Re: [boost] [gil] Ideas to make my extension more generic?

From: John Femiani (JOHN.FEMIANI_at_[hidden])
Date: 2007-04-23 21:31:25

Christian,

First off, I think the semantics of color_convert already handle this.
That function maps the min and max value of the source channel type to
the min and max value of the target channel type.  

1. I think you can use type traits. boost::is_integral<T>::value  tells
you if it is an integer data, and boost::is_signed tells you whether or
not it is -- well, signed.  Just #include<boost/type_traits.hpp>.

2. I think you can write a multifunction and pass it into
boost::gil::static_for_each. Create a class and make a templatized
member function operator()(T1 src, T2& dst) that does the special
conversion you want. Use enable_if and disable_if from
<boost/utility.hpp> to specialize the member function.

3. the xxx8_image_t is not a real type -- it is a typedef from
image<pixel<uint8, xxx_layout_t>, false>.  

John Femiani.

-----Original Message-----
From: boost-bounces_at_[hidden]
[mailto:boost-bounces_at_[hidden]] On Behalf Of Christian Henning
Sent: Monday, April 23, 2007 8:38 AM
To: boost_at_[hidden]
Subject: [boost] [gil] Ideas to make my extension more generic?

Hi there, in the course of getting more familiar with gil 2.0 I
started writing a new extension for downsampling (don't if that's the
right term) an source image to its unsigned 8 bit counterpart. You
would use such an extension when trying to save an image that is using
more than 8 bit per channel, e.g. rgb32_image_t. Some file formats
only take unsigned 8bit per channel.

Attached there is the code I have so far. But there are still some
areas where my extension isn't generic enough or even the algorithm
fails due to my own short comings.

Here are some questions I have:

1. How do I figure out if  the src channel type is signed? If it is I
would like to add half the value range to all channels and pixels.

2. When dealing with heterogeneous pixel types how can I initializes
them generically?

3. How do I compute the unsigned 8bit image type of the src image
type? I'm looking for xxx8_image_t.

Thanks ahead. Any ideas or suggestions are welcome.

Christian