Entity

Time filter

Source Type


Zhu J.,Hohai University | Zhu J.,Changzhou Key Laboratory of Sensor Networks and Environmental Perception | Cao N.,Hohai University | Meng Y.,Hohai University | Meng Y.,Changzhou Key Laboratory of Sensor Networks and Environmental Perception
International Journal of Distributed Sensor Networks | Year: 2013

A novel adaptive multihypothesis (MH) prediction algorithm for distributed compressive video sensing (DCVS) is proposed in this paper. In the proposed framework, consistent block-based random measurement for each video frame is adopted at the encoder independently. Meanwhile, a mode decision algorithm is applied in CS-blocks via block-based correlation measurements at the decoder. The inter-frame MH mode is selected for the current block wherein the interframe correlation coefficient value exceeds a predetermined threshold. Otherwise, the intraframe MH mode is worthwhile to be selected. Moreover, the adaptive search window and cross-diamond search algorithms on measurement domain are also incorporated to form the dictionary for MH prediction. Both the temporal and spatial correlations in video signals are exploited to enhance CS recovery to satisfy the best linear combination of hypotheses. The simulation results show that the proposed framework can provide better reconstruction quality than the framework using original MH prediction algorithm, and for sequences with slow motion and relatively simple scene composition, the proposed method shows significant performance gains at low measurement subrate. © 2013 Jinxiu Zhu et al.


Zhu J.,Hohai University | Zhu J.,Changzhou Key Laboratory of Sensor Networks and Environmental Perception | Cao N.,Hohai University | Yao Z.,Hohai University | Yao Z.,Changzhou Key Laboratory of Sensor Networks and Environmental Perception
Journal of Computational Information Systems | Year: 2013

Modeling and estimating the correlation noise between the original Wyner-Ziv frame data and its side information is the most important factors in distributed video coding system. One of the problems of the state-of-the-art correlation estimators is that their performance is not consistent satisfied with a wide range of video with different motion activities. To address this problem, this paper analyses theoretically the Laplace distribution model and demonstrates how a Laplace distribution of the residual coefficients can be derived by using a doubly stochastic model. This model also allows us to investigate that not only variance but also cross correlated energy of blocks could affect the shape of distribution. Based on this, a new correlation model able to adapt to changes in the content and GOP size is proposed by taking into account the cross correlated energy of blocks. Experimental results show that the proposed model can effectively improve the coding efficiency especially for the video sequence containing a significant amount of motion activity and for the DVC system with large GOP size. © 2013 by Binary Information Press.


Zhang Y.,Hohai University | Zhang Y.,Changzhou Key Laboratory of Sensor Networks and Environmental Perception | Zhu J.,Hohai University | Zhu J.,Changzhou Key Laboratory of Sensor Networks and Environmental Perception | And 4 more authors.
Journal of Computational Information Systems | Year: 2014

In this paper, considering the temporal and spatial correlations in video signals, the block-based Karhunen-Loève transform (KLT) basis which can get the most sparsity representation of current frame is introduced to reconstruct the frame to get better initial reconstruction, where the block-based KLT basis is generated by reference blocks which adaptively extracted from the search window that centered in the best matching block in previously reconstructed frame. In our proposed DCVS framework, the K-frame uses KLT-based intra multihypothesis (MH) prediction algorithm, while for the CS-frame, the KLT-based inter MH prediction algorithm is selected for the current block if the interframe correlation coefficient value exceeds a predetermined threshold. Otherwise, the KLT-based intra MH-prediction algorithm is worthwhile to be selected to generate a better side information (SI) for the sparse reconstruction. The experimental results show that our proposed framework can provide 1dB-4dB increase in PSNR compared to the conventional MH prediction for DCVS. © 2014 Binary Information Press.

Discover hidden collaborations