Low Dimensional Video Descriptors for Fast Video Content Description and Searching
|
Supervisor: Dr. Lochandaka Ranathunga, Assoc. Prof. Nor Aniza Abdullah Department: Information Technology Faculty: Information Technology Video content description in an abstract form is mandatory for fast video searching. However, the content description is now beyond the capability of human as some popular video streamers have reported more than 180 hours of total video duration uploaded per minute. Existing computer-based content description techniques are computationally expensive. |
|
|
Human visual system can exchange the spatial resolution with a smaller number of colors to create an illusion of having many more colors. Inspired from this natural phenomenon, a set of very low dimensional video descriptors have been invented for fully automated and fast video content description. These descriptors can be used in not only for video archiving and streaming services but also in many other real-time computer vision applications such as autonomous vehicles and robots.  Abstract
An Initiative by MoraMinds Follow us on Twitter @uomresearcher
|
||
