An overview and evaluation of the radiologists lounge, a semantic content-based radiographic images retrieval

摘要

With the emergence of professional online communities, professionals in the radiology field needed such technology to discuss and share radiographic images between each other. As medical communities rise, they suffer from semantic gap in the process of retrieving radiographic images. This paper presents The Radiologists Lounge, a professional online community specializing in connecting professionals in the radiology field. It has proposed the use of semantics in data retrieval and in structuring radiographic images, with the aim of retrieving these images on the basis of their semantic content. It is implemented using standard biomedical ontology known as the Unified medical Language System (UMLS). The primary goal of our study is to determine how well the Radiologist Lounge prototype functions as a semantic online community that retrieves images based on their semantic content. We study the performance of Radiologist Lounge measured in; indexing speed, retrieval speed, precision of results, recall of results and F-measure of results. Radiographic images are described on two conceptual layers; semantic and logical. The precision, recall and F-measure of retrieved results are measured on the different layers; logical, semantic, combination of logical and semantic. All results are later compared with the precision, recall and F-measure of the tradition tag-based retrieval. The paper describes the limitations of the Radiologist Lounge, ways to improve its performance, and increase its recall, precision and F-measure.