Sparse Representation-Based Multi-Focus Image Fusion Method via Local Energy in Shearlet Domain

Area of Science:

• Computer Vision
• Image Processing

Background:

• Multi-focus image fusion is crucial for computer vision applications.
• Existing methods can suffer from blurring and information loss.
• High-definition, information-rich fused images are the desired outcome.

Purpose of the Study:

• To propose a novel multi-focus image fusion method.
• To address blurring and information loss in image fusion.
• To achieve high-definition and information-rich fused images.

Main Methods:

• Decomposition of source images into low- and high-frequency sub-bands using shearlet transform.
• Fusion of low-frequency sub-bands via sparse representation.
• Fusion of high-frequency sub-bands using local energy.
• Reconstruction of the fused image using inverse shearlet transform.

Main Results:

• The proposed method was validated on the Lytro dataset (20 image pairs).
• Performance was compared against 8 state-of-the-art fusion methods using 8 metrics.
• Experimental results indicate the method's effectiveness in multi-focus image fusion.

Conclusions:

• The proposed shearlet domain fusion method effectively generates high-quality fused images.
• The combination of local energy and sparse representation yields superior fusion results.
• This approach offers a promising solution for enhancing multi-focus image fusion quality.