Showing papers on "Lossless JPEG published in 1970"

No-Reference JPEG Image Quality Assessment Using Haar Wavelet Decomposition

Abstract: This paper presents a novel method of no-reference image quality assessment for JPEG encoded images by means of multiresolution analysis using Haar wavelet decomposition. The proposed method takes advantage of the fact that JPEG encoded images are usually contaminated with blockiness artifacts. Blockiness artifact is modeled as a particular edge structure that transforms into a different edge structure when edge detection algorithm is applied. Subsequently after edge detection is performed, a 3-level Haar Wavelet Transform (HWT) is employed to construct an edge map, from which some features are derived. These features give meaningful information for blockiness distortions identification and quality assessment. The proposed quality metric was tested against publicly available JPEG subset of LIVE Image Database, whilst the detection algorithm was evaluated subjectively in terms of how well the automatic detection agrees with human’s perceived view. The detection algorithms as well as the proposed JPEG quality metric demonstrate satisfying performances.