A Novel DWT-based Method for Image Sharpness Measure using Edge-related Frequency Components
Journal
Computer Science Research Notes
Part Of
Computer Science Research Notes
Start Page
371
End Page
376
ISSN
2464-4625
Date Issued
2024
Author(s)
Yuan-Kang Lee
Abstract
In this paper, a novel no-reference image quality metric of sharpness is proposed. Our image quality metric is evaluated on two key attributes discerned during the assessment of image sharpness by the human visual system (HVS): 1. Image sharpness is principally contingent upon the salience of edges within the image. 2. With an increase in the decomposition level of the Discrete Wavelet Transform (DWT), the high-frequency coefficients correspond to higher spatial frequency information in an image. Experimental results show that in comparison to other state-of-the-art metrics, our method not only accurately assesses image sharpness in both defocus and motion blur scenarios but also showcases superior precision and broader applicability.
Event(s)
32nd International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision, WSCG 2024
Publisher
University of West Bohemia, Czech Republic
Type
conference paper