 OpenVL is our framework for providing computer vision methods to those not expert in the field. The framework is an abstraction based at the same level OpenGL is for computer graphics. |
||
 UCF is our framework for cross-platform uniform access to imaging devices and an abstraction over camera access and image formats. |
||
 Hive is our framework for transport and distributed processing of computer vision tasks using a peer-to-peer network with centralised control. |
||
 Part of our work requires the creation of metadata for effective video browsing; we are researching computer vision methods for detection and tracking of objects within video. |
||
 Navigating multi-view video requires research into methods for selecting moving objects, switching views and providing hyperlinks between videos. |
||
 The metadata required for navigating multi-view video can be partially created by computer vision methods, however this is not accurate enough for end-user application. To enhance the accuracy, we use an authoring tool for creation of in-video tags, adding information, comments and links. |
||
|
The advance of multimedia and HCI mechanisms for multi-view HD video requires accelerated GUI elements for animation and interaction. |
||
 Many visual effects in film and television use free-viewpoint video for innovative views of dynamic scenes. This is my research into efficient and high quality view optimisation from multi-view video capture of a dynamic scene. |
||
