A Haar wavelet-based perceptual similarity index for image quality assessment
作者：
Highlights：
•
• A computationally inexpensive perceptual image similarity measure is proposed.
• The HaarPSI yields state of the art correlations with human opinion scores (MOS).
• Basic assumptions about the HVS suffice to obtain a powerful similarity measure.
• The HaarPSI can be used to evaluate image and video processing methods.
• Replacing the Haar wavelet with other filters seems not to improve the performance.
摘要
•A computationally inexpensive perceptual image similarity measure is proposed.•The HaarPSI yields state of the art correlations with human opinion scores (MOS).•Basic assumptions about the HVS suffice to obtain a powerful similarity measure.•The HaarPSI can be used to evaluate image and video processing methods.•Replacing the Haar wavelet with other filters seems not to improve the performance.
论文关键词：Image quality,Perceptual similarity,Haar wavelets,Human visual system
论文评审过程：Received 8 May 2017, Revised 6 November 2017, Accepted 6 November 2017, Available online 13 November 2017, Version of Record 24 November 2017.
论文官网地址：https://doi.org/10.1016/j.image.2017.11.001
