Typical approaches to multi-label classification problem require learning an independent classifier for every label from all the examples and features. This can become a computati...
Social network contents are not limited to text but also multimedia. Dailymotion, YouTube, and MySpace are examples of successful sites which allow users to share videos among the...
Janice Kwan-Wai Leung, Chun Hung Li, Ting Keung Ip
We propose a discriminative learning approach for fusing multichannel sequential data with application to detect unsafe driving patterns from multi-channel driving recording data....
Identifying short repeated video clips, such as news program logo, station logo, TV commercials, etc., from broadcasting video databases or streams is important for video content ...
In this paper, we present a scene detection framework on continuously recorded videos. Conventional temporal scene segmentation methods work for the videos composed of discrete sh...