Perceptual Quality Assessment for Multi-Exposure Image Fusion

Area of Science:

• Computer Vision
• Image Processing
• Perceptual Quality Assessment

Background:

• Multi-exposure image fusion (MEF) enhances image quality but lacks robust perceptual quality assessment.
• Existing objective models for general image fusion are inadequate for MEF images.

Purpose of the Study:

• To develop and validate a novel objective image quality assessment (IQA) algorithm specifically for multi-exposure fused images.
• To address the limitations of current IQA models in predicting the perceived quality of MEF images.

Main Methods:

• Construction of a dedicated MEF image database.
• Conducting a subjective user study to gather perceptual quality judgments.
• Development of a new objective IQA model based on structural similarity and patch structural consistency.

Main Results:

• High agreement among human subjects regarding MEF image quality.
• No single MEF algorithm consistently yields the best results across all images.
• The proposed objective IQA model demonstrates strong correlation with subjective quality scores.
• The new model significantly outperforms existing general image fusion IQA models.

Conclusions:

• The developed objective IQA model accurately predicts perceived quality of MEF images.
• The proposed model offers a reliable tool for evaluating MEF algorithms and tuning their parameters.
• This work advances the field of perceptual quality assessment for specialized image fusion techniques.