LTI-Lib (C++ Computer Vision Library) Code

alvarado: @CHG: the global template casting function cond_round_cast had a very short life due to the lack of ANSI C++ overloading of Visual C++. The bug Q240871 vorbids the use of the static_cast or dynamic_cast similar syntax, so that the new function is called condRoundCastTo(from,to).

alvarado: First version of template files for a template functor.

alvarado: @CHG: The base classes for interpolators are not template classes any more. Therefore, the files ltiEquallySpacedSamplesInterpolator.cpp, ltiInterpolator.cpp and ltiVariablySpacedSamplesInterpolator.cpp have being added.

alvarado: @ADD: A simple class to represent line segments (tline<T>) has being added. Besides the storage of the begin and end points, this class provides some simple operations like detection of intersection with other lines or rectangles and minimal distance between a point and the line segment.

alvarado: @ADD: fastLine<T> class has being added. This from tline<T> inherited class stores more attributes of the line (like slope and delta (the direction vector) ), which should help to speed up the algorithms for intersection and distance detection. This first version provides only the interface, but the methods might still be too slow for real requirements. Work in accelerating them is still being done.

wickel: added explicit instantiation file

alvarado: @ADD: an object factory for scalarValuedInterpolation<T> objects has been added.

alvarado: @ADD: class nearestNeighborInterpolator added. It provides the simplest and fastest interpolation method, but with not very good results. If the bilinearInterpolation is too slow for your algorithm, you can try this one only if the quality of the results is not important.

alvarado: Not ltiGenerator accepts more @tokens in the header template. This was required towards template functors templates.

alvarado: Small changes towards a more generic implementation. This class must be completely reviewed, as it works only with float and double types, and should also work with ubyte, int and rgbPixel.