4_Student Final Year Projects >
Computer Science - Undergraduate Final Year Projects >
Please use this identifier to cite or link to this item:
|Title: ||Retrieval in large video database|
|Authors: ||Ng, Foon Wai|
|Department: ||Department of Computer Science|
|Issue Date: ||2003|
|Supervisor: ||Dr Ngo, C.W. First Reader: Wong, Hau San. Second Reader: Ip, Horace|
|Abstract: ||As the hard disk space now much larger than before, many people or commercial
likely to keep a huge amount of video files in the computers. However, most of them
don’t know how to retrieve the required shots within the large amount of videos.
Although they can write down some comments on each video, they cannot drop down
comments for each shot.
Many research works are done on the classification of video into predefined classes
by silence, non-silence, speech and music. Of course, the number of predefined
classes must be limited and usually small. This usually involves segmenting the
document into meaningful units, classifying each unit into a predefined scene type.
Less attention was paid to the content-based retrieval of similar video clip without the
limitation of fixed and predefined classes.
Therefore, it is the time to find out some approaches to solve this problem. The
similarities among video clips are based preliminary on their visual and audio content.
Hence, we divide the videos into two classes: color and audio. Color histogram
analysis can be used to retrieve video with similar colors; audio features analysis can
be used to retrieve video with similar audio contents. In this paper, I will present my
work in applying different approaches and methods for audio analysis. The visual
information extraction will be presented in my partner’s project report.|
|Appears in Collections:||Computer Science - Undergraduate Final Year Projects|
Items in CityU IR are protected by copyright, with all rights reserved, unless otherwise indicated.