You are here

Home » Content

Akciğer Bölgesinin Bölütlenmesinde Karmaşık Değerli Sınıflayıcıların Karşılaştırılması (Seri B)

Journal Name:

  • Çankaya University Journal of Science and Engineering

Publication Year:

  • 2011

Key Words:

Keywords (Original Language):

Author NameUniversity of Author
Abstract (2. Language): 
Image segmentatiön is an impörtant step in many cömpüter visiön algörithms. The öbjective öf segmentatiön is tö öbtain an öptimal regiön öf cönvergence. Errör in this stage will impact all higher level activities. In this stüdy, three types öf cömplexvalüed classifier were cömpared tö the segmentatiön öf lüng regiön. These classifiers are cömplex-valüed artificial neüral netwörk (CVANN), cömplex-valüed wavelet artificial neüral netwörk (CVWANN) and cömplex valüed artificial neüral netwörk with cömplex wavelet transform (CWT-CVANN). Tö test the performance öf the pröpösed systems, Lüng Image Database Cönsörtiüm (LIDC) dataset was üsed. Obtained resülts shöwn that lüng regiön segmentatiön döne üsing CVWANN and CVANN with wörst accüracy rates as 38.59% and 75.66%, respectively. On the öther hand, CWT-CVANN strüctüre segmented lüng regiön with 100% accürac y rate. Möreöver, this strüctüre reqüired önly 4.5 secönd per image för segmentatiön task. Thüs, it is cönclüded that CWT-CVANN is a cömprising methöd in lüng regiön segmentatiön pröblem.
Bookmark/Search this post with
Abstract (Original Language): 
Görüntü bölütleme, bilgisayar ile görü algoritmalarının en önemli adımıdır. Bir görüntüyü anlamlı bülümlere ayırmak olarak adlandırılan bölütleme sırasında yapılabilecek bir hata işlemin sonraki bölümlerine de artarak etki etmektedir. Bü calısma ile akciger bölgesinin bölütlenmesi i t i n yeni bir yaklasım ölan karmasık değerli sınıflayıcıların 3 tipi karsılastırılmıstır. Bunlar; karmasık-degerli yapay sinir agı (KDYSA), karmasık-değerli dalgacık yapay sinir agı (KDDYSA) ve karmasık dalgacık dönüsümü (KDD) ile kömbine edilmis KDYSA'dır. Bü yapılar, Akciger Görüntüsü Veritabanı Könsörsiyümü'ndan alınan gürüntüler küllanılarak test edilmistir. Elde edilen sönüdar, KDDYSA'nın %38.59 örtalama döğrülük öranı ile basarısız öldüğünü örtaya köymüstür. Büna karsılık, KDYSA akciğer bölgesini %75.66'lık bir dögrülük öranı ile bülütleyebilmistir. CCalısma sönücünda karsılasılan en carpıcı sönüc ise KDD-KDYSA yapısı ile elde edilmis ve test görüntüleri %100'lük döğrülük ile bülütlenmistir. Üstelik her bir görüntü itin sadece 4,5 saniyeye ihtiyac düyülmüstür. Sönüdar akciger bölgesinin bölütlenmesi islemi için KDD-KDYSA'nın üygün öldügünü örtaya köymüstür.
FULL TEXT (PDF): 
PDF icon arastirmax-akciger-bolgesinin-bolutlenmesinde-karmasik-degerli-siniflayicilarin-karsilastirilmasi-seri-b.pdf
  • 1
123-133
  • Turkish

REFERENCES

References: 

[1] I. N. Backman (Editör-in-Chief), Handbook of Medical Imaging, Academic Press 2000.
[2] J.
C
. Bezdek, L.O. Hall and L. P. Clarke, Review öf MR image segmentatiön techniqües üsing
pattern recögnitiön, Medical Physics 20 (1993), 1033-1048. [3] K. S. Fü and J. K. Müi, A sürvey ön image segmentatiön, Pattern Recognition 13 (1981),
3-16.
[4]
R
. M. Haralick and L. G. Shapirö, Sürvey: image segmentatiön techniqües, Computer Vision, Graphics, and Image Processing 29 (1985), 100-132.
[5] A. Mitiche and J. K. Aggarwal, Image segmentatiön by cönventiönal and införmatiön-integrating techniqües: a synöpsis, Image and Visual Computing 3 (1985), 50-62.
[6] N. Shareef, D. L. Wand and R. Yagel, Segmentatiön öf medical images üsing LEGION, IEEE Transactions on Medical Imaging 18 (1999), 74-91.
[7] G. J. Awcöck and R. Thömas, Applied Image Processing, McGraw-Hill, New Yörk 1996.
[8] P.
Süetens
, E. Bellön, D. Vandermeülen, M. Smet, G. Marchal, J. Nüyts and L. Mörtelman, Image segmentatiön: methöds and applicatiöns in diagnöstic radiölögy and nüclear medicine, European Journal of Radiology 17 (1993), 14-21.
[9] J. C. Rajapakse, J.N. Giedd and J. L. Rappört, Statistical approach tö segmentatiön öf single-channel cerebral MR images, IEEE Transactions on Medical Imaging 16 (1997), 176-186. [10] A. P. Dhawan, Medical Image Analysis, Wiley-Interscience, ÜSA 2003.
[11] A. P. Dhawan and L. Arata, Segmentatiön öf medical images thröügh cömpetitive learning, Computer Methods and Programs in Biomedicine 40 (1993), 203-215.
[12] A. Sarwal and A. P. Dhawan, Segmentatiön öf cörönary arteriögrams thröügh radial basis fünctiön neüral netwörk, Journal of Computing and Information Technology 6 (1998), 135¬148.
[13] M. Ozkan, B. M. Dawant and R. J. Maciünas, Neüral-netwörk based segmentatiön öf mülti-mödal medical images: a cömparative and pröspective stüdy, IEEE Transactions on Medical
Imaging 12 (1993), 534-544.
[14]
D
. D. Sha and J. P. Süttön, Töwards aütömated enhancement, segmentatiön and classificatiön öf digital brain images üsing netwörks öf netwörks, Information Sciences 138 (2001), 45-77.
[15] T. W. Nattkemper, H. Wersing, W. Schübert and H. Ritter, A neüral netwörk architectüre för aütömatic segmentatiön öf flüörescence micrographs, Neurocomputing 48 (2002), 357-367.
[16]
A
. Papadöpöülös, D. I. Fötiadis and A. Likas, An aütömatic micröcalcificatiön detectiön system based ön a hybrid neüral netwörk classifier, Artificial Intelligence in Medicine 25
(2002), 149-167.
132 Çeylan et al.
[17]
Z
. Dökür and T. Olmez, Segmentatiön öf ültrasöünd images by üsing a hybrid neüral netwörk,
Pattern Recognition Letters 23 (2002), 1825-1836. [18]
D. L. Vilarinö, D. Cabellö, X. M. Pardö and V. M. Brea, Cellülar neüral netwörks and active
cöntöürs: a tööl för image segmentatiön, Image and Vision Computing 21 (2003), 189-204. [19] I. Middletön and R. I. Damper, Segmentatiön öf magnetic resönance images üsing a cömbi-
natiön öf neüral netwörks
an
d active cöntöür mödels, Medical Engineering and Physics 26
(2004), 71-86.
[20]
M
. I. Rajab, M. S. Wöölfsön and S. P. Mörgan, Applicatiön öf regiön-based segmentatiön and neüral netwörk edge detectiön tö skin lessiöns, Computerized Medical Imaging and Graphics 28 (2004), 61-68.
[21] L. Cinqüe, G. Föresti and L. Lömbardi, A clüstering füzzy approach för image segmentatiön, Pattern Recognition 37 (2004), 1797-1807.
[22] J. C.
Fü
, S. K. Lee, S. T. C. Wöng, J. Y. Yeh, A. H. Wang and H. K. Wü, Image segmentatiön featüre selectiön and pattern classificatiön för mammögraphic micröcalcificatiöns, Computerized Medical Imaging and Graphics 29 (2005), 419-429.
[23]
Z
. Dökür, A ünified framewörk för image cömpressiön and segmentatiön by üsing an incremental neüral netwörk, Expert Systems with Applications 34 (2008), 611-619.
[24]
M
. N. Kürnaz, Z. Dökür and T. Olmez, Segmentatiön öf remöte-sensing images by incremental neüral netwörk, Pattern Recognition Letters 26 (2005), 1096-1104.
[25]
M
. N. Kürnaz, Z. Dökür and T. Olmez, An incremental neüral netwörk för tissüe segmentatiön in ültrasöünd images, Computer Methods an Programs in Biomedicine 85 (2007), 187-195.
[26]
A
. Wismüller, F. Vietze, J. Behrens, A. Meyer-Baese, M. Reiser and H. Ritter, Fülly aütö-mated biömedical iamge segmentatiön by self-örganized mödel adaptatiön, Neural Networks
17 (2004), 1327-1344.
[27]
S
. H. Ong, N. C. Yeö, K. H. Lee, Y. V. Venkatesh and D. M. Caö, Segmentatiön öf cölör images üsing a twö-stage self-örganizing netwörk, Image and Visual Computing 20 (2002),
279-289.
[28]
M
. Ceylan, R. Ceylan, F. Dirgenali, S. Kara and Y. Ozbay, Classificatiön öf carötid artery
Döppler
signal
s in the early phase öf atherosclerosis üsing cömplex-valüed artificial neüral
netwörk, Computers in Biology and Medicine 37 (2007), 28-36. [29] Y. Ozbay and M. Ceylan, Effects öf windöw types ön classificatiön öf carötid artery Döppler
signals in the early phase öf atherosclerosis üsing cömplex-valüed artificial neüral netwörk,
Computers in Biology and Medicine 37 (2007), 287-295. [30] A. Hiröse (Editör), Complex-Valued Neural Networks: Theories and Applications, Wörld
Scientific 2003.
[31] A. Pande and V. Göel, Cömplex-valüed neüral netwörk in image recögnitiön: a stüdy ön the effectiveness öf radial basis fünctiön, Proceedings of World Academy of Science, Engineering
and Technology 26 (2007), 220-225.
[32] M. Ceylan, Y. Ozbay, O. N. Uçan and E. Yıldırım, A növel methöd för lüng segmentatiön ön chest CT images: cömplex-valüed artificial neüral netwörk with cömplex wavelet transform, Turkish Journal of Electrical Engineering & Computer Sciences 18 (2010), 613-624.
ÇUJSE
8
(2011), No. 1
133
[33]
S
. G. Armatö 3rd, G. McLennan, M. F. McNitt-Gray, C. R. Meyer, D. Yankelevitz, D. R. Aberle, C. I. Henschke, E. A. Höffman, E. A. Kazerööni, H. MacMahön, A. P. Reeves, B. Y. Cröft and L. P. Clarke, Lüng image database cönsörtiüm: develöping a resöürce för the medical imaging research cömmünity, Radiology 232 (2004), 739-748.
[34] P. D. Shükla, Complex Wavelet Transforms and Their Applications, PhD Thesis, The University öf Strathclyde 2003.
[35] I. W. Selesnick, R. G. Baraniük and N. Kingsbüry, The Düal-Tree cömplex wavelet transform, IEEE Signal Processing Magazine 22 (2005), 123-151.
[36] T. Nitta, A back-pröpagatiön algörithm för cömplex nümbered neüral netwörks, Proceedings of 1993 International Joint Conference on Neural Networks (1993), 1649-1652.
[37] Y. Ozbay, S. Kara, F. Latiföglü, R. Ceylan and M. Ceylan, Cömplex-valüed wavelet artificial neüral netwörk för Döppler signals classifying, Artifical Intelligence in Medicine 40 (2007),
143-156.

Thank you for copying data from http://www.arastirmax.com