... lawA1
Moore's law tells us that the density of silicon integrated devices roughly doubles every 18 months. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... deltaB1
The Kronecker delta is defined as follows $\left\{ \begin{array}{c} \delta_{ij}=1 \mbox{ for } i=j \\ \delta_{ij} = 0 \mbox{ for } i \neq j\end{array}\right.$. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... equationC1
The Einstein convention is used (i.e. indices that are repeated should be summed over). 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... occurE1
Note that the approach would still work if the pure rotation takes place while observing a planar part. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... skewF1
In this case the skew should be given as an angle in the image plane. If the aspect ratio is also known, this corresponds to an angle in the retinal plane (e.g. CCD-array). 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... (IAC) F2
See Section 2.2.3 for details. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... rotationF3
In this case even a projective reconstruction is impossible since all the lines of sight of a point coincide. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... rotationF4
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... distributedF5
This is a realistic assumption since outliers should have been removed at this stage of the processing. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... occlusionsG1
As view point related occlusions we consider those parts of the object that are visible in one image only, due to object self-occlusion. 
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.