-
Various changes to make the plugin consistent with the current API. Switch a number of data structures to QT things instead of openCV equivalents. Make FRsdk::OpenCVImageBody construct either an rgb or grayscale representation depending on input image type instead of always building both. Mark failure cases in both the detection and enrollment transforms. Allow the enrollment transform to work even if eye detection hasn't been done yet (the sdk will automatically do detection using the default parameters).
-
This line was intended to check the value of done, not for a null pointer. The code could cause an infinite loop in some circumstances. Signed-off-by: caotto <ottochar@gmail.com>
-
Signed-off-by: caotto <ottochar@gmail.com>
-
The qPrintable macro was mangling a string in br::Context::contains (but only when using visual studio debug mode), but expanding the macro into two lines avoids the issue for some reason. Signed-off-by: caotto <ottochar@gmail.com>
-
PP5 building, bypassing _compare for now added backProject to Context and Transform made pixel happy fixed enroll some final changes to pittpatt added author information
