OCR Based Compiler for Java and C plus plus

The ?OCR (Optical Character Recognition)  based Compiler for Java and C++ using Android? makes users extract and edit the code which is hand written or in printed form into a mobile phone. This application can help students, teachers as well as programmers to quickly know whether the writt

2025-06-28 16:34:18 - Adil Khan

Project Title

OCR Based Compiler for Java and C plus plus

Project Area of Specialization Artificial IntelligenceProject Summary

The “OCR (Optical Character Recognition)  based Compiler for Java and C++ using Android” makes users extract and edit the code which is hand written or in printed form into a mobile phone. This application can help students, teachers as well as programmers to quickly know whether the written code is error free or not. The application allows user to select two basic languages that is JAVA and C++. After selecting language, user is be able to choose image format. The image format can be either in handwritten or in printed form. The image can be captured through a camera. The image written on the page or on whiteboard should be in neat and understandable form, so that Optical Character Recognition is able to extract and edit the code. It is an offline application needs not the presence of an internet. The main aim is to facilitate to the students while learning, teachers who find difficulty in checking papers etc. and for those needy students who cannot afford laptops and PC.

Project Objectives

The aim and objectives of the project are

Project Implementation Method

To implement “OCR based compiler for JAVA and C++” in Android the languages explored and used are C++, Java, XML and Python. The first challenge was to create a dataset of characters that is to be used by OCR in decision/comparison stage so, to train the dataset  the anaconda with jupyter notebook is configured that is used for accessing the google colab. Python is used on Jupyter notebook to train the dataset and to achieve the processing efficiency Google GPU’s and Floyd-hub  is used which is paid GPU’s for training the dataset. This method gives us the dataset in protobuff (.pb) and tflite (.tflite) format, which is used with real time SDK to improve the response time while extracting the code form image.

Benefits of the Project

Newbie programmers have lot of benefits through this application. Its reduces their time in understanding logic, it helps them dry run their code automatically. In short, it provides one click compilation.

Technical Details of Final Deliverable

Android Application   (APK file)

Trained Model           (.pb , .tflite)

Final Deliverable of the Project Software SystemType of Industry Education , IT , Others Technologies Artificial Intelligence(AI), OthersSustainable Development Goals Quality Education, Partnerships to achieve the GoalRequired Resources
Item Name Type No. of Units Per Unit Cost (in Rs) Total (in Rs)
Total in (Rs) 29500
FloydHUB Equipment11550015500
Real Time SDK Equipment140004000
Miscellaneous Miscellaneous 11000010000

More Posts