Improved estimation for just-noticeable visual distortion
Signal Processing
Perceptual Watermarking Using Pyramidal JND Maps
ISM '08 Proceedings of the 2008 Tenth IEEE International Symposium on Multimedia
Spatio-temporal just noticeable distortion profile for grey scale image/video in DCT domain
IEEE Transactions on Circuits and Systems for Video Technology
Robust Watermarking in DoG Scale Space Using a Multi-scale JND Model
PCM '09 Proceedings of the 10th Pacific Rim Conference on Multimedia: Advances in Multimedia Information Processing
Perceptual watermarking using a multi-scale JNC model
ACIIDS'10 Proceedings of the Second international conference on Intelligent information and database systems: Part II
A perceptually tuned subband image coder based on the measure of just-noticeable-distortion profile
IEEE Transactions on Circuits and Systems for Video Technology
Estimating Just-Noticeable Distortion for Video
IEEE Transactions on Circuits and Systems for Video Technology
Hi-index | 0.00 |
In this paper, a new Pyramidal Just-Noticeable-Distortion (PJND) model is proposed for video. This model incorporates the most relevant HVS properties such as: the spatio-temporal contrast sensitivity function, the influence of eye movements, the contrast masking effect and the saliency masking effect. The video sequence is first analyzed into multi-scales representation using the Laplacian pyramid decomposition (extended for 3D case). For each level, a sub JND threshold is then derived providing a global pyramidal JND map. To this end, we consider, for the first time, the influence of visual attention when designing JND model. For free viewing condition, human visual cortex is driven by a bottom-up mechanism so that it is attended only by selective salient regions. By this way, salient regions tend to mask non-salient regions. JND threshold is hence modulated by two masking mechanisms: contrast masking and "saliency masking". Recent JND models do not take into account this phenomenon and therefore do not completely exploit human visual system (HVS) limitation. Intensive experiments are carried out to demonstrate the proposed model's performance. Evaluation is performed in terms of distortion tolerance as well as perceptual transparency by means of PSNR and subjective tests. Compared to other state-of-art JNDs, our proposed PJND better exploits HVS properties by the fact that it can tolerate much more distortion while maintaining a good level of perceptual quality.