ѧ
JOURNAL OF SOFTWARE
1999 19 3  Vol.19 No.3 1999



BСͽԾ׷ȱƵ׳еӦ
潨Ҧ
ժҪһƵĿBС任ͽԾ׷ۣڴŹȱƵȡƵƵסָƵ,ؽͼ.ڴ˷УÿBС任ӵƵŹȡָȱٵĸƵϢȻýԾ׷ۣЩϢɸƵŹƵסָƵݣFFT(fast Fourier transform)ؽŹͼ.ʵģ,ؽͼз.
ؼʡԾ׷,ź,ź.
ͼšTP391
Application of B Wavelet and Step Spectrum Analysis 
in Truncated Data Imaging
LUO Jian-hua
Department of Biomedical Engineering Shanghai Jiaotong University Shanghai 200030)
YAO Min
Department of Computer Science and Technology Hangzhou University Hangzhou 310028
Abstract In this paper, the authors propose a fast B wavelet transform method in spectrum domain and step spectrum analysis theory for extracting completed magnetic resonance(MR) spectrum data from the MR truncated data, and the reconstruction image. In this method, the feature information is abstracted from the available low-frequency MR data using B wavelet method, and then high-frequency components are exerted from the feature information and whole spectrum data is recovered by using step spectrum analysis theory(SSAT), at last MR image is reconstructed from the recovered completed spectrum data by FFT(fast Fourier transform). Experiment and simulation results show that this method gives much better reconstructed image than the available method.
Key wordsStep spectrum analysis, signal fantastic points, signal fantastic degree.
ҽѧŹmagnetic resonance image,MRIһĿҽѧͼֶ.ʱ䳤ȺߣӶMRI豸ʮְÿ20200Ԫ֮䣬ԼĹ㷺Ӧ.ǺͼҪָ.ʹŹmagnetic resonanceMRͼ֯粿ֱģ.ȵķУ(1) 豸ָ꣬ʹ豸ɱ(2) źŵƽԭڵӹУйɵźŵõǿ.磺nεӣͳѧ۵㣬߱,ɼźʱn.ɼźʱҪܵ˵ĳƺͻ֯˶źŴĲӰ1.MRIʱɱʾΪt=nmt0nǵӴmһͼɨһͼt0ǵɨʱ.,Сmʮش.СmֻɼMRĵƵƵݣԼɨ.MRȱƵ׳.
MRIϵͳɨƵݲиʣʵͼƵף,øҶ任ʽʡɨ2.MRȱƵͼؽ״ǰδɼƵ㲹䣬ȻøҶ任ַ㷨,ڰƹ˾MRI豸.ȱͼαӰ.Ϊһ⣬Ŀǰҽѧչҵо.ҪķòɵõĵƵƵѵBackpropagationMultilayerԪ磬ȻԪЩMRƵƵݣӶָȱƵ3,4.ַһЧҪMRͼƵܴѵԪʱϳҽѧٴӦԶ.ýԾ׷СϣӵƵƵźȡĸƵָMRȱƵ׵ĸƵ.ǵǰMRûеģȽ.ĵ1ڽܽԾ׷ۣ2ڽMRȱƵ׵㼰ȡ3ڽBСͽԾ׷ؽȱƵMRͼʵּ;4ǽ.
1Ծ׷
Ծ׷ΪźŵƵźŵĸȫȷ.źŵźϵĵΪĵ㣬źŵҲźŵԳ̶ȣźϵĵС.֤κźŶźŵԱʾźƵҲźŵĸȫȷ.
κźŶźŵԱʾ.
x=x1,x2,...,xNTʾһźţNάԿռENеһENһNNάĵλ.ôxԱʾΪ

ei=ei1,ei2,...,eiNT ,  j=1,2,...,N;i=1,2,...,N.
1. i{1,2,...,N}ͪj{1,2,...,N}
ei=ei1,ei2,...,eiNT,
,ôSB={e1,e2,...,eN}NάԿռĽԾ.ʽ

źx=x1,x2,...,xNTڽԾSBµyʾx=SB*y.
y=.x,


׿һֱ任ǳΪԾ任,S.yʾźŵϵ.ǰyķΪybiЪi{1,...,Q}Qyзĸ,źڽԾֵһָӳ̶ȵı.ԣǶźڽԾϵĸӶȦΪ
.
log2Nǹ鷶ӣNźά.磬x=0,0,1,1,1,1,0,0T,

2Ӷ2/3ֱb1=3b2=7,ϵȷֱyb1=1yb2=-1.źxڽԾϵıʾΪ

xʽźȱʾʽ.Ƕ߽иҶ任õ

F(ebi)ǽԾϵĵbiƵףbi㣬ybibi.ԣκźŵƵ׿źŵ㼰Ψһȷ.
2MRȱƵ׵㼰ȵȡ
Ծ׷СϣӵƵƵźȡĸƵȱٸƵȱMRƵؽȫƵؽһMRͼ.սԾ׷ۣźŵ㼰Ⱦ͵ȼڻȡźŵƵ.ܴMRƵƵ׷ȡMRͼȣôܴMRƵƵ׷лָMRƵƵ׷,ӶʹȱƵ׳еõ.ǵ˼·,ѡǡСMRƵƵתΪС任ϵͨ߶ȲⶨСϵ(任ֵ)ļ㣬ͼźŵ5,6ݣǿ.
ʵֱIRϵƽɲ⺯ĿռΪL2(IR)L2(IR)ǶСfL2(IR)Сϵcj,k

Цj,k(x)=2j/2(2jx-k).fL2(IR)ôfFourier任Ϊ
():=(Ff)().
FourierParsevalʽL2(IR)УǵFourier任֮ϵParsevalʽΪ
f(x),(x)=(),j,k(),
()j,k(x)ֱf(x)j,k(x)Fourier任.Сϵcj,kʽ.
cj,k=(),j,k().
Nάɢź֤,Сϵcj,kд
(1)
ˣǿͨźŵƵźŵСϵ.
ȡj,k(u)ΪBСBСĸҶ任.B-õ߶Ⱥ߶ȹС7.m׻B-ΪNm(t)=Nm-1(t)*N1(t)m>2,N1(t)01).丵Ҷ任Ϊ

ʽ(1)õ߶ȺĸҶ任
(2)
Fm(.)ǹNmĹEuler-Frobenius Laurentʽ.߶ȴݺС(x)ĸҶ任
(3)
z=e-i/2.θһj,k(x)ųɵĶСj,k(x)=2j/2(2jx-k)ĸҶ任
j,k()=2-j/2e-i2-jk(2-j).(4)
ʽmBС任ĸҶ任.
ź󲿷ּڵƵ֣һMRƵݵĵƵСϵȷMRͼźŵ㣬磬Сڶ֮һֹƵʵĵƵ.ͼĳʱõ㽫С任ϵģļֵģļֵų߶ȵӶӣ8.С任ϵģļֵų߶ȵӶС8.ݴ˾Ϳȷ㣬Ӱ.
ͨ㷽.źx=x1,x2,...,xNTƵ׵ĵƵõΪbi,i=1,2,...,Q.ǿΪybi(u):=(Fx)(u)bi(u):=(Febi)(u)ֱźxͻebiƵʷǵĹϵд

Qɵǣy=yb1,yb2,...,ybQTѡQƵƵ׷һƵQ=(q1),(q2),...,(qQ)TQͿɾʽQԪ㷽飺
Q.y=Q,



y=-1QQ.
ȷȵĽǶȽڸӶΪǵźţֻҪЦlog2NźŵĵƵƵ׷ͿȫָźŵĸƵ.,Ƶ̫ٿźƯƣ,ƵҪ㹻ȷ.
3ʵּ
ǰýԾ׷MRƵƵ׷ָƵؽͼȱMRݳ񷽷Ծ׷񷽷.ؽ㷨Ĳһ̨586/133΢Ͻе.Ϊ˳ʶԾ׷񷽷Ч3ַʽַò㷨ؽͼ˱Ƚ:(1)ͼֱӱȽ,(2)ͼȽϣ9,10,(3)б9,10.
㷨õʵʴŹȱƵݺͼģŹȱƵݣģݵͼͷģͣͼߴ128128أͼ1(a)ʾ.ʵʴŹȱƵڰƹ˾ṩĺ˴ŹͷƵݣͼߴ256256.ȱƵ׽ȡԭƵݵĵƵ֣ԭƵһ.


(a)(b)(c)

(d)(e)(f)
ͼ1ģƵؽͼͼ
ͼ1У(b)(c)ǼģȱƵݷֱֱӸҶ任ͽԾ׷ͼ.ͼ1(d)(f)ֱͼ1(a)(c)ĵ42ͼ.1ڹһһͼؽʱ3棬Բ㷨ģ׼ŻȱƵʷ棬ģ׼Ż10ݣͽԾ׷бȽϵĽ.йһһ㶼ͼ1(a)Ϊ׼ͼ.
1 㷨Ƚ

㷨һһԳʱ
㷨0.239 878 20.490 902 41
ģ׼Ż0.017 939 70.035 427 82.56
Ծ׷0.000 016 30.000 031 72.56

ͼ2(a)ƵֱӸҶ任ؽͼ.ͼ2(b)(c)ʵȱƵݷֱֱøҶ任ͽԾ׷ͼ.ͼ2(d)(f)ֱͼ2(a)(c)85ͼ.2ڹһһͼؽʱ3棬Բ㷨ģ׼ŻͽԾ׷񷨽бȽϵĽ.йһһ㶼ͼ2(a)Ϊ׼ͼ. 
2 㷨Ƚ

㷨һһԳʱ
㷨0.303 366 20.515 402 71
ģ׼Ż0.018 687 60.039 221 42.56
Ծ׷0.009 063 40.015 821 12.56

ͼȽϿͼ1(c)ͼ1(b)ضӰ٣ӽͼ1(a)ͼ2(c)ͼ2(b)αӰ٣ӽͼ2(a).ͼͼ1(f)ͼ1(e)ضӰҪٵö࣬ͼ1(f)Ĳѻʧܽӽͼ1(d);ͼ2(f)ͼ2(e)ضӰ٣Ҳƽ;ӱָ꿴1ͱ2Ծ׷ͼָ궼ڲ㷨ģ׼Ż.˵Ծ׷ͼȱƵײ㷨ؽͼӽƵؽͼ.Ծ׷ľ;Ȳ㷨ľ;.


(a)(b)(c)

(d)(e)(f)
ͼ2ʵƵؽͼͼ
Ծ׷֮ԱȳĲ㷨ؽͼΪ㷨δɼƵֱ棬ܸƵͼϢ١СϾСͼϢУƵϢʧ.ʣ㷨ȫⲿͼϢ.Ծ׷ȴһȱƵ׶ӦͼıԵԣЩδɼƵݽлָ.Ͳ˲ֱʧͼϢں̶ܴƵײĶͼӰ.ٵ£ؽͼɼ㹤߾ϽƵ׼ȷָ.
4 
͵ǰŹ񾭳ҪȱƵݵͼؽ.ȱƵݵĲ㷨ͼؽͼαӰͼֲڣΪۻȡʱ.ͳĲ㷨ܽ.Ծ׷񣬲۶ʵʵNMRȱƵݣǶԼģNMRȱƵݵʵ,˽Ծ׷һָ߾ȵͼؽ֤ؽͼԭʼͼкܴԻӽ,ͼôͳؽͼ.ؽٶܹٴӦҪ.
оõȻѧϺͨѧƼ㽭ѧܹ߿Ƽչ.潨,1958,ʿڣҪоΪҽѧͼؽ.Ҧ,1954,ʿ,ҪоΪҽѧͼؽ.
ͨѶϵ:潨Ϻ200030,Ϻͨѧҽѧѧϵ
ߵλ潨ϺͨѧҽѧѧϵϺ200030
Ҧݴѧѧ뼼ϵݡ310028
E-mail: jhluo@mail1.sjtu.edu.cn
ο
1Bottomley P, Pykett I L. Biological and medical imaging by NMR. Journal of Magnetic Resonance, 1978,29:335340
2Shepp L. Computerized tomograghy and nuclear magnetic resonance. Journal of Computer Assisted Tomography, 1980,4:94101
3Hui Y, Smith M R. MRI reconstruction from truncated data using a complex domain backpropagation neural network. In: IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing. NewYork: IEEE Press, 1995. 513516
4Yan Hong, Mao Jin-tong. Data truncation artifact reduction in MR imaging using a multilayer neural network. IEEE Transactions on Medical Imaging, 1993,12(1):7377
5Mallat S, Hwang W L. Singularity detection and processing with wavelets. IEEE Transactions on IT, 1992,IT-38(2):617643
6Mallat S, Zhong S. Wavelet transform maxima and multiscale edges. IEEE Transactions on PAMI, 1992,PAMI-14(7):710732
7Chui Charles K. An Introduction to Wavelet. Boston: Academic Press, 1992
8Mallat S G, Hwang Wen Liang. Singularity detection and processing with wavelet. IEEE Transactions on Information Theory, 1992,38(2):617634
9Luo Jian-hua, Lu Wei-xue. Fuzzy multicriterion image reconstruction technology. Chinese Journal of Computers, 1996,19(8):609615
10Luo Jian-hua, Lu Wei-xue. Fuzzy multicriterion NMR image reconstruction technology. Acta Electronica Sinica, 1996,24(7):16
1998-03-24ո壩
