Removing JPEG blocking artifacts using machine learning
Authors:
- 1 April 2014
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 77-80
Abstract
JPEG is a commonly used image compression method. While it normally yields very good compression ratios, it also introduces blocking artifacts and quantization noise. In this paper, we present a method to remove noise and blocking effects from JPEG-compressed images. We use machine learning techniques to predict DCT coefficients and pixel values in a compressed image. Results show a decrease in mean square error between our predicted images and the original uncompressed images when compared to the compressed images, as well as a clear reduction of blocking artifacts.Keywords
This publication has 5 references indexed in Scilit:
- Denoising JPEG images by re-application of JPEG2002
- Vision Models and Applications to Image and Video Processing2001
- Removal of compression artifacts using projections onto convex sets and line process modelingIEEE Transactions on Image Processing, 1997
- A deblocking algorithm for JPEG compressed images using overcomplete wavelet representationsIEEE Transactions on Circuits and Systems for Video Technology, 1997
- The JPEG still picture compression standardCommunications of the ACM, 1991