# Image information optimization processing based on fractional order differentiation and WT algorithm

## TL;DR

This paper introduces a new image processing method using fractional order differentiation and wavelet transform to improve edge detection and image fusion, resulting in better image quality.

## Contribution

A novel image processing method combining fractional order differentiation and wavelet transform for enhanced edge detection and fusion.

## Key findings

- The proposed method achieved high precision in edge detection with 78.9% detection precision and 80% recall rate.
- The method showed minimum information entropy of 3.13 and mean square error of 152 in edge detection.
- The image fusion method outperformed others in average gradient, spatial frequency, and mutual information metrics.

## Abstract

As one of the most important ways for humans to perceive the world, images contain a wealth of visual information. Digital image processing is a technology that uses computer methods to process and enhance photographs in order to extract meaningful information and improve image quality. However, current image processing techniques have poor performance in processing complex images. To improve the quality of complex images, research proposes an image information optimization processing method based on fractional order differentiation and WT algorithm. Image edge detection and image fusion are important technologies in the field of image processing, with wide application value. Therefore, the study is based on wavelet transform algorithm and fractional order differentiation to perform edge detection and image fusion. The results revealed that when the study used the four evaluation metrics of information entropy, recall, mean square error, and precision to evaluate the effectiveness of image edge detection, the Sobel operator had the highest precision of detection recall, and the smallest information entropy and mean square error. The method achieved an 80% recall rate, a minimum information entropy of 3.13, a highest detection precision of 78.9%, and a minimum mean square error of 152. The average gradient, information entropy, spatial frequency, mutual information of the method adopted by the study for image fusion was compared with other methods in case of different groups of images. The method adopted by the study for image fusion provided the best results. The precision of the proposed method edge detection by the study was higher and the performance of image fusion was better and effective in improving the quality of the image.