{"title":"Modification of Objective Full Reference Metrics for Image Quality Measurement and Evaluation","authors":"R. Deliyski","doi":"10.1109/MMA52675.2021.9610944","DOIUrl":null,"url":null,"abstract":"This paper presents a modification of objective PSNR metric for image quality measurement and evaluation. In the modified metric the MSE is substituted by 3 different distance measures: MAE, Euclidean distance and Canberra distance. All metrics were tested with 12 different 3-channels, 8-bits images. Every test image was distorted by 10 different values of dispersion of white Gaussian noise with zero mean. The experiments show that the dispersion of the PSNR with Euclidean distance has the lowest standard deviation and thus improved the original PSNR. It is also suggested that due to normality of the data and their deviation, the results has to be accepted with some standard uncertainty.","PeriodicalId":287017,"journal":{"name":"2021 XXXI International Scientific Symposium Metrology and Metrology Assurance (MMA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 XXXI International Scientific Symposium Metrology and Metrology Assurance (MMA)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/MMA52675.2021.9610944","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 1
Abstract
This paper presents a modification of objective PSNR metric for image quality measurement and evaluation. In the modified metric the MSE is substituted by 3 different distance measures: MAE, Euclidean distance and Canberra distance. All metrics were tested with 12 different 3-channels, 8-bits images. Every test image was distorted by 10 different values of dispersion of white Gaussian noise with zero mean. The experiments show that the dispersion of the PSNR with Euclidean distance has the lowest standard deviation and thus improved the original PSNR. It is also suggested that due to normality of the data and their deviation, the results has to be accepted with some standard uncertainty.