Radiology doctors perform text－based image retrieval when they want to retrieve medical images. However

the accuracy and efficiency of such re⁃ trieval cannot keep up with the requirements. An innovative algorithm is being proposed to retrieve similar medical images. First

we extract the pro⁃ fessional terms from the ontology structure and use them to annotate the CT images.Second

thesemanticsimilaritymatrixofontologytermsiscalculat⁃ ed according to the structure of the ontology. Lastly

the corresponding se⁃ mantic distance is calculated according to the marked vector

which con⁃ tains different annotations. We use 120 real liver CT images (divided into six categories) of a top three－hospital to run the algorithm of the program. Resultshowsthattheretrievalindex“Precision”is80.81%

andtheclassi⁃ fication index“AUC(Area Under Curve) ”under the“ROC curve”(Receiv⁃ erOperatingCharacteristic)is0.945.