support for float16

Is there a possibility of adding support for a 16-bit float PixelType?  I believe the current floating point options are 32 or 64 bit.  Our microscopy data don't have that high of a resolution, so it would be nice to be able to save a factor of 2 in space.

Thank you!