tailieunhanh - Kết hợp đặc trưng thị giác và ngữ nghĩa trong truy vấn video số dựa trên mô hình phân cấp dữ liệu

This paper presents a method for automatic structural analysis of digital videos to generate the table of content and the index table of the given tables allow storing videos by a hierarchy of video, cluster of shots, shots, key-frames of shots, cluster of regions. Then a method for retrieving videos by using the input data as visual feature and semantic concepts is represented. | ’ Tap ch´ Tin hoc v` Diˆu khiˆ n hoc, , (2007), 27—38 ı e e . . a ` . . . ´ ˆ ˘ ´ ` ˜. KET HO P DAC TRU NG THI GIAC VA NGU NGH˜ IA . . . . ´ ´ ´ ˆ ˆ INH PHAN CAP DU. LIEU ˆ ˆ ˜ ˆ ˆ ˆ TRONG TRUY VAN VIDEO SO DU A TREN MO H` . . . . ˜ ´ ˆ ˜ ´ ˆ ´. NGUYEN LAM1 , LY QUOC NGOC2 , DU O NG ANH DU C2 . 1 Tru.`.ng o ´ Dai hoc Thˆ Vinh, Nam Dinh e . . . 2 Tru.`.ng Dai hoc Khoa hoc Tu. nhiˆn - DHQG - Hˆ Ch´ Minh ` ı o o e . . . . Abstract. Nowadays, digital video documents grow in both the number and the size of storing spaces. Therefore, it requires efficient management techniques and methods that allow retrieving video documents in an efficient way. This paper presents a method for automatic structural analysis of digital videos to generate the table of content and the index table of the given tables allow storing videos by a hierarchy of video, cluster of shots, shots, key-frames of shots, cluster of regions. Then a method for retrieving videos by using the input data as visual feature and semantic concepts is represented. The video retrieval is performed by two steps in off-line and on-line modes. The off-line step consists of decomposing a video sequence into elementary segments. Then, these elementary segments are classified by a hierarchical clustering algorithm. Finally, a table of content and an index table are generated for the given video sequence. The on-line step consists of retrieving videos based on a hierachical data base using the input data as video clip, shot ,key-frames, representatives of region’s clusters, then the retrieved results are filtered by semantic concepts. The obtained results show that the proposed model is more efficient than the traditional systems which are only based on global, local visual features or keywords. ’ o o . ´ ´ ´ o u a a e a a a T´m t˘t. Hiˆn nay d˜. liˆu video sˆ tr˜. v` ph´t triˆ n v´.i sˆ ng`y c`ng t˘ng, do o a e u e . . . .c quan l´ h˜.u hiˆu dˆ phuc vu viˆc truy t` ’ ˜ ´ ` a