Hi, I have had great success using gil, and I have some questions and feedback. I didn't see a way to create an image of floating point values. I use my own typedef as follows, but I would think this should be in gil: typedef boost::gil::pixel<float,gray_layout_t> value32f_pixel_t; It is quite common to use square filters in image processing, and I think it would be useful to overload subimage_view with a fixed offset from all 4 sides. template <typename View> inline View subimage_view(const View& src, int border) { return View(src.width()-border*2,src.height()-border*2,src.xy_at(border,border)); } I also had a need for an "image" constructor that would take a size and allocator, but not initialize the values: image(const point_t& dimensions, std::size_t alignment, const Alloc alloc_in = Alloc()) : _memory(0), _align(alignment), _alloc(alloc_in) { allocate_and_default_construct(dimensions); } The biggest change I would suggest is adding the ability to take a view of a view, a const_view of a const_view, and a const_view of a view. This would make writing functions much easier. Thanks, Chris