Commercial detection by mining maximal repeated sequence in audio stream

Efficient detection of commercial is an important topic for many applications such as commercial monitoring, market investigation. This paper reports an unsupervised technique of discovering commercial by mining repeated sequence in audio stream. Compared with previous work, we focus on solving practical problems by introducing three principles of commercial: repetition principle, independence principle and equivalence principle. Based on these principles, we detect the commercials by first mining maximal repeated sequences (MRS) and then post-processing the MRS pairs based on independence principle and equivalence principle for final result. In addition, a coarse-to-fine scheme is adopted in the acoustic matching stage to save computational cost. Extensive experiments both on simulated data and real broadcast data demonstrate the effectiveness of our method.

[1]  Lie Lu,et al.  Robust learning-based TV commercial detection , 2005, 2005 IEEE International Conference on Multimedia and Expo.

[2]  Chu-Song Chen,et al.  A Time Warping Based Approach for Video Copy Detection , 2006, 18th International Conference on Pattern Recognition (ICPR'06).

[3]  Ling-Yu Duan,et al.  Robust Commercial Retrieval in Video Streams , 2007, 2007 IEEE International Conference on Multimedia and Expo.

[4]  Wei Liang,et al.  A Histogram Algorithm for Fast Audio Retrieval , 2005, ISMIR.

[5]  Shumeet Baluja,et al.  Advertisement Detection and Replacement using Acoustic and Visual Repetition , 2006, 2006 IEEE Workshop on Multimedia Signal Processing.

[6]  John M. Gauch,et al.  Identification of new commercials using repeated video sequence detection , 2005, IEEE International Conference on Image Processing 2005.

[7]  Qun Liu,et al.  Fast commercial detection based on audio retrieval , 2008, 2008 IEEE International Conference on Multimedia and Expo.

[8]  Ton Kalker,et al.  A Highly Robust Audio Fingerprinting System , 2002, ISMIR.