Taken from their wiki:
libvips is a fully demand-driven, threaded image processing library with no image size limits and with good support for colour. It is mature and documented.
Comparison with other imaging package also available
Though not so pythonic their binding is fully featured and comparable with C command-line version. Personally I use it to manipulate HDR image (exr) and NASA Blue Marble Imagery but unfortunately whenever I build it with direct OpenEXR support, vips become broken. Though other HDR format available (radiance) the later is more widely used in scientific field than in multimedia world. As a workaround vips can utilize ImageMagick (as image import) which recent version support OpenEXR though arguably limited to 16bit whereas exr theoretically could support up to 32bit.
edit: dummy.. for exr to works it must be converted to native vips first a.k.a using im_exr2vips there 32bit preserved! will rebuild the runtime
pyvipsCC.7z Python 2.5-2.7 binding
Vips with ImageMagick (LGPL):
vips_im_std-18.104.22.168z standalone runtime
vips_im_min-22.214.171.124z smaller runtime for used with pygtk aio (extract to site-packages\gtk-2.17.2)
Vips with GraphicsMagick (LGPL):
vips_gm_std-126.96.36.199z standalone runtime
vips_gm_min-188.8.131.52z smaller runtime for used with pygtk aio (extract to site-packages\gtk-2.17.2)
Supported image format:
TIFF, JPEG (JFIF), PPM/PGM/PBM/PFM, PNG, CSV, EXR, HDR (Radiance), Matlab, FITS, VIPS(native) plus many libMagick image formats (SVG supported only through ImageMagick).