5
Software like Tesseract (open source) and Adobe Acrobat (proprietary) contain OCR tools that are preconfigured to extract text from a picture. You could look into using these. If you want to figure out and design algorithms to increase the text's readability, then you could go through some literature on image processing and write code to perform the transform functions on the given image.
3
You can use OCR software to convert images of text into true text - this will allow you to render the text at any level of quality you desire. OCR software is often included with image scanners and graphics suites.
Arabic OCR software is available. Even if the target platform doesn't support Arabic, you can OCR on a platform that does and then rasterise the text to a higher quality image for deployment.
If you want to apply transformations such as sharpening or edge detection you can use tools such as ImageMagick
1
Have you tried using something like Photoshop - it is designed for things like this. And, it allows you to batch process so it is ideal if you have multiple files.
Even Photoshop elements allows you to batch but it's limited compared to the full version (but easier on the wallet). Here is an example of batching with Elements.
You can get a free trial of Photoshop Elements.
1Could you show us an example? – Der Hochstapler – 2012-08-07T09:47:33.343
1Is this text only, or text with an image? EG, words over a picture. – Dave – 2012-08-07T09:48:00.350
Post updated with an example (I can also remove the surrounding and keep the text only) – TTem – 2012-08-07T11:23:18.747