I'm working on a project to help my visually impaired friend, a python script will first take a screenshot every second and whatever is on the image will be converted to text, and the character which is nearest to the coordinate of curser, will be the output.
User can move the curser anywhere on screen and nearest alphabet to curser will be the output of program.
Don't worry about the form of output, it will be in form audio. But for the sake of simplicity of question lets assume it's in the form of a single character text.
Every tutorial I could find explained how to use OCR dependencies just to convert all the text to a continuous text file.
For my particular application, each alphabet will be associated with a specific co-ordinate. But I just couldn't find a single resource to learn how to identify the location of converted character on the image.
Please enlighten me how to extract the coordinates of a character from an image.
This is a good project. But I think it is a chicken-and-egg problem. You need to have OCR performed by a capable OCR engine (most don't provide coordinates) and the result will have the text and associated coordinates. Your question "how to extract the coordinates of a character from an image" means perform OCR and get coordinates. If performing zonal OCR, i.e. Not the entire screen, you need to know what zone to OCR, and establishing this zone to make sure it includes all necessary text around your mouse location in that zone is probably the biggest challenge. My company at www.wisetrend.com builds such OCR-specialized projects per case. We'll be glad to help in this non-commercial project if you'd like to work jointly.