Media Landscape

In this project the goal was to continuously provide images and video content based on context information such as if anyone is watching the screen or how many people are paying attention, with many added video and image processing functionalities.

For this, the project included the implementation of a media gallery, motion and object detection algorithms, metadata extraction from videos and images, as well as of an intuitive GUI for the application.

It includes the detection of faces (Haar cascades), objects (ORB detectors and FLANN-based matching), edges, motion, rhythm and cuts. Different media items are displayed based on characteristics of the viewers (number of people watching, motion, etc.).

This project was made for the course of Multimedia Computing at FCT-UNL, for which a grade of 20 / 20 was attained.

Pedro Agostinho collaborated with me on the project.