Title :
Computer recognition of chest image orientation
Author :
Evanoff, Michael G. ; McNeill, Kevin M.
Author_Institution :
Dept. of Radiol., Arizona Univ., Tucson, AZ, USA
Abstract :
A review console displaying digital radiographs should automatically display chest images in their proper orientation to save radiologists´ time. To accomplish this feat, the authors have developed a simple, rapid algorithm that can be implemented in hardware. First, the view of the radiograph is determined. Two profiles from each chest image are compared against a standard to determine if the image contains a posterior-anterior (PA) or lateral view. In the case of PA, the top of the image is determined from the comparisons of the profile. The edge of the heart is found to make sure the image is not displayed as a mirror image. If the image is found to contain a lateral view of the patient, the profile indicates the spine but the top of the image remains to be detected
Keywords :
cardiology; diagnostic radiography; edge detection; image recognition; medical image processing; radiology; chest image orientation; chest image orientation recognition; chest image profiles; computerized image recognition; digital radiograph display; hardware-implementable algorithm; heart edge detection; image top determination; lateral view; mirror image; posterior-anterior view; radiology; review console; spine; standard image; Biomedical imaging; Computed tomography; Digital images; Displays; Hardware; Image recognition; Mirrors; Picture archiving and communication systems; Radiography; Radiology;
Conference_Titel :
Computer-Based Medical Systems, 1998. Proceedings. 11th IEEE Symposium on
Conference_Location :
Lubbock, TX
Print_ISBN :
0-8186-8564-6
DOI :
10.1109/CBMS.1998.701373