基于离散余弦变换的活性污泥显微图像融合 
下载： 879次
Activated Sludge Microscopic Image Fusion Based on Discrete Cosine Transform
沈阳化工大学信息工程学院, 辽宁 沈阳 110142
图 & 表
图 1. 基于DCT图像融合算法的框架
Fig. 1. Framework based on DCT image fusion algorithm
下载图片 查看原文
图 2. 第二组实验的融合图像。(a)原图1;(b)原图2;(c) DWT;(d) DTCWT;(e) CVT;(f) NSCT;(g) DCT+Variance+CV;(h)所提算法
Fig. 2. Fusion images of 2nd group of experiment. (a) Original image 1; (b) original image 2; (c) DWT; (d) DTCWT; (e) CVT; (f) NSCT; (g) DCT+Variance+CV; (h) proposed algorithm
下载图片 查看原文
图 3. 第三组实验的融合图像。(a)原图1; (b)原图2; (c) DWT; (d) DTCWT; (e) CVT; (f) NSCT; (g) DCT+Variance+CV; (h) 所提算法
Fig. 3. Fusion images of 3rd group of experiments. (a) Original image 1; (b) original image 2; (c) DWT; (d) DTCWT; (e) CVT; (f) NSCT; (g) DCT+Variance+CV; (h) proposed algorithm
下载图片 查看原文
图 4. 第四组实验的融合图像。(a)原图1; (b)原图2; (c) DWT; (d) DTCWT; (e) CVT; (f) NSCT; (g) DCT+Variance+CV; (h)所提算法
Fig. 4. Fusion images of 4th group of experiment. (a) Original image 1; (b) original image 2; (c) DWT; (d) DTCWT; (e) CVT; (f) NSCT; (g) DCT+Variance+CV; (h) proposed algorithm
下载图片 查看原文
表 1所提算法在不同的分块方式下的融合效果
Table1. Fusion effect of proposed algorithm under different block methods
| Method | MI | FMI | Q | MG | SF | Running time/s |
|---|
| 4×4 | 4.4587 | 0.5901 | 0.7162 | 7.2252 | 21.8384 | 9.4830 | | 5×5 | 4.4200 | 0.5869 | 0.6934 | 7.0108 | 21.0300 | 4.9600 | | 6×6 | 4.4414 | 0.5889 | 0.7064 | 7.1659 | 21.7830 | 4.5810 | | 7×7 | 4.4364 | 0.5881 | 0.6859 | 6.9053 | 20.9382 | 3.0770 | | 8×8 | 4.5077 | 0.5940 | 0.7018 | 7.0104 | 21.6701 | 3.0000 | | 12×12 | 4.5766 | 0.5951 | 0.6891 | 6.8962 | 21.5137 | 1.4340 |
|
查看原文
表 2第二组实验的融合效果
Table2. Fusion effect of 2nd group of experiment
| Algorithm | MI | FMI | Q | MG | SF | Running time/s |
|---|
| DWT | 1.7489 | 0.4440 | 0.6170 | 3.3297 | 8.9463 | 9.4520 | | DTCWT | 1.8356 | 0.4636 | 0.6352 | 3.0485 | 8.3517 | 9.6750 | | CVT | 1.7495 | 0.4452 | 0.6096 | 3.0424 | 8.3101 | 11.8140 | | NSCT | 1.8934 | 0.4863 | 0.6468 | 3.1316 | 8.3853 | 23.0330 | | DCT+Variance+CV | 3.9590 | 0.5901 | 0.7114 | 3.1438 | 8.5346 | 3.0470 | | Proposed algorithm | 4.5077 | 0.5940 | 0.7018 | 7.0104 | 21.6701 | 3.0000 |
|
查看原文
表 3第三组实验的融合效果
Table3. Fusion effect of 3rd group of experiment
| Algorithm | MI | FMI | Q | MG | SF | Running time/s |
|---|
| DWT | 1.3561 | 0.4324 | 0.5427 | 3.2179 | 7.4195 | 16.4360 | | DTCWT | 1.4271 | 0.4730 | 0.5645 | 2.8719 | 6.5640 | 18.1820 | | CVT | 1.3889 | 0.4576 | 0.5354 | 2.8895 | 6.5314 | 32.3250 | | NSCT | 1.4686 | 0.4910 | 0.5831 | 2.9925 | 6.6675 | 116.3610 | | DCT+Variance+CV | 3.6897 | 0.5952 | 0.6683 | 2.9919 | 6.8273 | 10.5250 | | Proposed algorithm | 3.7159 | 0.5902 | 0.6068 | 6.1086 | 15.1858 | 10.4190 |
|
查看原文
表 4第四组实验的融合效果
Table4. Fusion effect of 4th group of experiment
| Algorithm | MI | FMI | Q | MG | SF | Running time/s |
|---|
| DWT | 1.2648 | 0.4476 | 0.5617 | 3.7288 | 8.7925 | 12.4590 | | DTCWT | 1.3298 | 0.4805 | 0.5828 | 3.3432 | 7.7297 | 14.9850 | | CVT | 1.2965 | 0.4643 | 0.5529 | 3.3485 | 7.6582 | 31.7530 | | NSCT | 1.3647 | 0.4953 | 0.5976 | 3.4657 | 7.8333 | 99.2840 | | DCT+Variance+CV | 3.7660 | 0.5990 | 0.6811 | 3.4818 | 8.1006 | 10.4820 | | Proposed algorithm | 3.8221 | 0.5912 | 0.6308 | 7.0475 | 17.8417 | 10.4140 |
|
查看原文
赵立杰, 左越, 黄明忠. 基于离散余弦变换的活性污泥显微图像融合[J]. 激光与光电子学进展, 2020, 57(24): 241103. Lijie Zhao, Yue Zuo, Mingzhong Huang. Activated Sludge Microscopic Image Fusion Based on Discrete Cosine Transform[J]. Laser & Optoelectronics Progress, 2020, 57(24): 241103.
PDF全文
