A semantic analysis method based on concept lattice rules streamlined by negative samples

Video semantic concept detection plays an important role in video retrieval. On basis of the analysis that the videos contain rich semantic, this paper proposes a semantic analysis method based on concept lattice rules streamlined by negative samples, by introducing concept lattice, an important mathematical tool on knowledge processing and analysis of data. And the proposed method is applied to video semantic concept detection. Finally, we verify this method by news videos. Experiments show feasibility and validity of proposed method in video concept detection.