GitLab

than necessary when pulling out chips.

and const correctness.

to use it.

particular, this change allows complex valued images to be the output of
extract_image_chips().

incorrect outputs when the requested chip stretched the image unevenly
vertically or horizontally.  This is because we used the best similarity
transform rather than affine transform between the image and the output chip.

like cv_image as input.

--HG--
extra : rebase_source : 3875b192c4c15d82b8106a17020ed53ff922c8ef

unrotated chip.  Also added another constructor to chip_details that takes just
a rectangle.

lower lip since those are really mobile parts of the face.

detections.

correspondences between the chip and the original image.

This is the dlib/image_processing/generic_image.h file.  Then I changed all the
image processing functions so that they use this interface.  All the changes
are very minor, but there are just a lot of them.

Any user code that was using array2d objects to represent images will still
work.  However, this change makes it so that users can use their own custom
image objects with dlib by simply implementing a few global functions for their
image object.

a special version of it that is optimized for this case.

explicitly specify the dimensions of the extracted chips.

but not ENABLE_ASSERTS.

with full_object_detection objects in addition to rectangles.

  describes the exact transformation used.
- Fixed a bug in rotate_image(), it returned a point_transform_affine object
  that did the opposite of what the specification said it returns.

rotate_image_dataset().

rectangles instead of just one.