Sum Modified Laplacian-Based Image Fusion in DCT Domain with Super Resolution

摘要

Multi-focus image fusion in DCT domain are useful for Visual Sensor Network where the images have to be stored and transmitted in the encoded format. The drawbacks of existing DCT-based fusion methods are blurriness and blocking artifacts. In this paper, a novel multi-focus image fusion method is proposed by combining super resolution (SR) technique with the DCT. Single frame super resolution method is applied to the input images to avoid blocking artifacts. The contrast is chosen as a activity level measurement, and it is measured with SML. Based on the largest SML value, fusion is performed. The results obtained verify the efficiency of proposed scheme in terms of both subjective and quantitative analysis.