Improvement in Image Reconstruction of Biological Object by EXACT SIRT cell Scanning Technique from Two Opposite sides of the Target

Transcription

1 Vol. 3, Iue. 3, ay.-june pp ISSN: Improvement in Image Recontruction of Biological Object by EXACT SIRT cell Scanning Technique from Two Oppoite ide of the Target Kabita Purkait 1, Sanjib Sarkar 2, Kalyan Adhikary 3 1 (Aociate Profeor, Electronic & Communication Engineering, Kalyani Govt. Engineering College, Wet Bengal, India) 2 (. Tech, Electronic & Communication Engineering, Kalyani Govt Engineering College, Kalyani, Wet Bengal, India) 3 (Att. Profeor, Electronic & Communication Engineering, odern Intitute of Engineering and Technology,Wet Bengal, India) Abtract: In thi paper a cell technique from two oppoite ide of the target i propoed to recontruct the complex permittivity of the biological body uing Exact Simultaneou Iterative Recontruction Algorithm. The biological body i illuminated by an array antenna conit of 15x15 half wave dipole eparated by quarter wave pacing from each other with a beam width of 6 0 operating at 1 GHz. The field are meaured by 20 half wave dipole placed one ide of the biological model for 24 tranmitter poition on other ide of it. The poition of tranmitter and receiver are interchanged and view are taken from two oppoite ide of the target which improve the quality of the recontructed image. The accuracy to dicriminate the dieaed portion from the normal one increae by 5%-10% when recontruction of complex permittivity of the cell ha been done from two oppoite ide than that obtained from one ingle ide of the target. Recontruction of complex permittivity i imulated uing FORCE 209 and reult are preented uing color gradation cale. Keyword: SIRT, double ided, complex permittivity, exact algorithm. I. Introduction Tomography i the pictorial repreentation of unknown cro ection of an object. By thi proce viualization of the internal tructure of an object without the uperpoition of over- and under-lying tructure i poible. Low frequency microwave (about 1 GHz) can be ued for thi purpoe and that why it i called a non-invaive imaging technique. Each organ of a biological ytem ha a unique complex permittivity. It conit of a real part called dielectric contant and imaginary part called lo factor or dielectric conductivity. Complex permittivity depend on the tiue type and it condition. When microwave energy i paed through a biological body incident field at the cell vary with their complex permittivity. Again, complex permittivity in a cell increae with the increae of water content in it. A water content in a cancerou cell increae, complex permittivity alo increae in it compared to it normal tate. Hence recontructed complex permittivity can be ued to detect cancerou cell in early tage. Among different recontruction technique employed to recontruct the complex permittivity of the biological target, Simultaneou iterative recontruction technique how poitive reult. In the pat few year, iteration recontruction technique have been increaingly popular. According to Richmond moment method dielectric medium i divided into large no. of quare cell each of which ha contant electric field intenity and complex permittivity. A ytem of linear equation can be achieved by taking into account that at the centre of each cell total field i equal to incident field and cattered field. Uing perturbation technique [3-4] the received field can be ued to obtain tomographic image. Subequent modification are made in firt order and econd order algorithm [5]. It ha been oberved that the above algorithm fail to recontruct larger model with large number of higher order term (greater than two) and alo fail in cae of maller model with large perturbation. Conidering the limitation of above algorithm, a new exact algorithm [6] ha been developed. Thi algorithm i applied on normal model a well a an dieaed model and recontructed complex permittivity i oberved from ingle ide and two oppoite ide of the target. II. An Exact Algorithm For Large Perturbation The field ditribution in unperturbed homogeneou medium i expreed by the following equation:- [C].[E i ]=[E i ] (1) Where E i i repreented a the incident field at i th cell in the free pace and E i repreent the internal field at the ame i th cell when the medium i aumed to be a homogeneou one having known permittivity ditribution and [c] repreent the coefficient matrix of homogenou medium. When the homogeneou biological target i replaced by the inhomogeneou one, the permittivity value of the cell are perturbed imultaneouly by mall amount i (i=1,2,3..n) and if the correponding change in the internal field are E i then [C ].[E i + E i ]=[E i ] (2) Where [C ] i the coefficient matrix of the inhomogeneou medium Subtracting eq 1 from 2 the change in the electric field can be given by 1597 Page

2 Vol. 3, Iue. 3, ay.-june pp ISSN: n Ei xi Ei x je j ji(0) (3) Where E i i the modified field in the i th cell under perturbed condition, (0) and ji(0) are the determinant and cofactor of (j, i) th element of unperturbed coefficient matrix [C] repectively. x i i the requiite fractional change in the complex permittivity of the cell with repect to aline water i.e x ( ) / ( 1), w are the complex permittivity of i w w aline water(74-j40) and model cell repectively. There will be a reultant change in the cattered field at a particular receiver location owing to the change of internal field at the different cell of the medium caued by perturbation of complex permittivity ditribution. The net change in the cattered field, (k) at the R th receiver location correponding to the k th beam can be determined from the equation E n jr, (0) R x je j (4) Since x i =0 for all receiver cell a they are located in aleline water. If E Rml (k) denote the cattered field intenity at the R th receiver location for the kth beam in the inhomogeneou numerical model and E Rol (k) denote the calculated cattered field intenity at the ame receiver location for the ame k th beam for the aumed known homogeneou permittivity_ ditribution for the object, then the reultant change in cattered field intenity, E ( k) at a particular receiver location i expreed in term of the unknown variable xj (i.e. the k requiite fractional change of unknown permittivity from the aumed initial trial olution of permittivity), relevant cofactor and determinant of coefficient matrix correponding to the homogeneou medium and perturbed internal field correponding to the inhomogeneou model. Therefore, olving previou equation the total change in the cattered field at + m jr, (0) ER ( k) ERml ( k) ERol ( k) xe j j (5). III. Comparative Study between Single Sided Scanning and ultiple Sided Scanning 3.1 Numerical model The numerical model under tudy i a biological object, rectangular in hape. The model contain 360 cell of ize 1 q.cm and conit of different human organ viz. liver (46-j10), mucle (50-j23), mucle type material (40-j23) and fat (25- j5). It i urrounded by 340 cell of aline water. It i illuminated by 24 tranmitter antenna which i deigned with ( each of 15x15 dipole array antenna) of beam width 6 0 and the radiation i received by 20 half wave dipole acting a receiver. The ditance between the tranmitter and the receiver i 50cm. The total arrangement i immered in water to get better impedance matching and mall antenna ize.[5]. E R Fig 1 Block diagram of the propoed experimental et up 3.2 Two Oppoite Sided of the Biological Target The quality of the recontructed image i improved further by incorporating two oppoite ided technique which i dicued below in brief:for the different poition of tranmitter and receiver complex permittivity i calculated and average i done. A modified cell technique i adopted [7-8] where the beam width of the tranmitting antenna i taken a 6 0. Thu the number of cell where change in internal field take place due to change of complex permittivity in a particular cell i reduced and thereby reduce the error caued by the proce of SIRT algorithm itelf [3,4,5,6] Page

3 Vol. 3, Iue. 3, ay.-june pp ISSN: Next, the poition of the tranmitter and receiver are interchanged and recontructed complex permittivity of each cell i calculated again by uing equation(5). The image of the biological target i recontructed by uing the average value of the complex permittivity in each cell obtained by cell technique from two oppoite ide of numerical biological object. Employing thi, different value of recontructed complex permittivity are found for the two different poition of tranmitter and receiver. Then the average value i calculated. 4.2 Figure and Table Uing the recontructed algorithm for all above cae, the experimental data are imulated [9] and correponding image are hown below: Fig 2 color gradation for real and imaginary parameter repectively IV. Recontruction of normal model Fig 3.1 Fig 3.2 Fig 3.3 Fig 3.4 Fig 3.5 Fig 3.6 Fig 3.1 &3.2: Real and imaginary value of complex permittivity of normal model; Fig 3.3 & 3.4: Recontructed real and imaginary value of complex permittivity of normal model uing ingle ided view; Fig 3.5 & 3.6 Recontructed real and imaginary value of complex permittivity normal model uing two oppoite ide view; 1599 Page

4 Vol. 3, Iue. 3, ay.-june pp ISSN: V. Recontruction of Dieaed model The model under tudy i the ame a that conidered in earlier cae, except it liver region i aumed to be affected and hence characterized by a different value of complex permittivity (48-j12) [1] where a for normal liver it i aumed to the (46-j10). Fig 4.1 Fig 4.2 Fig 4.3 Fig 4.4 Fig 4.5 Fig 4.6 Fig 4.1 &4.2: Real and imaginary value of complex permittivity of dieaed model; Fig 4.3 & 4.4: Recontructed real and imaginary value of complex permittivity of dieaed model uing ingle ided view; Fig 4.5 & 4.6 Recontructed real and imaginary value of complex permittivity dieaed model uing two oppoite ide view; Table 1 Average value of permittivity in different organ of the normal model for different cae Different organ of model Average value of complex permittivity of different organ Normal model Recontructed normal model Uing ingle ide Uing two oppoite ide Fat 25-j j j4.56 ucle 53-j j j25.62 ucle aterial 35-j j j13.92 Liver 46-j j j9.71 Water 76-j40 76-j10 76-j Page

5 Vol. 3, Iue. 3, ay.-june pp ISSN: Table 2 Average value of permittivity in different organ of the dieaed model for different cae Different organ of model Average value of complex permittivity of different organ Dieaed model Recontructed normal model Uing ingle ide Uing two oppoite ide Fat 25-j j j4.56 ucle 53-j j j29.07 ucle aterial 35-j j j15.43 Liver 48-j j j11.13 Water 76-j40 76-j10 76-j10 VI. CONCLUSION In thi paper an overall improvement in recontructed image of the biological object ha been obtained when Exact SIRT cell technique i applied from two oppoite ide of the target.from normal model, the accuracy in recontructed image of the real part of complex permittivity i increaed from ( %) to (92.59%-99.31%) when view are taken from two oppoite ide. In cae of dieaed model the recontructed image of the imaginary part of the complex permittivity for liver region i far better in double ided cell technique (99.13%) than obtained from ingle ide (70%). The improvement how that the image quality will be more accurate it will be taken from all ide. REFERENCES [1] Suan R Smith & Kenneth R Foter, "Dielectric propertie of low-water-content tiue", Phy. ed. Biol., Vol. 30, No. 9, pp ,1985. [2] H.Richmond, IEEE Tran.Antenna Prop. Vol. AP-B,pp ,1965. [3] A N Dana & B. Bondopadhyay. Proc. of IEEE; Vol. 74, pp [4] A N Datta & B. Bondopadhyay.Innov Tech Biol. ed, Vol. 8. pp [5] K.Purkait. A.N.Datta, An Improved form of Iterative Recontruction Algorithm for Firt Order and Second Order icrowave Image Recontruction. Indian Journal of Pure & Applied Phyic, CSIR, Vol.34.pp [6] K.Purkait & A.N.Datta, "An Exact Algorithm for icrowave Tomography", preented at ympoium HOT-2003,1NRAPHEL.C.U.3rd-5th Feb [7] Kabita Purkait and Kalyan Adhikary Application of odified Cell Scanning Technique in Exact Algorithm for edical Diagnoi, IJTS Vol- 21,pp ,October [8] Kabita Purkait and Kalyan Adhikary, odification in field meaurement applied on exact algorithm for biomedical imaging, IJER, Vol. 2, Iue- 3, pp , ay [9] Force 209 FORTRAN Compiler, Page