SALIENCE PRESERVING MULTIFOCUS IMAGE FUSION WITH DYNAMIC RANGE COMPRESSION

This paper proposes a novel multifocus image fusion method. Different from most state-of-the-art approaches, e.g., multiscale decomposition (MSD), region selection (RES) and learning based methods, our proposed method is based on salience preserving gradient and it can better emphasize the structure details of sources. We firstly measure the salience map of the gradient from each source, and then use the saliency to modulate their contributions in computing the global statistics. Gradients with high saliency are properly highlighted in the target gradient, and thereby salient features in the sources are well preserved. Furthermore we handle the dynamic range problem (DRC) by applying range compression on the target gradient. In this way, halo effect is effectively reduced. In addition, we show that the method can be easily extended to color domain by exploiting the relationships among each chromatic channels using importance-weight based trigonometric average (ITA). Extensive experiments on several datasets of multifocus images have demonstrated the superiority of our method, both in terms of visual effect and objective evaluation criteria.