Prosody-based automatic segmentation of speech into sentences and topics
Speech Communication - Special issue on accessing information in spoken audio
Detecting Faces in Images: A Survey
IEEE Transactions on Pattern Analysis and Machine Intelligence
Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns
IEEE Transactions on Pattern Analysis and Machine Intelligence
A Multi-Modal Approach to Story Segmentation for News Video
World Wide Web
Integrating prosodic and lexical cues for automatic topic segmentation
Computational Linguistics
Personalized Digital Television: Targeting Programs to Individual Viewers (Human-Computer Interaction Series, 6)
An Unsupervised Algorithm for Anchor Shot Detection
ICPR '06 Proceedings of the 18th International Conference on Pattern Recognition - Volume 02
A Two-level Method for Unsupervised Speaker-based Audio Segmentation
ICPR '06 Proceedings of the 18th International Conference on Pattern Recognition - Volume 04
A Multimodal Scheme for Program Segmentation and Representation in Broadcast Video Streams
IEEE Transactions on Multimedia
IEEE Transactions on Circuits and Systems for Video Technology
Laplacian Eigenmaps for Automatic Story Segmentation of Broadcast News
IEEE Transactions on Audio, Speech, and Language Processing
| Hi-index | 0.00 |
Broadcast news video has been playing a more and more important role in our daily life. However, how to effectively organize the broadcast news video into story semantic units is still a challenge issue. In this paper, we propose a novel style learning based story boundary detection method (SL-SBD) to explore the boundary and structural style of each program and segment the broadcast news video into story units. Compared with traditional methods, SL-SBD calculates the appearing-candidate range of news story boundary based on topic caption tracking techniques for more reliable boundary detection. Parallel to this, SL-SBD makes use of a wealth of boundary description features to explore the boundary characteristics of each program, and proposes a two-level style learning strategy including a detector and a refiner, to enhance the learning process with strong combination of boundary style and structural style collectively. We evaluate our method on Chinese News Vision dataset, and the encouraging experimental results demonstrate the effectiveness of SL-SBD over traditional story boundary detection methods.