An innovative of pyramid-based fusion for generating the HDR images in common display devices

In this paper, a method for generating the HDR images in the common devices is introduced. It can be done without computing the step of tone mapping compression. By using the dodge and burn technique with the pyramid-based technique, the resulted image can achieve both in global and local information. Firstly, a set of different exposed images in each pixel is measured for representing the well-exposed value. Then, the high intensity of the input images with dramatically changing values during the sequence of image will be reverted the occurrence, white becomes dark. Then, the pyramid-based blending technique is applied to fuse and reconstruct the set of input images. During the fusion step, the selection between inputA and inputB are used to improve the local information and avoid the visible seam. Finally, the HDR image is generated to show in the common devices. The superior of the algorithm is it does not require any information of camera response or the exposure values. Our experiments show the efficient results comparing to the existing methods.

1 Introduction

In the past decades, high dynamic range (HDR) technique has been one of significant techniques for expanding the range of the luminance levels. By following the acquisition pipeline of HDR image, a set of different exposed images, from the darkest to the brightest, is used to combine and capture the detail. It also required the function of camera response and the information of exposure values to linear the intensity of each exposed image. Anyway, most of the obtained results still have the wider dynamic range than the common devices. In the other word, it cannot directly display the HDR images in the common devices [1]. To overcome the problem, various tone mapping techniques have been implemented for compressing the dynamic range to match the ranges of display devices.