Project Description
I am looking for a solution to OCR (Optical Character Recognition)
From an image file (athlete in running or cycling), the software should automatically identify the number.
The images are in JPG. Resolution: bigger than 1200x1600 pixels.
The numbering of athletes varies with each event, ie are not standardized and can vary the number of the source, size, color, background color, sometimes skewed, more than one number on page occurs (than all numbers should be found - of course if the are meaningful).
Process should look like this (needs to be covered by solution):
#1 Scripts downloads photos to Amazon S3
#2 OCR is triggered and performed (does not have to be real-time) - Amazon S3
#3 Found numbers are written to the database table (MySQL) + status of scanning is change in db (from to be "to-be-scanned" -> processed -> "finished")
Please include technology which would be used - tesaract(other?) (and additional software if required)
Demo would be appreciated (ie. what technology, additional details) I do not expect "working" demo
Please feel free to ask if you have any questions or need example photos.
Dear sir.
I have read your job posting carefully and become very excited.
As you can see from my portfolio, I am an OCR expert and have rich experience.
"OCR" is my main subject and I have good OCR engine.
In fact, recognizing number on images is not easy, but my engine is great.
I hope to work with you.
Look at my portfolio carefully.
Best regards.
I would use python. tesseract with the pytesseract package, mysql connector, PIL for the basic stuff and opencv if the images are not very basic. in addition to learning machine vision in uni, I also have a working tesseract demo in my ubuntu machine. cheers from hungary!
Hello,
I am .NET developer with 10 years of experience. I have worked on OCR tasks in the past on this site. You may check my feedback section for details.
I have worked on OCR to read captchas, read scanned driving license card in my previous projects. I had used tesseract OCR which is an open source component along with .NET. I can make the solution as web or windows application as per your requirement. Tesseract gives 90% accurate results if the image is clear enough. Kindly conract me to discuss more on the project.
Looking forward to hear from you.
Thanks,
Ramya