Multi-focus image fusion using a guided-filter-based difference image.

Area of Science:

• Computer Vision
• Image Processing

Background:

• Multi-focus image fusion aims to combine multiple partially focused images into a single, completely focused image.
• Existing methods often struggle with preserving fine details and achieving optimal fusion quality.

Purpose of the Study:

• To propose a novel multi-focus image fusion method utilizing a guided filter and salient feature extraction.
• To enhance the performance and quality of all-focused image generation.

Main Methods:

• A guided filter is employed for salient feature extraction and to obtain smoothing images highlighting sharp regions.
• A mixed focus measure, combining intensity variance and gradient energy, is used to generate an initial fusion map.
• Morphological filtering and guided filter optimization are applied to refine the fusion map for superior results.

Main Results:

• The proposed method demonstrates marked improvements in fusion performance compared to existing techniques.
• Experimental results show the method is competitive with, and often superior to, state-of-the-art fusion approaches.
• Both subjective visual quality and objective metrics confirm the enhanced fusion outcomes.

Conclusions:

• The novel guided filter-based multi-focus image fusion method effectively integrates information from multiple images.
• The presented salient feature extraction and fusion map optimization techniques lead to high-quality all-focused images.
• This approach offers a competitive and improved solution for multi-focus image fusion applications.