Plating — New Prospects for an Old Art.pdfVIP

A Multi-exposure Fusion Method Based on Locality Properties 1 1, 1 1 1 Yuanchao Bai , Huizhu Jia , Hengjin Liu , Guoqing Xiang , Xiaodong Xie , Ming Jiang2 , and Wen Gao1 1 National Engineering Laboratory for Video Technology, Peking University, Beijing 100871, China {yuanchao.bai,hzjia,hjliu,donxie,wgao}@pku.edu.cn 2 Department of Information Sciences, School of Mathematical Sciences, Peking University, Beijing 100871, China ming-jiang@ieee.org Abstract. A new method is proposed for fusing a multi-exposure se- quence of images into a high quality image based on the locality proper- ties of the sequence. We divide the images into uniform blocks and use variance to represent the information of blocks. The richest information (RI) image is computed by piecing together blocks with largest variances. We assume that images in the sequence are high dimensional data points lying on the same neighbourhood and borrow the idea from locally linear embedding (LLE) to fuse a result image which is closest to the RI image. The result is comparable to the state-of-art tone mapping operators and other exposure fusion methods. Keywords: Exposure fusion, High dynamic range, Locality properties, LLE. 1 Introduction Digital cameras are developing quickly in producing higher-resolution images, but the limited dynamic range is still a challenge. Currently, the vast majority of images are represented in 8 bits per pixel for each colour channel, which is called low dynamic range (LDR) image, in contrast with the high dynamic range (H