This topic and its replies were posted before the current version of LEADTOOLS was released and may no longer be applicable.
#1
Posted
:
Saturday, September 1, 2007 12:10:16 AM(UTC)
Groups: Registered
Posts: 5
hi
Im trying to find the source code for the cleanup demo (one of the document imaging demos) but i cant find it. i would appreciate it if u help me with that.
another question if you dont mind, im trying to ocr a fax document. could u show me what are the most effective image cleanup operations which should be done before ocring.
Thanks in advance
#2
Posted
:
Monday, September 3, 2007 1:36:36 AM(UTC)
Groups: Registered, Tech Support
Posts: 1,326
Was thanked: 1 time(s) in 1 post(s)
Hello,
If you are talking about "API Document Clean-Up Application", the source code of this demo doesn't ship with the SDK. If you would like to ask about the source code of the API Document Clean-Up Application, please ask our sales department at
sales@leadtools.com.
About the cleanup operations, the functions that you need to use depend on the image that you want to OCR.
However, you may try to use the following functions to enhance and cleanup the image before OCR it:
- You can remove the border around the image by using the BorderRemove (1-bit images) or AutoTrim (all types of images) functions.
- To reduce the noise in the image. If the image is colored, you can use one of the following filters:
- Average Filter Bitmap
- Gaussian Filter Bitmap
- Median Filter Bitmap
- Smooth Edges Bitmap
If the image is 1-bit, you can use one of the following filters:
- Despeckle Bitmap
- Smooth Bitmap
- Border Remove Bitmap
- Color Bitmap Rgn
- Dot Remove Bitmap
- Hole Punch Remove Bitmap
- Inverted Text Bitmap
- Line Remove Bitmap
Thanks,
Maen Badwan
LEADTOOLS Technical Support
#3
Posted
:
Tuesday, September 4, 2007 4:01:21 AM(UTC)
Groups: Registered
Posts: 5
hi maen
you are right we need to do diffenet cleanup operations according to the quality of the image. but how can we determine if an image is actually low quality or high quality to decide what actions to do. i guess image resolution (dpi) isn't enough. is there any other measurments which leadtools offer to determine the quality of the image for ocring.
#4
Posted
:
Tuesday, September 4, 2007 11:27:10 PM(UTC)
Groups: Registered, Tech Support
Posts: 1,326
Was thanked: 1 time(s) in 1 post(s)
Hello,
When OCR any image, there are three different factors must be under consideration:
- The shape and order of characters: the characters in the image should be ordered and clear.
- The colors in the image:
Colors in the image decrease the accuracy of the OCR engine when recognizing the Text. To get more accurate results, you need to convert colored images to 1-bit Black and white (2 colors).
You can do this by changing the image to Grayscale 8-bit image using the L_GrayScaleBitmap function, and then change the bits-per-pixel to 1-bit by using the L_ColorResBitmap
- The resolution of the image: The OCR engine requires the images to be clear and high resolution (300 - 400 DPI). If the image resolution is low, you need to increase the DPI of the image.
In general, there is no automatic way to scan the image and decide if it's low or high quality.
Thanks,
Maen Badwan
LEADTOOLS Technical Support
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.