Revision as of 06:16, 30 November 2010 by Ksoong (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)



Solution to Q3 of Week 14 Quiz Pool


a. According to the table, we have

$ \begin{align} h[m,n]=&-\frac{1}{8}\delta [m+1,n-1]+\frac{1}{2}\delta [m,n-1]-\frac{1}{8}\delta [m-1,n-1] \\ &-\frac{1}{4}\delta [m+1,n]+\delta [m,n]-\frac{1}{4}\delta [m,n-1] \\ &-\frac{1}{8}\delta [m+1,n+1]+\frac{1}{2}\delta [m,n+1]-\frac{1}{8}\delta [m-1,n+1] \end{align} $

Replace $ \delta [m,n] $ with general input signal $ x[m,n] $ we get the difference equation of the filter.

$ \begin{align} y[m,n]=&-\frac{1}{8}x[m+1,n-1]+\frac{1}{2}x[m,n-1]-\frac{1}{8}x[m-1,n-1] \\ &-\frac{1}{4}x[m+1,n]+x[m,n]-\frac{1}{4}x[m,n-1] \\ &-\frac{1}{8}x[m+1,n+1]+\frac{1}{2}x[m,n+1]-\frac{1}{8}x[m-1,n+1] \end{align} $

b. Place the center of filter (i.e. where m=0,n=0) upon the pixel of image. Multiply h[m,n] with x[m,n] of the correspondent position and sum the value. We can get

$ \begin{align} y[0,0]=&0*h[-1,1]+0*h[0,1]+0*h[1,1]+ \\ &0*h[-1,0]+1*h[0,0]+0*h[1,0] \\ &1*h[-1,-1]+1*h[0,-1]+1*h[1,-1] \\ =&1-\frac{1}{8}+\frac{1}{2}-\frac{1}{8} \\ =&\frac{5}{4} \end{align} $

c. Notice that

$ \mathbf{h}[m,n]= \begin{pmatrix} -\frac{1}{8} & \frac{1}{2} & -\frac{1}{8} \\ -\frac{1}{4} & 1 & -\frac{1}{4} \\ -\frac{1}{8} & \frac{1}{2} & -\frac{1}{8} \end{pmatrix} = \begin{pmatrix} \frac{1}{2} \\ 1 \\ \frac{1}{2} \end{pmatrix} \begin{pmatrix} -\frac{1}{4} & 1 & -\frac{1}{4} \end{pmatrix} $

Therefore \mathbf{h}[m,n] can be separated as outer product of two column vector given by

$ \mathbf{h}[m,n]=\mathbf{h}_1[m]\mathbf{h}_2[n]^T $

where $ \mathbf{h}_1[m]= \begin{pmatrix} \frac{1}{2} \\ 1 \\ \frac{1}{2} \end{pmatrix} $ and $ \mathbf{h}_2[n]= \begin{pmatrix} -\frac{1}{4} \\ 1 \\ -\frac{1}{4} \end{pmatrix} $

Then compute the CTFT of $ h_1,h_2 $ we get

$ H_1(u)=\frac{1}{2}e^{-ju(-1)}+e^{-ju(0)}+\frac{1}{2}e^{-ju(1)} $

$ H_2(v)=-\frac{1}{4}e^{-jv(-1)}+e^{-jv(0)}-\frac{1}{4}e^{-jv(1)} $

According to the Separability property of CSFT, we get

$ H(u,v)=H_1(u)H_2(v)=(1-\frac{1}{2}cosu)(1+cosv) $



Back to Lab Week 14 Quiz Pool

Back to ECE 438 Fall 2010 Lab Wiki Page

Back to ECE 438 Fall 2010

Alumni Liaison

Ph.D. on Applied Mathematics in Aug 2007. Involved on applications of image super-resolution to electron microscopy

Francisco Blanco-Silva