block-Sample CT Image：块样品的CT图像.pptVIP

Sample CT Image Central Section OrProjection Slice Theorem Back-projected impulse response Instead of converting to frequency domain, use convolution in the image domain π ∞ ∫ d? ∫ [ g? (R) * c (R) ] ? ( x cos ? + y sin ? - R) dR 0 -∞ Each projection is convolved with c (R) and then back projected. Describe c (R) C (p) = | ? | c (R) = lim 2 (?2 - 4π2R2) / (?2 + 4π2R2)2 ? ?0 ? 0 - ? 0 c (R) = ?0 [ 2 sinc (2 ?0 R) - sinc 2 (?0 R)] Actual Filters : Ram-Lak : Shepp-Logan 1.0 a) Ramachandran-Lakshminarayanan kernel b) Shepp-Logan kernel. Solid lines represent h(?) and circles hk c) Representation of the filter functions corresponding to hR and hS We won’t ask the other player if there is a battleship in a certain peg. Write down how many squares you see a battleship along each column. Then write down how many times you see a battleship along each row. I started it for you. We won’t ask the other player if there is a battleship in a certain peg. Write down how many squares you see a battleship along each column. Then write down how many times you see a battleship along each row. I started it for you. * 2 % attenuation change detectable in film 0.2 % change in attenuation coefficient ∑ ∑ ∑ ∑ ∑ CT, like X-ray measure line integrals x y f(x,y) g(R) = CT detector array output g(R) Radon Transform 1917 Central Section Theorem - Bracewell The transform of each projection forms a line, at that angle, in the 2D FT of f(x,y) f(x,y) g(R) g(R) = ∫ f(x,y) dy We will skip the Algebraic Reconstruction Technique {f(x,y)} = F (u,v) = ∫∫ f(x,y) e -i 2π (ux +vy) dx dy F(u,0) = ∫ [ ∫ f(x,y) dy] e -i 2π ux dx So F(u,0) is the Fourier transform of the projection formed from line integrals along the y direction. v u F(u,0) First, let’s think of our experience on the meaning of F(u,0) in the Fourier transform. Uses a collimator to keep exposure to a slice Builds image from multiple projections We will assume p