Layout< T > Class Template Reference

Represents tiled layouts of arbitrary numbers of images of different sizes. More...

#include <Image/Layout.H>

Collaboration diagram for Layout< T >:

[legend]

List of all members.

Public Types
enum	Dir { H, V }
Public Member Functions
	Layout (const Image< T > &p)
	Layout (const Layout< T > &p1, const Layout< T > &p2, Dir d)
	Layout (const Image< T > *imgs, const size_t nimgs, Dir d)
	Layout (const ImageSet< T > &imgs, Dir d)
	Layout (const Layout< T > *imgs, const size_t nimgs, Dir d)
bool	initialized () const
Dims	getDims () const
int	getWidth () const
int	getHeight () const
Dir	getDir () const
size_t	numParts () const
const Layout< T > &	part (size_t i) const
const Image< T > &	leafImage () const
Image< T >	render () const
Image< T >	render (const T &bgcol) const
void	renderInto (Image< T > &x, const Point2D< int > p) const

---

Detailed Description

template<class T>
class Layout< T >

Represents tiled layouts of arbitrary numbers of images of different sizes.

Previously, when we wanted to build up a big image of different individual images (e.g. for display), we had two choices:

• use a series of nested concatX() and concatY() calls, which is convenient syntactically but is inefficient because each pixel is copied multiple times, once for each concat call
• make a single output image and use a series of inplacePaste() or inplaceEmbed() calls to put the various pieces in place; this is efficient but requires manual tabulation of the various corner locations of the different tiles

Now, with the Layout class, we get both efficiency and syntactical convenience. Instead of concatX() and concatY(), we use hcat() and vcat() which construct Layout objects (with arbitrary nesting); none of these calls involves any pixel copying. Instead we just keep track of the tile structure that is desired. Then when the final Layout has been constructed, it can be rendered into an Image in a single pass with the render() call.

For example, where previously you would have done

Image<byte> = concatX(e, concatY(concatX(a,b), concatX(c,d)));

now you would do

Image<byte> = hcat(e, vcat(hcat(a,b), hcat(c,d))).render();

Definition at line 76 of file Layout.H.

---

The documentation for this class was generated from the following file:

• src/Image/Layout.H