Integration of face processing functionalities into relational database system Mimer SQL

Abstract: In a relational database it is possible to store a multitude of different types of data. It would be an advantageous if the database manager provides functionality for handling specialized data (e.g. images and sounds). Currently, an increasing number of commercial relational databases provide functionalities to support clients and manipulate images. This master’s thesis project is a prototype of larger project at Mimer Information Technology AB. And the project has the goal to implement a general framework for adding multimedia functionalities to Mimer SQL. This project in Mimer SQL focused mainly on face detection and face recognitions instead of general image processing, which is the common focus of its major competitors. This project made use of external image libraries to implement image functionalities, such as face detection and face recognition, into the Mimer relational database. A User-defined-type Image with built-in image functionality methods was designed and implemented to make face detection and face recognition available in Mimer SQL for users with the use of standard SQL language. And this project set up a solid foundation for the future multimedia extensions in Mimer SQL.