Research of Multi-focus Image Fusion Algorithm Based on Sparse Representation and Orthogonal Matching Pursuit - Advances in Image and Graphics Technologies - page 64

Table 4. Objective Evaluation Indicator Data to Multi-focus Fusion Image of Group D

Q AB/F Algorithm MI AG_A AG_B AG_F CC_A CC_B DD_A DD_B

8*8DCT

7.843 3.912 4.402 5.225 0.997 0.996 2.827 0.983 0.731

8*8KSVD

7.582 3.912 4.402 5.178 0.976 0.996 2.702 0.975 0.727

FSD

5.640 3.912 4.402 4.428 0.979 0.988 3.702 2.779 0.675

Contrast

7.059 3.912 4.402 5.472 0.981 0.993 2.919 1.566 0.691

DWT

6.454 3.912 4.402 5.616 0.981 0.991 2.904 1.710 0.666

SIDWT

6.789 3.912 4.402 5.257 0.984 0.992 2.750 1.544 0.695

SF

8.049 3.912 4.402 5.135 0.977 0.996 2.815 0.910 0.732

4 Conclusion

Through the careful observation of four groups of multi-focus fusion image by Matlab
and comparing with the corresponding data for objective evaluation index, we can
draw the following conclusions: from the view of fusion image, the results of image
fused by various algorithms differ slightly. But from the details of clarity and texture,
the fused images by the multi-focus image fusion algorithm based on sparse
representation and orthogonal matching pursuit have better clarity and texture details,
with more information from the original what?. From the objective evaluation index,
the multi-focus image fusion algorithm based on sparse representation and orthogonal
matching pursuit presents higher mutual information, larger average gradient value
AG, minimum distortion coefficients DD and higher Q ab / f values, which means that
this image fusion algorithm can maintain more original information with the smallest
distortion, reflecting the edge information and the importance of the original image.
Therefore, it makes better effects of image fusion than other multi-source image
fusion algorithms.

References

1. Pohl, C., Van Genderen, J.L.: Multisensor Image Fusion in Remote Sensing: Concepts,
Methods and Applications. International Journal of Remote Sensing 19(5), 823-854 (1998)
2. Eichmann, G.: Pyramidal Image Processing Using Morphology. Applications of Digital
Image Processing Xl. Proceedings of the SPIE 974, 30-37 (1998)
3. Goutsias, J., Heijmans, H.M.: Nonlinear Multi-resolution Signal Decompositions
Secheme-Partl: Morphological Pyramids. IEEE Trans. on Image Processing 9(11),
1862-1876 (2000)
4. Burt, P.J., Kolczynski, R.J.: Enhanced Image Capture Through Fusion. In: Fourth
International Conference on Computer Vision, pp. 173-183. IEEE Press, Berlin (1993)
5. Qiguang, M., Baoshu, W.: Multi-Sensor Image Fusion Based on Improved Laplacian
Pyramid Transform. Acta Optica Sinica 29(9), 1605-1610 (2007)