacular aneurysm area in first retrieved image to cerebral
sacular aneurysm area in the query image for the eight query process which have been
done, average resolution was 98%. Results indicates that the designed content based image
retrieval could be used to calculate unknown cerebral saccular aneurysms area from a
cerebral saccular aneurysms database images whose areas are known.