Steady-state analysis of computational load in correlation-based image tracking
Sungjun Hong · IEE Proceedings - Vision Image and Signal Processing · 2002
Computational load in correlation-based image tracking is analysed in terms of the image frame rate. The computational load is defined as the number of arithmetic and data management operations per unit time required for correlation-based image tracking. The optimal frame rate is presented for minimising the computational load. The optimal rate is derived based on steady-state analysis under the assumption that the motion of target image is driven by the continuous white acceleration process in the image co-ordinates. The analysis shows that the optimal frame rate is a simple function of the ratio of the spectral density of the white acceleration process to the variance of observation noise.