A Spatio-temporal CRF Model for Multi-Focus Image Fusion

• Main Authors: Te-Hsiu Liao, 廖得琇
• Other Authors: Kai-Lung Hua
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/54662200708570469232

碩士 === 國立臺灣科技大學 === 資訊工程系 === 103 === The purpose of multi-focus image fusion is to integrate multiple images of different focusing goal at the same scene into a composite focusing sharp image that is more informative. Many existing methods provide good solutions for strict static scenes that do not contain any moving objects. However, in practice, these assumptions are not always realistic, and the existing methods result in ghosting artifact by the presence of moving objects. To tackle this issue, in this paper, a novel multi-focus image fusion algorithm for dynamic scenes is proposed. The proposed algorithm first detects the potential moving objects based on motion information, and then formulates the multi-focus image fusion problem into a spatial-temporal conditional random field model that considers motion, focus, and coherence factors. Finally, maximum a posteriori (MAP) and fuzzy set theory are employed to compute the weight of each pixel of each input image for the fusing the input images into an all-in-focus image. Experimental results show that the proposed method outperforms the state-of-the-art methods for both static and dynamic scenes.