kirtanc25 / gujarati_handwritten_character_recognition

Gujarati Handwritten Character Recognition Project.

Gujarati Handwritten Character Recognition

Gujarati Handwritten Character Recognition project having a good accuracy to predict the given character.
For detailed information, check: https://youtu.be/uxZnQrkTGak

Project Summary:

The aim of this project is to develop a mobile application that takes scanned gujarati handwritten image as input and extracts characters (alphabets, digits, symbols) from it. The image can be of a handwritten gujarati document or printed document. The application will recognize the text from the image and convert it into the readable text. Along with this, our application will keep a history of scanned records.

Purpose:

We have found that some people find difficulty in reading gujarati handwritten documents and cannot able to recognize what is written. So, our main purpose is to recognize the gujarati handwritten text and convert it to readable text.

Objective:

1. Collect gujarati handwritten dataset
2. To decide the machine learning model according to the dataset and then train it
3. Testing of the model
4. Create mobile application
5. Integrate model with mobile application

Scope:

This mobile application can be used by anyone to recognize handwritten gujarati text and convert it to a readable format. The mobile application might recognize text with an accuracy up to 90%. The mobile application depends on the scanning quality of camera, if quality is good then it can able to recognize the text easily and with greater accuracy.

Further uses:

The character recognition technique is further used in word recognization. Word recognization is difficult task in Gujarati Handwritten Words, but first word segmentation is done and after that recognition of one-one character might be possible to achieve whole word recognition.
This mobile application can be used by anyone to recognize handwritten gujarati text and convert it to a readable format. The mobile application might recognize text with an accuracy up to 90%. The mobile application depends on the scanning quality of camera, if quality is good then it can able to recognize the text easily and with greater accuracy.