ScanFile LuraDocument

Mixed documents are segmented to generate the following three images:

• A binary image containing text
• A foreground image preserving the text colour
• A background image with the text removed

The three images are then compressed using the most appropriate, efficient techniques.

Suitability of LuraDocument

The LuraDocument is particularly suitable for documents scanned at a resolution of 300dpi or higher, containing large image regions together with text.

Separating Documents:

The separation stage uses the segmentation to separate the document into three images of different resolutions.

The text is removed from the original document to generate a colour or grayscale image containing the image regions and the text background. This image is compressed using a specially adapted version of the “LuraWave” wavelet based image coder. The removal of text regions – extremely difficult structures for image coders to compress – enables a significantly higher Compression performance to be achieved.

The removal of image regions from the Bi tonal text image makes high compression rates possible. The Bi tonal text image is compressed with a Luratech developed context-based coder, generating compressed data 10% to 40% smaller than that achieved by CCITT Fax Group 4.

A third image, containing colour for text regions, is generated and compressed using the “LuraWave” wavelet coder. The Bi tonal text image is compressed with the original document resolution while the background and text colour images are coded with a reduced resolution.