Code for paper “Text-MFF: Degradation multi-focus image fusion using multi expert text constraints”.
训练、测试框架由 MFFT(EAAI, 2024) 构建而来。
网络结构由 ArtFlow(CVPR, 2021) 和 Text-IF(CVPR, 2024) 启发而来。
感谢上述作者所作出的杰出工作。
The training and testing framework is built by MFFT(EAAI, 2024).
The network structure is inspired by ArtFlow(CVPR, 2021) and Text-IF(CVPR, 2024).
Thank you to all the authors mentioned above for their outstanding work.
| Method | Code | Paper | Status |
|---|---|---|---|
| MSI-DTrans (2024) | ✅ Published | ||
| FusionGCN (2025) | ✅ Published | ||
| Frame-MFF (N/A) | 🙅 Unrevealed | ||
| Text-MFF (2026) | ✅ Published |
- The generation of statements is limited by a fixed vocabulary.
- Only cosine similarity may mislead the network into producing incorrect statements.
- Unable to effectively address strong and variable degradation interference.
- 仅提供关键代码和权重。
- 完整代码构建可参考FusionGCN项目。
- 仅需简单替换即可完成。
- Only provide key codes and weights.
- The complete code construction can refer to the FusionGCN project.
- Simply replace it to complete.
- ESWA-D-25-15335: STJ(6.17)→WE(6.18)→UR(7.5)→DIP(9.15)→Revise(9.16)
- R1: STJ(9.18)→WE(9.18)→UR(9.24)→DIP(10.6)→Revise(10.8)
- R2: STJ(10.8)→WE(10.8)→UR(10.16)→DIP(12.24)→Revise(12.26)
- R3: STJ(1.7)→WE(1.7)→UR(1.11)→DIP(1.14)→Revise(1.19)
- R4: STJ(1.19)→WE(1.19)→UR(1.22)→DIP(1.26)→Acepted(1.26)
@article{OUYANG2026131369,
title = {Text-MFF: Degradation multi-focus image fusion using multi expert text constraints},
journal = {Expert Systems with Applications},
volume = {311},
pages = {131369},
year = {2026},
issn = {0957-4174},
doi = {https://doi.org/10.1016/j.eswa.2026.131369}
}
Ouyang Y, Zhai H, Jiang J, et al. Text-MFF: Degradation multi-focus image fusion using multi expert text constraints[J]. Expert Systems with Applications, 2026: 131369.
E-mail addresses: 2023210516060@stu.cqnu.edu.cn (Y. Ouyang)