During my thesis I gathered a lot of useful knowledge about mobile software development and image recognition problems. First, I got to know the Android development environment and studied different image recognition technologies for recognizing the Bolyai team competition's answer sheet. Through these I got acquainted with the development libraries of OpenCV, Google Vision API, Vuforia and Tesseract. I had to find a solution for the communication with the MySQL database, therefore I had to learn about Java Database Connectivity (JDBC) and PHP Web Server Implementations.
As the next step, I started to develop the application, including the answer sheet recognition as the main function. To the increase the efficiency and accuracy, I made some modifications on the answer sheet with the consent of the organizers. Once detection has reached the required accuracy, the other features of the application got implemented, like manual input, query results, password change and calendar function. The last step in the development process was to communicate with the server. For this I had to write separate classes from the Android side and separate PHP scripts from the server side.
After the program was completed, both the recognition and the application had to be tested. By creating a separate database and answer sheets it was first done in laboratory conditions. After the application performed well, I had an opportunity to test the program during an international final. The program is made for active usage, and hopefully it will be at good service from the next season onwards.