I have developed an Android Application, which gives user the opportunity to localize the text regions in the Scene Image. For the localization purpose, I have used Morphological Operations-based algorithms, as found by my research and implemented that using OpenCV and Java. I have also used MSRA-TD 500 dataset to develop a Weka-trained module, from the features extracted by the GLCM for the classification of Chinese and English texts.

You can see my whole project :

https://github.com/djdhar/App-for-Scene-Text-Detection-and-Classification

Link of the research paper published :

https://www.igi-global.com/article/multilingual-scene-text-detection-using-gradient-morphology/258252

I have developed an Web Application and also a Tkinter-based Desktop GUI. It takes doctor's prescription as input to extract out possible text regions by removing out the lines, tables and other noises by efficient Morphology based operations. Then with the help of supervised learning (RandomForest), on the basis of five features, it classifies, printed, handwritten, mixed texts and non-texts with a high accuracy, as obtained by my research.

You can see the whole project for Desktop GUI :

https://github.com/djdhar/Handwritten-and-Printed-Text-Classification-in-Doctors-Prescription

You can see the whole project of the Web App I Developed :

https://github.com/djdhar/Web-App-for-Prescription-Text-Classification

The Web App Link : https://hpdoc.herokuapp.com/

I have developed an Desktop Application, where the graph-path finding algorithms can be visualized. In this application, user can set a start node and an end end node and the obstacles, and depending upon the algorithms chosen by the user, we can visualize the path-way of the traversal.

DFS, BFS and Bellman–Ford’s, Floyd–Warshall, and Dijkstra's Shortest Path Algorithm can be visualized (Algorithms are implemented from scratch)

You can see the source code of the whole project :

https://github.com/djdhar/GRAPATH

Using my Data Structure and Algorithms knowledge of Huffman Encoding and Decoding, and serialization and de-serialization of binary tree, I have developed the Desktop Application. It will give the user an interface to input a file and save the compressed output file along with the tree-file. With the help of these two file, user can also recover the original file back. (Algorithms are implemented from scratch)

You can see the source code of the whole project :

https://github.com/djdhar/A-Desktop-Application-for-Lossless-File-Compression

It is space-consuming to store the scanned-image based pdfs such and letters and circulars. Different OCR tools can read the files but proper format of the letters and circulars are destroyed. I have developed a Desktop Application, which first extract the page-images from the pdf. After localizing the text regions in each image, I have OCRed them and the output pdf will contain all the pages, where the relative locations of the texts and the order of the pages will remain unchanged. As the image-format pdf is converted into the text-format pdf, hard drive memory is saved.

You can see the source code of the whole project :

https://github.com/djdhar/A-Desktop-Application-for-conversion-of-PDF-of-Scanned-Images-to-PDF-of-Texts

This game is developed by implementing the basic computer graphics algorithms from scratch.

Here there will be a fixed base polygon in the Game along with a game polygon. With different shape-transformation, rotation, and translation, the gamer can win the game by matching one of the edges of the base polygon exactly with one of the edges of the game polygon.

You can see the source code of the whole project :

https://github.com/djdhar/A-Qt-Game-based-on-Basics-of-Computer-Graphics

I have developed an Web App where

1. User can register / login.

2. Registered user can anonymously ask any question.

3. Registered user can publicly answer any question.

MicroWeb FrameWork : flask

DBMS : SQLite (Interacted by Flask – SQLAlchemy python library)

Database Schema is implemented from scratch, no third party server is involved, database existing locally on the web server.

You can see the source code of the whole project :

https://github.com/djdhar/A-Qt-Game-based-on-Basics-of-Computer-Graphics

The Web App is hosted at

https://miniquora.herokuapp.com/

I have developed the Web App which allows the creator to give a Group-ID (unique) and User-Name to enter the chat-pane. Other user can enter the chat-pane by typing the same Group-ID. Multiple such group chat panes can run simultaneously without interfering one other.

With the help of python-socketio and Flask-SocketIO python library the application back-end is developed.

You can see the source code of the whole project :

https://github.com/djdhar/A-Group-Chat-Web-App