An electronic digital image stabilizer based on stationary wavelet transform (SWT)

In this paper we present a new SWT based algorithm for electronic digital image stabilization application. In our proposed algorithm, we first estimate the translational motion by projecting the vertical and horizontal details of stationary wavelet decomposition in level 2. These parameters are used as initial estimation of motion's parameters. Then, we use several feature points and a gradient based algorithm to estimate the precise motion parameters. We compared our algorithm with two other algorithms, i.e. Moriomoto and Guestrin's stabilization algorithms. Experimental results indicate that our proposed algorithm has superior performance than these two.