Pornographic image/video recognition plays a vital role in network information surveillance and management. In this paper, its key techniques, such as skin detection, key frame extraction, and classifier design, etc.,...Pornographic image/video recognition plays a vital role in network information surveillance and management. In this paper, its key techniques, such as skin detection, key frame extraction, and classifier design, etc., are studied in compressed domain. A skin detection method based on data-mining in compressed domain is proposed firstly and achieves the higher detection accuracy as well as higher speed. Then, a cascade scheme of pornographic image recognition based on selective decision tree ensemble is proposed in order to improve both the speed and accuracy of recognition. A pornographic video oriented key frame extraction solution in compressed domain and an approach of pornographic video recognition are discussed respectively in the end.展开更多
Rate-distortion optimization greatly improves the performance of compression coding system so that it pervades all of the source coding from an informationtheoretic standpoint and for the design of practical coding sy...Rate-distortion optimization greatly improves the performance of compression coding system so that it pervades all of the source coding from an informationtheoretic standpoint and for the design of practical coding systems. For the case of rate-distortion optimization, Lagrange multiplier method provides the efficient and nearly optimal solution. In this paper, a fast and efficient algorithm is proposed to solve the optimal slope λ＊ of the rate-distortion curve at the given bit budget. Based on Lagrange multiplier method, the presented algorithm find λ＊ using the golden-ratio search. Compared with the Bisection method that only adapts to the system with the dense operational points on the rate-distortion curve, the proposed algorithm can be adapted to the system whether the operational points are populated densely or not. Thus it can be applied to both the wavelet coding system and the video coding standards such as H. 264, where Bisection method can not work well. In particular, the algorithm has been verified on the platform of the quadtree classified and trellis coded quantized （QTCQ） wavelet image compression system and the newest video coding standard H. 264. The experimental results are provided to demonstrate the efficiency of the algorithm. The proposed algorithm can improve the performance. A gain abour 0.6 - 0.7 dB can be achieved with the same rate in H. 264. In addition, it converges as fast as Bisection method, with almost the same ctinplexity.展开更多
基金Supported by the National Natural Science Foundation of China (No.60772069)863 High-Tech Project (2008AA01A313)
文摘Pornographic image/video recognition plays a vital role in network information surveillance and management. In this paper, its key techniques, such as skin detection, key frame extraction, and classifier design, etc., are studied in compressed domain. A skin detection method based on data-mining in compressed domain is proposed firstly and achieves the higher detection accuracy as well as higher speed. Then, a cascade scheme of pornographic image recognition based on selective decision tree ensemble is proposed in order to improve both the speed and accuracy of recognition. A pornographic video oriented key frame extraction solution in compressed domain and an approach of pornographic video recognition are discussed respectively in the end.
基金Special Foundation of Outstanding Young Teacher of ShanghaiShanghai Educational Development Foundation,China (No.2007CG66)+1 种基金Shanghai Key Research Project,China ( No.071605125,No.08160510600)Innovation Program of Shanghai Municipal Education Commission,China(No.09ZZ185,No.09YZ337)
文摘Rate-distortion optimization greatly improves the performance of compression coding system so that it pervades all of the source coding from an informationtheoretic standpoint and for the design of practical coding systems. For the case of rate-distortion optimization, Lagrange multiplier method provides the efficient and nearly optimal solution. In this paper, a fast and efficient algorithm is proposed to solve the optimal slope λ＊ of the rate-distortion curve at the given bit budget. Based on Lagrange multiplier method, the presented algorithm find λ＊ using the golden-ratio search. Compared with the Bisection method that only adapts to the system with the dense operational points on the rate-distortion curve, the proposed algorithm can be adapted to the system whether the operational points are populated densely or not. Thus it can be applied to both the wavelet coding system and the video coding standards such as H. 264, where Bisection method can not work well. In particular, the algorithm has been verified on the platform of the quadtree classified and trellis coded quantized （QTCQ） wavelet image compression system and the newest video coding standard H. 264. The experimental results are provided to demonstrate the efficiency of the algorithm. The proposed algorithm can improve the performance. A gain abour 0.6 - 0.7 dB can be achieved with the same rate in H. 264. In addition, it converges as fast as Bisection method, with almost the same ctinplexity.
