A fast motion sensing algorithm for vision based mobile phones user interface design

A fast motion sensing algorithm for vision based mobile phones user interface design is proposed in this paper. In this paper, video frames captured using camera incorporated on mobile phone will be used to determine the phone's motion, instead of using the accelerometer motion sensor. In our proposed motion sensing algorithm, gray-level grouping and comparison (GLGC) is firstly used to reduce the number of possible feature blocks, followed by feature block selection using cross first order image derivative (CFOID). Our motion sensing algorithm is then followed by using diamond search (DS) incorporating integral-successive elimination algorithm (ISEA) for block matching. Lastly, simple dominant global motion estimation (SDGME) is used to estimate a global motion. Compared with the existing motion sensing techniques, our proposed algorithm is able to reduce the number of computation significantly without sacrificing the sensing accuracy.