Pixel-level Multi-focus Image Fusion Based On Complex Wavelet Transform

Multi-focus image fusion aims to produce a new sharp image by some means from a set of images with the same scene but different focuses. The fused image is closer to the vision characters of human and machine, and is benefit to further applications. Multi-focus image fusion is an important part of image fusion, and it is widely used in military, medical and other fields. Wavelet transform can provide multi-scale and multi-resolution analysis in the time domain and frequency domain, so it can be used in Multi-focus image fusion.Firstly, we study the model and framework of multi-focus image fusion in wavelet domain. We study on two wavelet transform, DWT and DTCWT, including the transforming scheme, the structures of the filters in detail. We introduce the performance in image fusion of DWT.Secondly, we discuss the algorithms of fusing the approximate and detail subbands in wavelet domain separately. After studying many common used image fusion methods, we propose a new algorithm about fusing detail subbands based on the concept of the neighborhood sharpness. The sharpness is the weighted averaging value of all wavelet coefficient subbands energy. It measures the correlation of all the detail subbands in the same level. Experiments show that our algorithm with Q-Shift DT-CWT is effective compared with some traditional fusion algorithms.Thirdly, we discuss about the quality evaluation index of fused image. The quality evaluation is an important work after the images are fused and it aims to test the effectiveness of the image fusion algorithm and the quality of the fused image. Objective and subjective evaluation methods are used to evaluate the experimental results. Compared with the traditional methods, our approach has a batter result.