Document with Decorative Font
Situation description: too many errors during recognition.
(Training.tif) |
FineReader can successfully read texts set in practically any standard font
(with such texts the application is nearly insensitive to printing quality) but it is not meant for
recognising
of texts set in decorative fonts such as: FuturisShadowC,
PragmaticaShadowC, CyrillicGoth etc. Solution: There is a
special recognition mode to increase the recognition quality of this
document: Recognition
with training. This process of training a pattern is: first, recognize a page or two in
the special training mode. The training result is a pattern.
Then FineReader will use this pattern to help to recognize the
rest of the text. The created pattern is only useful for recognising texts
printed in the same
font, in the same size and scanned in the same resolution as the document
you trained your pattern on.
- Draw blocks.(menu
Process>Analyze page layout or click the
button on the Image Tool toolbar).
- Start the Train user pattern mode - click the Train user
pattern
on the Recognition tab, menu Tools>Options,
in the Training group. The default pattern name
("Default") will be displayed in the status line.
- Set the recognition language (English).
- Click the 2-Read button.
- Train your pattern - recognize a page or
two in the Train user pattern mode.
Trained characters are saved in the default pattern. After you have finished
training the pattern, FineReader saves the created pattern (Default.ptn) in
the current batch folder.
- Edit your pattern.
- Turn off the train mode (click the Use user pattern item on the Recognition tab).
|
|