Consider estimating the F$-$matrix from the sets of corresponding points, $\{$vec$($x$)\_($k$)^($L$)\}\_($k$)=1^($K$)$ and $\{$vec$($x$)\_($k$)^($R$)\}\_($k$)=1^($K$)$ in

the left and right images of a stereo pair. Here vec$($x$)\_($k$)^($L$)$ is a $2-$vector specifying the pixel coordinates in

the left image of the k$^($th $)$ scene point, and vec$($x$)\_($k$)^($R$)$ provides the pixel coordinates for the same scene point

in the right image.

a$)[2$ Point$]$ if there was no noise in the image point vec$($x$)\_($k$)^($L$)$ and vec$($x$)\_($k$)^($R$),$ write out an equation in terms

of F which relates these two positions,

b$)[2$ Point$]$ In the presence of noise positions of the corresponding points, write out an

expression for the squared distance between vec$($x$)\_($k$)^($R$)$ and the epipolar line specified by the

corresponding point vec$($x$)\_($k$)^($L$)($ignoring the noise in vec$($x$)\_($k$)^($L$))$