Related Experiment Video
Updated: Sep 9, 2025
13:01
Using Light Sheet Fluorescence Microscopy to Image Zebrafish Eye Development
Published on: April 10, 2016
34.1K
MLP-MFF: Lightweight Pyramid Fusion MLP for Ultra-Efficient End-to-End Multi-Focus Image Fusion
Yuze Song1,2, Xinzhe Xie3, Buyu Guo4,5
1National Ocean Technology Center, Tianjin 300112, China.
Sensors (Basel, Switzerland)
|August 28, 2025
Summary
This study introduces MLP-MFF, a novel deep learning model for multi-focus image fusion (MFF). It efficiently synthesizes all-in-focus images, overcoming limitations of existing methods for clearer optical imaging.
Area of Science:
- Computer Vision
- Image Processing
- Deep Learning
Background:
- Limited depth of field in optical imaging leads to partially focused images.
- Multi-focus image fusion (MFF) synthesizes all-in-focus images from multiple focal planes.
- Existing deep learning MFF methods struggle with long-range dependencies or high computational costs.
Purpose of the Study:
- To propose a novel, lightweight, end-to-end MFF network.
- To overcome limitations of CNNs, Transformers, and Mambas in MFF.
- To enhance fusion performance and realize a true global receptive field.
Main Methods:
- Developed MLP-MFF based on the Pyramid Fusion Multi-Layer Perceptron (PFMLP) architecture.
- Introduced a Dual-Path Adaptive Multi-scale Feature-Fusion Module based on Hybrid Attention (DAMFFM-HA).
- Designed to handle flexible input scales and learn multi-scale feature representations.
Main Results:
- MLP-MFF achieves competitive and often superior fusion quality compared to state-of-the-art methods.
- Demonstrated effective capture of long-range dependencies and multi-scale features.
- Maintained a lightweight and efficient architecture.
Conclusions:
- MLP-MFF offers an effective solution for multi-focus image fusion.
- The proposed architecture and fusion module significantly enhance MFF performance.
- This method provides a lightweight yet powerful approach to synthesizing all-in-focus images.
