- Objective: To research on the mechanism of detection of handwritten alphabets and conversion into editable text with an aim to improve the accuracy of conversion.
- Tools: The Coding of the program was done in Python and script is written in Visual Studio Code with the execution done in Windows Power Shell. Libraries such as OpenCV, Tesseract, Tkinter, Numpy etc. have been used.
- Outcome: As per our objective, we discovered many underlying algorithms and methods in character recognition. We narrowed down to few algorithms which will result in lower error rate in conversion. The accuracy for printed text in image file to text file was an average of 95%, however for handwritten text it came down to 70%, which can be increased by feeding series of training data to the program.
