Table of Contents
Advertisement
g. Text Extraction
In this block, text is extracted from an image to generate text signals (representing thickness of lines
and hue data). Executing UCR processing based on such text signals enables reproduction of naturally
emphasized edges and fine, black lines.
I
Image area separation
Text signals consist of the following elements:
1. Image Area Separation
In this block, shaded areas and continuous density areas of an image are distinguished, enabling
reproduction of naturally emphasized edges in dotted areas.
2. Line Width Identification
In this block, the width of the area (text) to which a specific pixel belongs is identified, thereby
enabling reproduction of naturally emphasized edges suited to the thickness of characters.
3. Edge Detection
In this block, the position of a specific pixel is identified as being or not being along the edge of a
character through comparison against adjacent pixels.
4. Color Identification
In this block, the hue of a specific pixel is identified. The result of identification may be black, or other
color.
The Bk signal (explained later) is generated based on the above characteristics.
Text extraction
Identification signal
Line width identification
Figure 3-322
I
Text signal
Text
signal generation
Identification signal
Advertisement
Table of Contents
