Image Recomposition

The 'Image recomposition ' include two steps: separate objects from original photo (Digital matting) and place it into others (composition). The generalized 'Image Processing' include two aspects: 'Image Enhancement/Adjustment' and 'Image recomposition'. The first aspect change the representation of a image while the second one change the contents.

This software 'Recomposit' (Recompose it) focus on the second aspect: change contents of image, while another software of us: 'Turbo Photo' do the 'Enhancement/Adjustment' work.

There are two basic conceptions you should know about image recomposition: background and foreground(the object). In most case, background is a regular rectangle image, and the foreground have nonregular shape which is seperated from its original image called matting result. The foreground object is hard to be understanded by beginer, it brings a new conception: the mask.

The mask is a channel( usually it is called the alpha channel) with same width and height of foreground, which defines where the foreground pixel should be transparent and where be opaque, or between two status( semitransparent). For transparent pixel, the observer will see the background color; while the opaque pixel, he will see the foreground color.For semitransparent pixel it will blend both foreground color and background color together as output. As a image recomposition software, in its data structure, the foreground object with its mask are contained in a structure called the 'Layer' . The background is a special layer with no mask called 'Base Layer'. The foreground objects, sometime there are more than one objects, exist in different layers over base layer.

In the software, user can manipulate any layers separately, the current one affected called 'Current Layer'. After edit work is done, a merge operation will flat all these layers into one image.

The following figure shows the relationship between layers in 3D space and its representation in user interface.