Combining Convolutional Attention Module and Convolutional Auto-encoder for Detail Injection Remote Sensing Image Fusion

Table 1. CC for $P_{L H}$ and $\tilde{M_{H}}$
View table
View in Article
Table 1. CC for $P_{L H}$ and $\tilde{M_{H}}$
Values Laplace Mean Morphological Gaussian
CC 0.051 2 0.104 2 0.178 1 0.333 8

Table 2. Fusion image metrics with different filters

View table

View in Article

Table 2. Fusion image metrics with different filters

Values	Laplace	Mean	Morphological	Gaussian	Ideal
CC	0.723 1	0.681 8	0.899 0	0.921 8	1
SAM	2.119 0	2.304 3	3.765 8	2.099 0	0
ERGAS	8.921 6	8.150 0	5.452 6	4.901 8	0
UIQI	0.802 1	0.792 2	0.871 5	0.893 0	1
AG	0.016 5	0.029 3	0.063 6	0.062 9	1
RASE	8.690 2	8.732 9	10.222 0	8.183 0	0

Table 3. Performance comparison of fusion results of Roma source images

View table

View in Article

Table 3. Performance comparison of fusion results of Roma source images

Methods	ERGAS	RASE	SAM	UIQI	AG	CC
IHS	7.885 7	15.851 7	1.611 0	0.754 3	0.029 5	0.754 5
BDSD	4.756 2	9.518 1	2.188 5	0.927 1	0.028 2	0.932 2
MTF-GLP-HPM	4.643 5	9.874 5	1.433 4	0.930 3	0.036 5	0.935 4
SR-D	7.368 2	27.234 2	1.788 9	0.795 7	0.019 9	0.846 6
PNN	4.654 8	9.781 9	1.516 7	0.901 1	0.031 0	0.928 1
Di-PNN	4.086 5	8.568 9	1.361 5	0.942 2	0.036 0	0.921 1
GAN	3.354 7	8.056 9	3.056 9	0.923 5	0.041 4	0.943 5
CAE	10.315 5	38.779 0	2.922 5	0.650 0	0.029 8	0.673 9
Proposed	3.242 8	8.137 2	1.316 7	0.929 1	0.043 6	0.934 1
Ideal	0	0	0	1	1	1

Table 4. Performance comparison of fusion results of United Arab Emirates source images

View table

View in Article

Table 4. Performance comparison of fusion results of United Arab Emirates source images

Methods	ERGAS	RASE	SAM	UIQI	AG	CC
IHS	5.562 4	12.301 5	0.450 1	0.842 1	0.014 4	0.841 2
BDSD	2.551 6	5.454 6	0.965 9	0.879 8	0.009 2	0.924 9
MTF-GLP-HPM	4.124 0	9.647 0	0.970 6	0.919 4	0.016 8	0.923 2
SR-D	8.411 7	15.447 9	1.251 5	0.859 4	0.011 9	0.886 0
PNN	4.096 8	9.601 9	1.134 1	0.920 0	0.016 7	0.924 3
Di-PNN	3.432 9	5.646 8	1.362 1	0.941 8	0.007 7	0.939 8
GAN	3.757 3	8.649 6	0.992 2	0.932 7	0.015 5	0.936 0
CAE	6.745 2	19.509 5	2.009 4	0.766 6	0.014 5	0.773 2
Proposed	1.078 9	4.757 8	0.264 1	0.953 9	0.007 7	0.944 0
Ideal	0	0	0	1	1	1

Tools

Get Citation

Copy Citation Text

Ming LI, Fan LIU, Jingzhi LI. Combining Convolutional Attention Module and Convolutional Auto-encoder for Detail Injection Remote Sensing Image Fusion[J]. Acta Photonica Sinica, 2022, 51(6): 0610005

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites