Improved contralateral subtraction images by use of elastic matching technique

被引:39
作者
Li, Q [1 ]
Katsuragawa, S [1 ]
Doi, K [1 ]
机构
[1] Univ Chicago, Dept Radiol, Kurt Rossmann Labs Radiol Image Res, Chicago, IL 60637 USA
关键词
contralateral subtraction; generalized Hough transform; snake model; elastic matching; nonlinear image warping;
D O I
10.1118/1.1287112
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 [临床医学]; 100207 [影像医学与核医学]; 1009 [特种医学];
摘要
A contralateral subtraction technique has been developed to assist radiologists in the detection of asymmetric abnormalities such as lung nodules on a single chest radiograph. With this technique, a contralateral subtraction image is obtained by subtracting a right/left reversed ''mirror'' image from the original one. The lesions in the subtraction image may be enhanced because most of the symmetric skeletal structures, such as peripheral ribs, are eliminated. Although the quality of the previous contralateral subtraction images is relatively good, severe misregistration artifacts, mainly due to serious asymmetry of the ribs in the two lungs of the original image, were observed in some cases, and minor misregistration artifacts were also observed in many cases. In this study, we employed three image warping techniques. An initial global warping technique was applied to reduce severe misregistration artifacts in the subtraction image caused by asymmetric rib structures. Additional two iterative warping techniques based on an elastic matching technique were used for accurate registration of the local structures of ribs, so that minor artifacts present in many subtraction images obtained with the previous technique were greatly reduced. With the new technique, the percentage of chest images, which were rated as being of adequate, good, or excellent duality of subtraction images by use of a subjective evaluation method, was improved from 91% to 97%. In particular, the number of cases with excellent quality was greatly increased from 15% to 42%. The contralateral subtraction technique can be used for detection of asymmetric abnormalities, such as lung nodules, pneumothorax, pneumonia, and emphysema, on peripheral lungs in single chest radiographs, and it therefore has potential utility in a large proportion of abnormal chest images. (C) 2000 American Association of Physicists in Medicine. [S0094-2405(00)02308-7].
引用
收藏
页码:1934 / 1942
页数:9
相关论文
共 22 条
[1]
USING DYNAMIC-PROGRAMMING FOR SOLVING VARIATIONAL-PROBLEMS IN VISION [J].
AMINI, AA ;
WEYMOUTH, TE ;
JAIN, RC .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1990, 12 (09) :855-867
[2]
MULTIRESOLUTION ELASTIC MATCHING [J].
BAJCSY, R ;
KOVACIC, S .
COMPUTER VISION GRAPHICS AND IMAGE PROCESSING, 1989, 46 (01) :1-21
[3]
BAJSCY R, 1982, P 6 INT C PATT REC, P351
[4]
Ballard D.H., 1982, Computer Vision
[5]
GENERALIZING THE HOUGH TRANSFORM TO DETECT ARBITRARY SHAPES [J].
BALLARD, DH .
PATTERN RECOGNITION, 1981, 13 (02) :111-122
[6]
Image registration for DSA quality enhancement [J].
Buzug, TM ;
Weese, J .
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS, 1998, 22 (02) :103-113
[7]
Davis L.S., 1975, Comput. Graph. Image Process., V4, P248
[8]
Iterative image warping technique for temporal subtraction of sequential chest radiographs to detect interval change [J].
Ishida, T ;
Katsuragawa, S ;
Nakamura, K ;
MacMahon, H ;
Doi, K .
MEDICAL PHYSICS, 1999, 26 (07) :1320-1329
[9]
JAIN AK, 1989, FUNDAMENTALS DIGITAL, P253
[10]
DIGITAL IMAGE SUBTRACTION OF TEMPORALLY SEQUENTIAL CHEST IMAGES FOR DETECTION OF INTERVAL CHANGE [J].
KANO, A ;
DOI, K ;
MACMAHON, H ;
HASSELL, DD ;
GIGER, ML .
MEDICAL PHYSICS, 1994, 21 (03) :453-461