Genetic Algorithms for Blind Digital Image Stabilization Under Very Low SNR
Amine Naït‐Ali · IEEE Transactions on Consumer Electronics · 2007
The aim of this paper is to present an efficient technique devoted basically for digital image stabilization (DIS). As it is well known, image stabilization algorithms are implemented in video camera systems in order to remove unwanted shaking phenomena. We have considered in this study the case where the sequence is acquired in low signal-to-noise ratio (SNR). In such conditions, classical techniques don't work properly. Therefore, we propose in this work a new approach for optimal image synchronization based on genetic algorithms (GA). We show that this technique allows, at the same time, both stabilization and denoising. The proposed method has been evaluated on simulated case as well as on real sequences.