All-in-focus Photo image Creation by Wavelet Transform

Abstract

In this paper, we describe a deblurring method using Image-Fusion of multiple images taken with different object-distances. Our method uses the wavelet transform for the fusion, but we create new wavelet coefficients by mixing each pixel value by using "reference image usage rate (IUR)". And we present simple and effective algorithms "Fusion rule for the lowest frequency subband" "Noise reduction algorithm" "Blur reduction algorithm" that use this IUR. I. INTRODUCTION When we take a photo, we adjust the object-distance (focus position) manually or automatically to take the clear image of the photographic subject. In other words, we put the subject into the zone of focus that is called depth-of-field. However, in the case of the short object distance, the zone becomes too narrow to cover the whole region of the subject(s) and the image of such regions become unclear. But, clear regions are also contained in a photo image, so we can correct the subject image by synthesizing (fusing) images that are taken with different object distance (See Fig.1).