近紅外光學(xué)斷層圖像重建算法研究.pdf

1、蘇州大學(xué)碩士學(xué)位論文近紅外光學(xué)斷層圖像重建算法研究姓名：吳孔培申請學(xué)位級(jí)別：碩士專業(yè)：通信與信息系統(tǒng)指導(dǎo)教師：王加俊20080501StudiesonAlgorithmsforNearinfraredOpticalTomographicImageReconstructionAbstractAbstractNearinfraredopticalimagingtechniqueisanewimagingmodalitydevelopedre

2、centlyformedicalpurposewherethenearinfraredlightsourcesareusuallyusedtoirradiatethetissuestobeimagednisimagingmodalitycallbeemployedtoobtainimagescorrespondingtotheopticalparameterssuch硒thescatteringandtheabsorptioncoeff

3、icientsSincethesecoefficientsarecloselyrelatedtostructuralandfunctionalchangesinthetissues，suchatechniqueCallbeusedespeciallyforearlydiseasediagnosisThisimagingmodalityhaSsomeparticularadvantagesduetoitsfollowingcharacte

4、ristics：nadiationhazards，continuousmonitoringofthepatients，andtheavailableofportableinstrumentsThistechniqueCallbepotentiallyusedforbloodoxygenationmeaSurements，genetherapyandsmallanimalimagingandhenceitisofgreatresearch

5、valuenlccontributionsofthisthesiscanbesummarized嬲follows：Firstlyaccordingtothefactthatimagereconstructionofthefluorescentopticaltomographyinvolvesrepeatedlysolvingofthelarge—dimensionmatrixequationscorrespondingtotheforw

6、ardandtheinverseproblemswhicharecomputafionallyexpensive，especiallyforthecaSewhentherearelargedeviationsintheopticalpropertiesbetweenthetargetandthereferencemedium，awaveletbasedmultiresolutionreconstructionapproachWaSpro

7、posedforthefluorescentopticaltomographicimagereconstructionincombinationwitllaparallelforwardcomputingstrategyinwhichboththeforwardandtheinverseproblemoffluorescentopticaltomographywassolvedinthewaveletdomainSimulationre

8、sultsdemonstratethattheproposedapproachCansignificantlyspeedupthereconstructionprocessandimprovetheimagequalityofthefluorescentopticaltomographySecondlythetechniquesof3Dnearinfraredopticalimagereconstructionwerealsostudi

9、ed111eforwardproblemdescribedbythe3DdiffusionequationWassolvedintheframeworkofthefiniteelementmethodForsuchapurpose，thevolumeofthetissuetobeimagedwasdividedintovoxelsofthetetrahedronswithoutmutualoverlappingFortheinverse

10、problemofthe3Dnearinfraredopticalimaging，aconjugategradientbasedimagereconstructionalgorithmWaSpresentedToenhancetheadaptabilityofthealgorithm，animprovedstrategyfortheconstructionofthesearchingdirectionsWaSpresentedandth