| |
Program Functions
113
1, 2, 4, 8, 16, 32, 64, 128. A control point distance of 1 means that every pixel is a control
point; a distance of 128 that there are only four control points at the edges of the image. The
restriction that there have to be control points at the edges of an image might be given up if
more variations of control point distances are needed. So far, the available options were suffi-
cient.
If the number of control points is smaller than the number of pixels, which is usually the
case, this calculation results in a loss of information for the general case. Only for special
cases, namely where the full displacement fields were calculated from the control points’ dis-
placement fields of the same size as the resulting size, there is theoretically no loss of
information involved when going from the full displacement fields to the control point
displacement fields. Practically, there might still be some changes due to rounding. Please also
refer to vResizeCuCv.
vXy2CtrlPtIndex.m
This function is called by the function that handles manually warping. In that function, the
user has to select a control point. The MATLAB function ginput returns y and x coordinates
within an image. These coordinates are given to vXy2CtrlPtIndex, together with information
about distances between control points as well as the number of control points in y and x di-
rections. Based on these inputs, the linear control point index of the closest control point is
calculated. This is therefore the inverse function of vCtrlPtIndex2xy if the input coordinates
that are given to vXy2CtrlPtIndex are control point coordinates. The linear control point index
is defined as described in the description of vCtrlPtIndex2xy.
6.5 Documentation
The documentation directory contains two html (Hyper Text Markup Language) files, which
can be displayed with a web browser, and a pdf (Adobe’s Portable Document Format) file,
|  |
|
| |
|
|
