Computer graphics relies on the human visual perception and cognition to convey computer-generated visual information. Of particular interest are the possibilities and limitations of technology and design. Computer graphics involves complex and multi-disciplinary tasks. For instance, for a long time, the film industry has used computer graphics mixed with real imagery. The complexity of generating the necessary geometry, material and lighting models for this has given rise to a range of techniques such as image-based modelling and lighting, etc. In order to be successful in computer graphics, you must master several of these techniques. You can work with e.g. animation, 3D graphics and motion capture, reproduction, manipulation, augmented and virtual reality, computer vision and visualisation of information, etc.
project examples from the computer graphics specialisation
body-ownership of wings in virtual reality
In the fall of 2015, a group of four 1st semester students decided to investigate how people can achieve agency and body-ownership of virtual wings in virtual reality. Using a head-mounted display and motion tracking of the head, upper arms, and torso, test participants were able to control their avatar in virtual reality. Large virtual mirrors in the virtual environment allowed the participants to see the virtual wings on the backs of their avatars without turning their heads. The test participants compared three different scenarios: one was without motor control over the virtual wings; one was with motor control over the wings; and one was with motor control and tactile feedback from vibrators on the backs of the test participants. The results showed that motor control was necessary for establishing agency and body-ownership of the wings and that the additional vibrotactile feedback significantly enhanced the agency and body-ownership of the wings. A few months after the exam, the students presented their results at the Virtual Reality International Conference 2016.
augmented reality for self-service guides in a museum
In the spring of 2016, a group of three 2nd semester students tried to find out how museums’ self-service guides could benefit from augmented reality (AR) technologies. Specifically, they collaborated with the architectural museum Utzon Center in Aalborg and developed two AR self-service guides for two exhibits at Utzon Center. One of the AR guides used a smartphone to display additional information about the exhibits on top of the camera view. The second AR guide used smart glasses (Epson Moverio BT-200) to display additional information directly in the field of view of the user. The user test not only revealed advantages and disadvantages of both guides, but it also started a discussion at Utzon Center about using AR technologies in future self-service guides for smartphones.
programming of games in real time
In the fall of 2015, several 3rd semester students participated for one semester in the education offered by the National Academy of Digital Interactive Entertainment (DADIU). Together with students from various Danish universities and academies, they formed teams and developed games for Android tablets. In particular, one student in the Computer Graphics specialization signed up as “Programmer” in the DADIU education and was, therefore, responsible for the graphics programming. Since the graphics performance of tablets is a lot worse than the graphics performance of desktop computers, he had to work hard on optimizations to make sure that the game is running in real time. Furthermore, he had to work with the art director and CG artists on the team (who lacked his technical knowledge) to find ways to bring their visions to life in spite of the limitations of the hardware. His team called their game “Game Changer” and published it as a free app on the Google Play store.
Master's thesis example: new rendering technique for light field displays
In 2016, two 4th semester students decided to develop a new rendering technique for light field displays, which are one of the very few display technologies for animated, computer-generated holograms, i.e., these displays provide even more depth cues and are more comfortable to look at than 3D television. Light field displays usually require several dozens of computer-generated images from slightly different viewpoints to show a single frame of an animated hologram. Unfortunately, computing that many images in real time requires a lot of graphics performance. To reduce the required performance, the students developed an algorithm that requires only four images from four different viewpoints. From the pixels of these images, the algorithm can compute the remaining images without processing any other data that would usually be necessary to render the images. The students implemented their algorithm for graphics processing units (GPUs) and evaluated not only the performance of the algorithm but also the visual quality of the images as perceived by human users. The results showed that the visual quality is sufficient but the performance requires further optimizations. Maybe more importantly, their algorithm inspired a follow-up project about how to improve the rendering of computer-generated 360 degree panoramas in head-mounted displays.