Compressed Image Quality Distortion Problem due to Repeated Rotation in Spatial Domain and Its Solution by Changing Image Information in Frequency Domain
{"title":"Compressed Image Quality Distortion Problem due to Repeated Rotation in Spatial Domain and Its Solution by Changing Image Information in Frequency Domain","authors":"Kang Yi, Jeong-Hyun Joo, Kyungmi Kim","doi":"10.1109/DRBC.2015.12","DOIUrl":null,"url":null,"abstract":"Repeated rotation of compressed images may result in image quality distortion. According to our preliminary experimental results further quality distortion of JPEG image occurs as the more number of rotation. We observe that the quality degradation becomes so serious when the number of rotation exceeds a threshold. In this paper, we propose a distortion-less image rotation method for JPEG images. The key idea of our proposed method is based on the finding that a simple modification of information in frequency domain of images results in geometric transformation in spatial domain of images. We demonstrated the performance of our idea by implementing JPEG rotator and measuring the rotated image quality. The experimental results show that the proposed method achieves a perfect lossless JPEG image rotation.","PeriodicalId":426478,"journal":{"name":"2015 8th International Conference on Disaster Recovery and Business Continuity (DRBC)","volume":"141 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2015-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2015 8th International Conference on Disaster Recovery and Business Continuity (DRBC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/DRBC.2015.12","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
Repeated rotation of compressed images may result in image quality distortion. According to our preliminary experimental results further quality distortion of JPEG image occurs as the more number of rotation. We observe that the quality degradation becomes so serious when the number of rotation exceeds a threshold. In this paper, we propose a distortion-less image rotation method for JPEG images. The key idea of our proposed method is based on the finding that a simple modification of information in frequency domain of images results in geometric transformation in spatial domain of images. We demonstrated the performance of our idea by implementing JPEG rotator and measuring the rotated image quality. The experimental results show that the proposed method achieves a perfect lossless JPEG image rotation.