Read text from image without using tesseract using python3 - opencv

I have a task to read text from image(.png format). I researched that it is possibile using opencv module, tesseract_OCR application, pytesseract module.
As I am on a strict client environment I won't be able to install tesseract_OCR (.exe) application on the host.
I am searching for an approach if it can be done without installing this OCR application. Please suggest me the approach.
For example i have below image:
I want out put as a text:
img=IMG.Convert('RGBA')

Related

ROI selection cv2.selectROI not working in Google Colab

I’m learning opencv for object detection and trying to do all the coding using google colab. And so far it has been working well: I can mount the gdrive and upload file and read the video file just fine.
The selecting ROI by passing video frame into cv2.selectROI() crashed the Colab. The runtime log complained about a plugin xcb not initialising. Further reading reveals the plug in require to be run on local machine, so I wonder if there is a way to run cv2.selectROI or equivalent on colab?

How to convert CorelDraw .WI wavelet-compressed image

I have a large sample of .WI images I need to convert to e.g. JPEGs, but the format now seems defunct.
The mimetype is image/wavelet.
The compression algorithm was developed by Summus, a US company that also now seems defunct.
The last CorelDraw support for the format was under 32-bit Windows. If I go down the hardware route I need to be able to make calls to a server via e.g. REST.
I think under *nix djvulibre might be able to open the files, but I haven't been able to test this yet.
Another option is to re-implement the codec myself.
It would be a nice-have to be able to script this.
Here's an example file http://www.wolfgang-rolke.de/graphics/wavelet.wi

Access exif data from gimpfu plugin

I am trying to access an images exif data from inside a gimpfu plugin, specifically the date a photo was taken, I can do this with PIL,
from PIL import Image
date_taken = Image.open(file)._getexif()[36867]
But when I try to import PIL, the plugin no longer appears inside the gimp menu. No error is being produced, and GIMP starts normally.
Is there a way to get the exif data with gimpfu? Or is there a way to import PIL into the plugin?
If the plugin no longer appears in the Gimp menu, it is likely because there is some problem in it that prevents it from running to the end to register. It can be a syntax error or, in this case, a problem importing PIL. If you are on Linux or OSX this is easy to debug, just start Gimp in a terminal session, error messages from the python interpreter will be displayed there. On Windows you can start Gimp with a --verbose parameter to make it run with an additional console windows for messages, but I'm not sure the Python errors show up there. If they don't, bracket you whole code in a try/except and dump the Exception error to a file. You can use the traceback module to get more info (line number of error, etc) but it could itself be the cause of other import errors...
To answer your initial problem, you can get the EXIF data from Gimp, using:
exifData=image.parasite_find('exif-data').data
However, this is just an array of bytes, and you have to do the parsing yourself (but I'll sure you'll find python code snippets to extract the data you want).

How would you build a YouTube Downloader that doesn't require any client-side install or permission (to run Java for example)?

I am trying to develop an application in java where I don't need anything should be installed or ask permission to install from client side? user just need to install app and thats it.
Open to try with php, jquery as well. so any solutions or hint within these domain will be helpful.
You could go for a php approach fetching the flash video files from youtube and present them to the user (Link to Videofile). Furthermore you could also do serverside flv to xyz conversion if you want to convert the video to another format.

LaTeX: Are there any web frontends available?

As of now the workflow is something like, I import an SVN or a CVS repository and then compile a document locally on my machine to get either a ps or a pdf file. But I was wondering if there is a Web front-end to do all the stuff, like for instance, an editor using which you can edit the file online and then download just the pdf file by compiling it?
Any suggestions?
http://www.scribtex.com/pages/index
http://code.google.com/p/latex-lab/
latex-lab will build on top of the google apps editor base...
scribtex is hosted only it looks like.
Another to add to the list is TeXonWeb.
If you mean online LaTeX compilers, then there are two I know of - at baywifi.com (to PDF) and at ScienceSoft (to several formats). Haven't seen any full editors, though.
There is a CMS based on Latex out there at www.osreviews.net.
The best site I found to produce PDF from LaTeX online is PC Shows.
Verbosus offers an Online LaTeX Editor that supports PDF preview, HTTPS, syntax highlighting, code completion, templates, etc. (Additionally it offers an editor for Octave/Matlab)
This is less of a web-based interface than a simple drag-and-drop cgi script that converts latex syntax to a graphic... www.forkosh.com/mimetex.html
latex-online is a simple open source web service that compiles latex sources/public git repos and returns pdf's. It has both a simplistic web front-end and a command-line tool for interacting with the service - you might find it interesting.
One rather new possibility is https://texlive.net/
You can either interactively edit your documents or you can pass your document via the url to it. E.g. a simple hello world document can be constructed as
https://texlive.net/run?%5Cdocumentclass%7Barticle%7D%0A%5Cusepackage%7Bamsmath%7D%0A%5Cbegin%7Bdocument%7D%0AHello%20world!%0A%5Cend%7Bdocument%7D

Resources