A method is described for the detection and localisation of benign tumours from head MR images. Detection is carried out at the pixel level using a neural network based approach. The results of this are fused with regions, formed by a robust edge-region segmentation, to enable identification of regions corresponding to part of a tumour. The method described achieves a high sensitivity and specificity, and all tumours in the example set are detected.
References
-
-
1)
-
M. Ozkan ,
B.M. Dawant ,
R.J. Maciunas
.
Neural network based segmentation of multi-modal medical images:a comparative and prospective study.
IEEE Trans.
,
3 ,
534 -
544
-
2)
-
B.M. Dawant ,
A.P. Zijdenbos ,
R.A. Margolin
.
Correction of intensity variations in MR images for computer-aidedtissue classification.
IEEE Trans.
,
4 ,
770 -
781
-
3)
-
J.F. Canny
.
A computational approach to edge detection.
IEEE PAMI
,
6 ,
679 -
698
-
4)
-
J.C. Bezdek ,
L.O. Hall ,
L.P. Clarke
.
Review of MR image segmentation techniques using pattern recognition.
Med. Phys.
,
4 ,
1033 -
1048
-
5)
-
Gerig, G., Martin, J., Kikinis, R., Kubler, O.: `Automating segmentation of dual echo MR head data', Proceedings of the 12th international conference IPMI, 1991, p. 175–187.
-
6)
-
Kohonen, T., Kangas, J., Laaksonen, J., Torkkola, K.: `LVQ_PAK: a program package for the correct applicationof learning vector quantisation algorithms', International joint conference on Neural networks, 1992, p. 725–730.
-
7)
-
R.C. Gonzalez ,
R.E. Woods
.
(1993)
Digital image processing.
-
8)
-
M.E. Brummer ,
R.M. Mersereau ,
R.L. Eisner ,
R.R.J. Lewine
.
Automatic detection of brain contours in MRI datasets.
IEEE Trans.
,
2 ,
153 -
166
-
9)
-
L.O. Hall ,
A.M. Bensaid ,
L.P. Clarke ,
R.P. Velthuizen ,
M.S. Silbiger ,
J.C. Bezdek
.
A comparison of neural networkand fuzzy clustering techniques in segmenting magnetic resonance images of the brain.
IEEE Trans.
,
5 ,
672 -
682
-
10)
-
T. Pavlidis ,
Y.T. Liow
.
Integrating region growing and edge detection.
IEEE PAMI
,
3 ,
225 -
233
http://iet.metastore.ingenta.com/content/journals/10.1049/ip-vis_19960690
Related content
content/journals/10.1049/ip-vis_19960690
pub_keyword,iet_inspecKeyword,pub_concept
6
6