Premium
An automated patient recognition method based on an image‐matching technique using previous chest radiographs in the picture archiving and communication system environment
Author(s) -
Morishita Junji,
Katsuragawa Shigehiko,
Kondo Keisuke,
Doi Kunio
Publication year - 2001
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.1373403
Subject(s) - histogram , radiography , correlation , artificial intelligence , matching (statistics) , computed radiography , receiver operating characteristic , computer vision , picture archiving and communication system , computer science , pattern recognition (psychology) , identification (biology) , template matching , image (mathematics) , medicine , radiology , mathematics , statistics , image quality , geometry , botany , biology
An automated patient recognition method for correcting “wrong” chest radiographs being stored in a picture archiving and communication system (PACS) environment has been developed. The method is based on an image‐matching technique that uses previous chest radiographs. For identification of a “wrong” patient, the correlation value was determined for a previous image of a patient and a new, current image of the presumed corresponding patient. The current image was shifted horizontally and vertically and rotated, so that we could determine the best match between the two images. The results indicated that the correlation values between the current and previous images for the same, “correct” patients were generally greater than those for different, “wrong” patients. Although the two histograms for the same patient and for different patients overlapped at correlation values greater than 0.80, most parts of the histograms were separated. The correlation value was compared with a threshold value that was determined based on an analysis of the histograms of correlation values obtained for the same patient and for different patients. If the current image is considered potentially to belong to a “wrong” patient, then a warning sign with the probability for a “wrong” patient is provided to alert radiology personnel. Our results indicate that at least half of the “wrong” images in our database can be identified correctly with the method described in this study. The overall performance in terms of a receiver operating characteristic curve showed a high performance of the system. The results also indicate that some readings of “wrong” images for a given patient in the PACS environment can be prevented by use of the method we developed. Therefore an automated warning system for patient recognition would be useful in correcting “wrong” images being stored in the PACS environment.