Library for working with images in C++?


Warning: count(): Parameter must be an array or an object that implements Countable in /home/styllloz/public_html/qa-theme/donut-theme/qa-donut-layer.php on line 274
0 like 0 dislike
23 views
I decided to try to write a program for image recognition (for ebook), or to be more precise — digits (MNIST database). I read that write different people on this topic and decided that you choose a anyhow and that — you need to find a good Libu, then not to rewrite the code 10 times and not curse yourself in a bad choice.

What IDE to write while also not decided on the choice I have Borland C++ Builder 6 and MS Visual Studio 2008 Pro || 2010 Express.


I want to hear the answer to the question what library/class to select on the basis of the following:

1. Easy to install in the described IDE or complicated, but good (read: step by step) disassembled somewhere off site.

2. Speed of operation with images

3. Preferably Russian documentation, can be an option: Russian a bit outdated (the book, for example, or just a website) and very fresh English

4. Good documentation, not just a couple of lines of each function, but also a description of the arguments, returned parameters and

5. Necessarily examples, examples, examples.


PS: if you have ever seen the comparison of lib to work with images, then please share. I vyguglil only a few old articles, and still there until the pile of subjectivity.
by | 23 views

4 Answers

0 like 0 dislike
Honestly I do not quite understand, you need a library to work with graphics or a library which already sharpened by neural networks?
\r
As for the graphics, definitely ImageMagick.
\r
As for your ebook, here to help you an old article, but very clear.
\r
\rNeural networks, pattern recognition (journal of the programmer No. 10, 2001)
by
0 like 0 dislike
OpenCV is the first thing that you need to pay attention, a lot of good examples in the library itself, should be enough for almost everything
by
0 like 0 dislike
Well then, ImageMagick, fast, reliable and constantly developing widely used on large projects.
by
0 like 0 dislike
Do not say anything about the library, but I will tell you a story. A good friend of mine moonlighting that wrote programs for diplomas and korovikov. One girl approached him with a similar problem — she was asked to write a program for text recognition.
\r
The first thing he did was break the text lines by analyzing the density of points in each scanline.
\r
The second step was splitting the rows on the marks, and he took into account that in modern fonts the width of different characters may be different.
\r
Finally found the signs he showed to the user and offered to type a letter or sign, of the image. The programme can thus be studied.
\r
\rThe result was amazing — after the training program showed better results than any existing at that time, the text recognition system (1995).
\r
The program had two drawbacks — a hard reference to the font(s) on which(s) were training. The more fonts understood the program, the lower was the accuracy of recognition.
\r
Since then many years have passed. It is a pity that friend will never come back to this topic. Its a simple program even now would make sense to recognize a great big text, printed in one font, such as books.
\r
By the way, do not understand why the author of the question needs a library to work with images — use monochrome BMP input "raspoznavalka". The meaning of a term paper is not in the graphic libraries, namely, text recognition.
by

Related questions

0 like 0 dislike
5 answers
0 like 0 dislike
3 answers
asked Mar 24, 2019 by Bo_bda
0 like 0 dislike
2 answers
asked Mar 25, 2019 by dtcDev
0 like 0 dislike
2 answers
110,608 questions
257,186 answers
0 comments
27,959 users