Using A Bag of Words for Automatic Medical Image Annotation With A Latent Semantic

Authors:
  • RBRiadh Bouslimi
    Riadh Bouslimi
  • AMAbir Messaoudi
    Abir Messaoudi
  • AAAndjalel Akaichi
    Andjalel Akaichi

Abstract
We present in this paper a new approach for the automatic annotation of medical images, using theapproach of "bag-of-words" to represent the visualcontent of the medical image combined with textdescriptors based approach tf.idf and reduced by latent semantic to extract the co-occurrence betweenterms and visual terms. A medical report is composed of a text describing a medical image. First, we areinterested to index the text and extract all relevant terms using a thesaurus containing MeSH medicalconcepts. In a second phase, the medical image is indexed while recovering areas of interest which areinvariant to change in scale, light and tilt. To annotate a new medical image, we use the approach of"bag-of-words" to recover the feature vector. Indeed, weuse the vector space model to retrieve similar medicalimage from the database training. The calculation of the relevance value of an image to the query image isbased on the cosine function. We conclude with an experiment carried out on five types of radiologicalimaging to evaluate the performance of our system of medical annotation. The results showed that ourapproach works better with more images from the radiology of the skull