how to read a text from a image file.

Question

Ayushi Saxena 2016-2-29

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/270571-how-to-read-a-text-from-a-image-file

回答： Sagar Hukkire 2021-1-7

I have a scanned document & I want to extract the text from it & wanted to store the same in some text file... I have no clue how to start this. would be grateful if anyone could help me on this..

Please suggest me is there any way to convert an image file into a text file. Moreover I want to do this without using image processing toolbox.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Walter Roberson 2016-2-29

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/270571-how-to-read-a-text-from-a-image-file#answer_211640

ocr() from the Computer Vision toolbox.

3 个评论
显示 1更早的评论隐藏 1更早的评论

Ayushi Saxena 2016-2-29

Moreover when I check for pixel intensities I am getting 255 for all pixels.

Walter Roberson 2016-2-29

I have not been able to determine why you are doing all of this.

If you are a student working on a student project, then your Student Version license covers Computer Vision and Image Processing toolkits, unless you went for the cheapest possible license, in which case the price for them is about $US10 each.

If you are a Home license user, then the toolkits are about $US45 each.

If you are working on a Professional / Commercial license, then although the cost of the toolboxes is much more, you should take into account the cost of your time to write good OCR functionality. OCR is something that people have been working on for decades -- the Post Office was working on OCR in the 1970s for example. OCR is not easy to do well. It will take many many hours of your time to implement, after having spent many many hours researching OCR papers and books (1.2 million hits in Google Scholar I think it was.)

I do not recommend implementing OCR yourself. Buy the functionality. Or buy third-party software that you can invoke on scanned images to do the OCR. Or look for already implemented OCR in C++ or in Java. It is not worth the effort to do it yourself.

请先登录，再进行评论。