A New Visual Comfort-Based Stereoscopic Image Retargeting Method

In this paper, we present a novel stereoscopic image retargeting method of reducing both visual discomfort and the geometric distortion of objects in an image. Given a specific target display size, our method adaptively deforms the input stereoscopic image using color and disparity contrast based on a saliency analysis method. The deformation is defined and performed by solving an energy minimization problem. We propose an energy function combining geometric distortion and alignment consistency. We also consider excess disparity and misalignment to reduce visual discomfort of the resized stereoscopic image. To validate our stereoscopic image retargeting method for addressing visual comfort, we perform a user study. Experimental results show that our method is efficient and offers more user-friendly retargeting results than previously proposed methods.