qns on http requests, images and OCR using C/C++

[guest_user]

hi, am doing a project which requires me to GET images using http protocol and then perform optical character recognition(OCR) on it.
Q1) Are images normally encoded using some form of encoding like Base64, etc such that I have to decode it before processing?
Q2) If so, what forms of encoding are normally used?
Q3) How do I store the received bytes, what data structure can I use, etc?
Q4) Are there any good books/references on OCR algorithms?

[Zugzwang]

Dear OP,

is there any particular reason why you want to reinvent the wheel? If you want to get images from a HTTP server, that's fine - use a library having the required functionality, for example libcurl. If you really want to do it the hardware, read some basic network book that covers HTTP.

The same goes for OCR - Wouldn't it suffice to run some open-source OCR tool over the image and just collect the output?

[guest_user]

Dear OP,

is there any particular reason why you want to reinvent the wheel? If you want to get images from a HTTP server, that's fine - use a library having the required functionality, for example libcurl. If you really want to do it the hardware, read some basic network book that covers HTTP.

The same goes for OCR - Wouldn't it suffice to run some open-source OCR tool over the image and just collect the output?

Its a school project so I can't use libraries

[Zugzwang]

Its a school project so I can't use libraries

Are you sure? I'm asking because implementing OCR algorithms doesn't look like something you could do in a couple of days or even weeks.

[worksofcraft]

hi, am doing a project which requires me to GET images using http protocol and then perform optical character recognition(OCR) on it.
Q1) Are images normally encoded using some form of encoding like Base64, etc such that I have to decode it before processing?
Q2) If so, what forms of encoding are normally used?
Q3) How do I store the received bytes, what data structure can I use, etc?
Q4) Are there any good books/references on OCR algorithms?

The images are mostly transferred as jpeg or png and these require considerable decoding to decompress into actual pixel data that you can use.

Besides that you would have to know the image file names and paths that you want on the server and that would mean interpreting the HTML which is not trivial.

Your best bet would be to select the area on screen and extract the pixels you want that way.

Now OCR is far from easy because characters can all be different fonts and sizes and different styles too. It is not the sort of thing that can be done as a school project IMO.

[guest_user]

Are you sure? I'm asking because implementing OCR algorithms doesn't look like something you could do in a couple of days or even weeks.

Its a final year project...

[Zugzwang]

Its a final year project...

If you really want to do it the hard way, i.e., implementing all of the following steps yourself

Using HTTP to obtain image files
Decoding PNG and/or JPG image files
Making OCR on it

then you are well-advised by getting books on these subjects. But please check with your advisor whether this is really necessary. Especially the second step, as worksofcraft pointed out, might be a lot of work, and if the aim of the project is to do OCR, also seems to be an unnecessary hassle.

Now to the books. The first stuff in the list is probably well-covered by the Tanenbaum standard book on networks. It should cover HTTP in enough detail to get you started.

For the rest, you can probably start by browsing your local university library.