ࡱ> 8:7bjbjUU >"??  0----   prrrrrrRr-    r--    --p  p   -  \0       P        rr                   :Resumo O computador e as tcnicas de processamento de imagens desempenham um papel fundamental no mbito da medicina, dado a que facilitam procedimentos manuais que geralmente so tediosos e demandam muito tempo. Na busca de um tratamento mdico eficaz para prevenir ou reverter as leses provocadas pela Neurofibromatose tipo 1 (NF1), enfermidade com evoluo progressiva e imprevisvel, tm sido conduzidos vrios estudos clnicos com diferentes drogas, os quais so acompanhados sistemicamente pelos especialistas atravs da quantificao das leses para verificar se os ensaios clnicos feitos conseguem reduzir os efeitos dessa doena. O atual processo de quantificao dessas leses feito de forma manual, o que resulta uma tarefa rdua, demorada e sujeita a variabilidade do examinador, problema que aumenta na medida em que os indivduos com NF1 tm maior nmero de neurofibromas presentes. O objetivo deste trabalho consiste em desenvolver um procedimento que permita identificar e contar leses em pacientes com NF1 de forma semiautomtica a partir de um conjunto de 275 imagens digitais com diversidade nas caractersticas da cor da pele, concentrao e formato da leso, do banco de dados privado do Ncleo de Estudo de Neurofibromatose do Hospital Universitrio Antnio Pedro (HUAP) de Niteri, Rio de Janeiro, Brasil. A abordagem proposta consiste na segmentao automtica dos neurofibromas cutneos a partir de suas bordas. Para isso foi desenhado um banco de filtros de Gabor e adaptado o algoritmo geral de Level Set conhecido como DRLSE (Distance Regularized Level Set Evolution), onde a partir da seleo visual adequada de um parmetro  QUOTE   obtida a contagem das leses dentro da moldura de papel da imagem de NF1. Tanto quanto sabemos, esta a primeira tentativa de contar de forma semiautomtica leses relacionadas NF1 em imagens. Avaliamos nossa abordagem comparando o nmero de leses detectadas com a contagem manual realizada por dois especialistas treinados em 20 imagens do banco, onde a segmentao da moldura foi bem-sucedida, no se incluem artefatos visveis e apresentam moderada a alta concentrao de neurofibromas cutneos e pequena ocorrncia de leses subcutneas. Para medir a relao bivariada da contagem automtica com a medio manual dos especialistas, utilizou-se um Coeficiente de Correlao Interclasses (Intraclass Correlation Coefficient ou ICC) e o teste t de Student. Os resultados estatsticos mostram que houve concordncia razovel entre os procedimentos semiautomtico e manuais com ICC=0,51 e ICC=0,58 para os dois avaliadores respetivamente. Palavras-chave: Neurofibromatose tipo 1, Filtros de Gabor, Level Set, deteco de bordas, segmentao. Abstract Computer and imaging techniques play a key role in medicine as they facilitate manual procedures that are often tedious and time-consuming. In the search for an effective medical treatment to prevent or reverse the lesions caused by Neurofibromatosis type 1 (NF1), a disease with progressive and unpredictable evolution, several clinical studies with different drugs have been conducted, which are systematically followed up by the specialists through the quantification of the investigate whether the clinical trials have been successful in reducing the effects of this disease. The current process of quantification of these lesions is done manually, which results in an arduous, time-consuming and subject to variability of the examiner, a problem that increases as individuals with NF1 have a greater number of neurofibromas present. The objective of this work is to develop a procedure that allows to identify and count lesions in patients with NF1 in a semiautomatic way from a set of 275 digital images with diversity in the characteristics of the skin color, concentration and shape of the lesion, the database of the Neurofibromatosis Study Center of Antnio Pedro University Hospital (HUAP), Niteri, Rio de Janeiro, Brazil. The proposed approach consists in the automatic segmentation of cutaneous neurofibromas from their borders. For this, a Gabor filter bank was designed and the general algorithm of Level Set known as DRLSE (Distance Regularized Level Set Evolution) was adapted, where from the appropriate visual selection of a parameter , the lesion count was obtained within the frame role of the NF1 image. As far as we know, this is the first attempt to count semiautomatic lesions related to NF1 in images. We evaluated our approach by comparing the number of lesions detected with the manual counting performed by two trained specialists in 20 images of the bank, where the segmentation of the frame was successful, no visible artifacts were included and presented moderate to high concentration of cutaneous and small neurofibromas occurrence of subcutaneous lesions. In order to measure the bivariate relation of the automatic counting with the manual measurement of the specialists, an Interclass Correlation Coefficient (ICC) and Student's t test were used. The statistical results show that there was a reasonable agreement between the semi-automatic and manual procedures with ICC= 0.51 and ICC = 0.58 for the two evaluators respectively. Keywords: Neurofibromatosis type 1, Gabor filter, Level Set, edge detection, segmentation. >^L {|eg9Habt}z|PY̿Ӳި졎|nhykhc5\mH sH  hhchykhcmH sH %h#jhc5;CJ\aJmH sH  h}hchj\hc5\jhl]hcEHUjhl]hcEHU hl]hcjhl]hcU hj\hc h[ShchchIhc5;CJ\aJ"9P$Vh]^V`a$gd^V]^V`gd^$Uh]U^`a$gdj\<P1h:pc. A!"#$% DpDd xD  3 A"b ঵錍3'{XSDEn঵錍3'{XSPNG  IHDR*iQsRGB pHYs+tIDAT(Sc?6U(88%- Vng=!m¶@ V t0޾v*[Hmi iwn]aRe31}.ì41`Lb+J%IENDB`Dd xD  3 A"b ঵錍3'{XSEn঵錍3'{XSPNG  IHDR*iQsRGB pHYs+tIDAT(Sc?6U(88%- Vng=!m¶@ V t0޾v*[Hmi iwn]aRe31}.ì41`Lb+J%IENDB`j 666666666vvvvvvvvv666666>6666666666666666666666666666666666666666666666666hH6666666666666666666666666666666666666666666666666666666666666666662 0@P`p2( 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p8XV~ OJPJQJ_HmHnHsHtH``` j\Normal$dh`a$$CJOJPJQJ_HaJmHsHtH DA D 0Default Paragraph FontRiR 0 Table Normal4 l4a (k ( 0No List PK![Content_Types].xmlj0Eжr(΢Iw},-j4 wP-t#bΙ{UTU^hd}㨫)*1P' ^W0)T9<l#$yi};~@(Hu* Dנz/0ǰ $ X3aZ,D0j~3߶b~i>3\`?/[G\!-Rk.sԻ..a濭?PK!֧6 _rels/.relsj0 }Q%v/C/}(h"O = C?hv=Ʌ%[xp{۵_Pѣ<1H0ORBdJE4b$q_6LR7`0̞O,En7Lib/SeеPK!kytheme/theme/themeManager.xml M @}w7c(EbˮCAǠҟ7՛K Y, e.|,H,lxɴIsQ}#Ր ֵ+!,^$j=GW)E+& 8PK!Ptheme/theme/theme1.xmlYOo6w toc'vuر-MniP@I}úama[إ4:lЯGRX^6؊>$ !)O^rC$y@/yH*񄴽)޵߻UDb`}"qۋJחX^)I`nEp)liV[]1M<OP6r=zgbIguSebORD۫qu gZo~ٺlAplxpT0+[}`jzAV2Fi@qv֬5\|ʜ̭NleXdsjcs7f W+Ն7`g ȘJj|h(KD- dXiJ؇(x$( :;˹! I_TS 1?E??ZBΪmU/?~xY'y5g&΋/ɋ>GMGeD3Vq%'#q$8K)fw9:ĵ x}rxwr:\TZaG*y8IjbRc|XŻǿI u3KGnD1NIBs RuK>V.EL+M2#'fi ~V vl{u8zH *:(W☕ ~JTe\O*tHGHY}KNP*ݾ˦TѼ9/#A7qZ$*c?qUnwN%Oi4 =3ڗP 1Pm \\9Mؓ2aD];Yt\[x]}Wr|]g- eW )6-rCSj id DЇAΜIqbJ#x꺃 6k#ASh&ʌt(Q%p%m&]caSl=X\P1Mh9MVdDAaVB[݈fJíP|8 քAV^f Hn- "d>znNJ ة>b&2vKyϼD:,AGm\nziÙ.uχYC6OMf3or$5NHT[XF64T,ќM0E)`#5XY`פ;%1U٥m;R>QD DcpU'&LE/pm%]8firS4d 7y\`JnίI R3U~7+׸#m qBiDi*L69mY&iHE=(K&N!V.KeLDĕ{D vEꦚdeNƟe(MN9ߜR6&3(a/DUz<{ˊYȳV)9Z[4^n5!J?Q3eBoCM m<.vpIYfZY_p[=al-Y}Nc͙ŋ4vfavl'SA8|*u{-ߟ0%M07%<ҍPK! ѐ'theme/theme/_rels/themeManager.xml.relsM 0wooӺ&݈Э5 6?$Q ,.aic21h:qm@RN;d`o7gK(M&$R(.1r'JЊT8V"AȻHu}|$b{P8g/]QAsم(#L[PK-![Content_Types].xmlPK-!֧6 +_rels/.relsPK-!kytheme/theme/themeManager.xmlPK-!Ptheme/theme/theme1.xmlPK-! ѐ' theme/theme/_rels/themeManager.xml.relsPK] "{#8@0(  B S  ?}[SH,;Ij\2c4egyks#jl]^c@(@@@@4@UnknownG*Ax Times New Roman5Symbol3" Arial7.@CalibriACambria Math"cgcg:r02$Pj\!xxRESUMO jvramiresHelioOh+'0|  8 D P\dltRESUMO jvramiresNormal_WordconvHelio2Microsoft Office Outlook@F#@߆@߆:r՜.+,0 hp|   RESUMO Title  !"#$%&()*+,-.01234569Root Entry F;Data 1Table WordDocument>"SummaryInformation('DocumentSummaryInformation8/CompObjy  F'Microsoft Office Word 97-2003 Document MSWordDocWord.Document.89q  F#Documento do Microsoft Office Word MSWordDocWord.Document.89q