Search or ask a question
Pricing
Login
Sign up
Home
Notebook
BETA
Profile
Citation booster
Literature Review
Copilot
Citation generator
Paraphraser
AI Detector
Chrome Extension
Talk with us
Use on ChatGPT
All figures (15)
Fig. 1. An out-of-focus fusion example using the “Disk” dataset available by the Image Fusion server [3]. We compare the TopoICA-based fusion approach and the proposed Diffusion scheme.
Table 3 Performance evaluation of the fused with Isotropic diffusion and the combined fusion - restoration approach in terms of PSNR (dB) and Q0.
Fig. 9. Overall fusion improvement using the proposed fusion approach enhanced with restoration. Experiments with the “pebbles” dataset.
Fig. 5. If there exist blurry parts in all input images, common Image Fusion algorithms cannot enhance these parts, but will simply transfer the degraded information to the fused image. However, this area of degraded information is still identifiable
Fig. 12. Overall fusion improvement using the proposed fusion approach enhanced with restoration and comparison with the TopoICA fusion scheme. Experiments with the “noisy-Porto” dataset.
Table 1 Performance evaluation of the Diffusion approach and the TopoICA-based fusion approach using Petrovic [18] and Piella’s [15] metrics.
Fig. 7. Overall fusion improvement using the proposed fusion approach enhanced with restoration. Experiments with the “leaves” dataset.
Fig. 11. Overall fusion improvement using the proposed fusion approach enhanced with restoration and comparison with the TopoICA fusion scheme. Experiments with the “Porto” dataset.
Table 2 Performance evaluation in the case of a multimodal example from the Toet database. The TopoICA-based approach is compared with the proposed fusion approach.
Fig. 4. Comparison of a multimodal fusion example using the TopoICA method and the Diffusion approach. Even though the metrics demonstrate worse performance, the diffusion approach highlights edges giving a sharper fused image.
Fig. 3. The weights w1, w2 highlight the position of high quality information in the input images.
Fig. 2. Convergence of the estimated fusion weight w1 using the proposed fusion algorithm in terms of ||∂w1/∂t||2.
Fig. 10. Overall fusion improvement using the proposed fusion approach enhanced with restoration. Experiments with the “British Airways (BA747)” dataset.
Fig. 8. Convergence of the restoration part and the final estimated h(r) for the common degraded area in the “leaves” example. The directivity of the estimated mask indicates the estimation of motion blur.
Fig. 6. Another example of degraded area identification in “fused” images.
Book Chapter
•
DOI
•
12 – Enhancement of multiple sensor images using joint image fusion and blind restoration
[...]
Nikolaos Mitianoudis
1
,
Tania Stathaki
1
•
Institutions (1)
Imperial College London
1
01 Jan 2007
-