Research On Video Watermarking System On The Android Platform

With the rapid development of information technology and mobile internet,the transmission of digital information is becoming more and more convenient.The development of smart phones and tablets is much more quickly and the market share of Android system is also increasing.Generally,mobile devices are applied to record videos,and these videos are uploaded through mobile internet.However,the contents of videos are easy to be copied and distributed,so they are easy stolen and tampered illegally by pirates.It infringe the interests and copyright of the video owner seriously.In this paper,we combine the development of Android system multimedia and digital watermarking technology which can protect the copyright of videos,and propose the video watermarking system on the Android platform.The proposed video system framework is based on the OpenMax IL.The whole process of video system includes video capturing,recording,encoding,packaging,storage,decapsulation and decoding.Furthermore,the codec is the key to the whole video system.OpenMax IL components can use GPU to encode,so coding efficiency is higher.We first analyze the traditional video watermarking algorithms,but traditional transform domain algorithms are too complex to satisfy the real-time requirement.Traditional spatial domain algorithms can't bear compression,so its robustness is poor.To improve the robustness of the watermark,we propose four kinds of video watermarking algorithm based on spacial domain under three different video bit rate successively.The top two algorithms are based on correlation detection,and the latter two algorithms are based on the energy contrast of the specific location.We compare the watermarking extraction accuracy and image quality of the four algorithms under different bit rate,different embedding capacity.Experimental results show that with the increase of bits' number per frame,the robustness of the watermark is worse.The robustness of the last algorithm is better than the previous three algorithms and the traditional video watermarking algorithms based on spatial domain,and it can satisfy the real-time and invisibility requirements of the watermark.