Estimating SSIM from MSE for DCT-Based Compressed Images
2026-08-03 • Multimedia
MultimediaComputer Vision and Pattern Recognition
AI summaryⓘ
The authors show a way to estimate the quality of JPEG-compressed images using a simpler measure called SSIM, but without needing detailed error info at every small part of the image. They do this by redistributing a global error score (MSE) with some statistics from the original image, like variance or standard deviation. Their methods work well on test image sets and give better results than just using the global error alone. They suggest this approach could also help with video quality measurement by reusing these image statistics across frames.
Discrete Cosine TransformJPEG compressionStructural Similarity Index (SSIM)Peak Signal to Noise Ratio (PSNR)Mean Square Error (MSE)variancestandard deviationimage quality assessmentvideo encoding
Authors
Luc Trudeau, Maria G. Martini
Abstract
Efficient and perceptually meaningful quality assessment is a fundamental requirement for image and video processing, compression, and streaming systems. This article shows that, in the context of Discrete Cosine Transform ( DCT)-based compressed images, Structural Similarity Index ( SSIM ) can be approximated from global Peak Signal to Noise Ratio (PSNR) or Mean Square Error ( MSE) using local statistics derived only from the reference image. While prior work assumes access to local MSE, we propose two approaches to approximate local MSE by redistributing the global MSE using variance or standard-deviation-based weighting. Experiments on the Kodak and Xiph Subset1 datasets across a range of JPEG quality levels demonstrate that both approaches provide accurate and robust SSIM approximations, substantially outperforming the global MSE baseline. The proposed framework is designed to extend naturally to video, where reference-derived statistics can be amortized across multiple encodes of the same content.