Multi-Exposure Image Fusion Algorithm Evaluation: A Perceptual Study

This paper investigates the perceptual evaluation of multi-exposure image fusion algorithms. Three datasets were utilized for this study:

1. 'LDR-HDR Dataset': This dataset contains 20 pairs of LDR (low dynamic range) and HDR (high dynamic range) images. It was used to assess the algorithms' ability to fuse images with varying dynamic ranges.

2. 'Multi-Exposure Dataset': This dataset comprises 20 multi-exposure image sequences, each containing 3 images captured at different exposure times. It allowed for the evaluation of algorithms' performance in fusing images from a single scene with varying exposures.

3. 'Synthetic Dataset': This dataset consists of 20 synthetic images created to evaluate the algorithms' performance across different scenarios and image characteristics.