الفهرس 
The ContextOnly 14 pages are availabe for public view
 |  | from | 422 |  |  |
| | | from | 422 | | |
Abstract
The interest in Three-Dimensional Video (3DV) technologies has grown considerably in both the academic and industrial worlds in the recent years. With the rapid evolution of multimedia technologies, the Three-Dimensional Video-plus-Depth (3DV+D) applications became very popular. It is expected that the 3DV+D technology will emerge as a prime application for consumer electronics in the near future. A successful 3DV system needs synergistic integration of various technologies such as 3DV acquisition, compression, transmission, and rendering. The 3DV+D is composed of variable-length stream sequences captured via diversified cameras surrounding an object. The transmission of 3DV over wireless channels has become a hot issue because of the limited resources and the existence of severe channel random and burst errors. Thus, it is an urgent task to accomplish sufficient encoding to be compatible with incoming bandwidth demands, while achieving a recommended 3DV reception performance.
In the 3DV coding prediction structure, the coded bit-streams may be dropped down due to error transmissions. Therefore, in the 3DV compression framework, the lost Macro-Blocks (MBs) might propagate into the following frames and the adjoining views. Because it is not possible to retransmit all erroneous or lost packets due to delay and bandwidth constraints on real time 3DV+D transmission, error control techniques like Error Resilience (ER) and Error Concealment (EC) are efficient techniques to ameliorate the MBs in the 3DV communication system. Thus, it is obligatory to avoid error propagation by recovering the corrupted MBs at the decoder through the utilization of appropriate pre-processing ER techniques and post-processing EC techniques. In this thesis, we focus on addressing the challenge of reliably transmitting and securing compressed 3DV+D over heavy-noisy wireless channels.
The existing error control algorithms fundamentally exploit the temporal, inter-view, and spatial matching within the 3DV+D frames and views to reconstruct the Disparity Vectors (DVs) and Motion Vectors (MVs) of the corrupted MBs. Unluckily, in the state of high severe corruptions and heavily erroneous MBs, these error control algorithms are predominantly unreliable and might give unreliable 3DV quality. Thence, in this thesis, we propose a lot of efficient and adaptive hybrid error control algorithms for reliable 3DV+D transmission over error-prone wireless channels. Moreover, the security of the transmitted 3DV+D is a critical
ABSTRACT vi
issue for protecting its information and copyright content. Digital watermarking and encryption are promising and efficient methodologies for protecting the transmitted 3DV+D data. Therefore, in this thesis, we also propose efficient multi-level security algorithms for securing the transmission of 3D Multi-view Video Coding (3D-MVC) and 3D High Efficiency Video Coding (3D-HEVC).
In the first part of this thesis, we suggest different error control algorithms for reliable transmission of 3DV over wireless channels. We propose the utilization of the Outer Block Boundary Matching Algorithm (OBBMA) to estimate the MVs and the Directional Interpolation EC Algorithm (DIECA) to estimate the DVs of the erroneous MBs. After that, the Bayesian Kalman Filter (BKF) is employed because of its efficiency to filter out the inherent errors in the previously predicted DVs and MVs to accomplish better 3DV performance. Experimental results on standard 3DV sequences demonstrate that the suggested BKF-based EC scheme is more powerful with heavy losses. It subjectively and objectively outperforms the traditional concealment techniques at severely random and bursty Packet Loss Rates (PLRs).
Furthermore, we propose hybrid ER-EC algorithms for efficient 3DV transmission over error-prone wireless channels. At the encoder, adaptive pre-processing ER mechanisms are proposed through employing the Context Adaptive Variable Length Coding (CAVLC) entropy, Slice Structured Coding (SSC) modes, and Explicit Flexible Macro-block Ordering (EFMO) mapping. They are employed to assist the suggested EC techniques at the decoder to accurately reconstruct the erroneous MBs and frames. At the decoder, an efficient post-processing EC technique with multi-proposition methods is proposed to dynamically select the convenient EC hypothesis method based on the size of the lost MBs, the faulty view, and the frame type. It conceals the received erroneous MBs of intra- and inter-encoded frames of the transmitted 3DV by exploiting the temporal, spatial, and inter-view correlations among frames and views. To further improve the decoded 3DV quality, a Weighted Overlapping Block Motion and Disparity Compensation (WOBMDC) technique is utilized to reinforce the performance of the suggested ER-EC techniques. Experimental results on various 3DV streams prove that the suggested hybrid techniques have considerably acceptable subjective and objective 3DV performance compared to the conventional error control algorithms at heavy PLRs.
To further enhance the received 3DV quality, we suggest several optimized schemes to reconstruct the lost and erroneous MBs of inter-encoded and intra-encoded frames. A hybrid method combining the Circular Scan order Interpolation Algorithm (CSOIA) and Partitioning
ABSTRACT vii
Motion Compensation Algorithm (PMCA) is proposed for intra-frames EC. For inter-corrupted MBs, a hybrid method comprising Directional Textural Motion Coherence Algorithm (DTMCA) and DIECA is suggested. Experimental outcomes on several 3DV frames demonstrate that the suggested optimized EC algorithms have more robustness to both heavy random and slice losses. They objectively and subjectively work better than the state-of-the-art EC techniques.
In the second part of this thesis, we suggest different efficient hybrid error control algorithms for reliable transmission of 3DV+D over wireless channels. We propose a new Encoder-Independent Decoder-Dependent Depth-Assisted EC (EIDD-DAEC) algorithm. It invests the depth correlations between the temporally, spatially and inter-view adjoining MBs to conceal the erroneous streams. At the encoder, the existing inter-view, temporal and spatial matching are exploited to efficiently compress the 3DV+D streams and to estimate the DVs and MVs. At the decoder, the gathered MVs and DVs from the received coded streams are used to calculate additional depth-assisted MVs and DVs, which are afterwards combined with the collected candidate texture color MVs and DVs groups for concealing the lost MBs of inter- and intra-encoded frames. Finally, the optimum DVs and MVs concealment candidates are selected by the DIECA and Decoder Motion Vector Estimation Algorithm (DMVEA), respectively. Experimental results on several standardized 3DV+D sequences verified the efficacy of the proposed EIDD-DAEC algorithm by achieving ameliorated efficacious objective and subjective results without generating and transporting depth maps at the encoder. The proposed work achieves high 3DV+D quality performance with an improved average Peak Signal-to-Noise Ratio (PSNR) gain compared to the state-of-the-art error concealment algorithms, which do not employ depth-assisted correlations at different Quantization Parameters (QPs) and PLRs.
In addition, to improve the transmission of color-plus-depth 3DV data, we propose the utilization of OBBMA and DIECA to recover the MVs and the DVs of the lost color-frames of the transmitted 3DV+D. For the lost 3DV depth-frames, an EIDD-DAEC algorithm is proposed. It exploits the recovered color MVs and DVs to estimate more additional concealment depth-assisted MVs and DVs. After that, the initially suggested concealment color-plus-depth DVs and MVs candidates are selected among the previously whole predicted ones using the DIECA and the DMVEA, respectively. Finally, the proposed BKF scheme is efficiently employed to filter out the inherent errors inside the selected concealment color-plus-depth MVs and DVs candidates; to achieve better 3DV quality. Extensive experimental results
ABSTRACT viii
on different standard 3DV+D sequences demonstrate that the proposed color-plus-depth schemes are more robust against heavy losses, and they achieve high 3DV quality performance with an improved average PSNR gain. They objectively and subjectively outperform the state-of-the-art error recovery techniques, especially at severe PLRs.
Moreover, for further enhancement of the color-plus-depth 3DV transmission, we propose efficient hybrid ER-EC algorithms for H.264 3DV+D transmission over error-prone channels. At the encoder, content-adaptive pre-processing ER mechanisms are implemented by applying the CAVLC, the SSC, and the EFMO. At the decoder, a post-processing EC algorithm with multi-proposition schemes is implemented to recover the lost 3DV color frames. The convenient EC hypothesis is adopted based on the lost MBs size mode, the faulty view, and the frame types. For the recovery of the lost 3DV depth frames, an EIDD-DAEC algorithm is suggested. It exploits the previously-estimated color DVs and MVs to estimate more additional depth-assisted MVs and DVs. After that, the optimum color-plus-depth DVs and MVs are accurately selected by employing the DIECA and the DMVEA. Finally, a WOBMDC scheme is utilized to reinforce the performance of the proposed hybrid ER-EC algorithms. Experimental results on standard 3DV+D sequences show that the proposed hybrid algorithms have superior objective and subjective performance indices.
Furthermore, to greatly improve the communication of the color-plus-depth 3DV content, we suggest optimized hybrid techniques to reconstruct the erroneous MBs of color-plus-depth inter-encoded and intra-encoded 3DV frames. A hybrid approach of CSOIA and PMCA is proposed for the color intra-frames loss concealment. For the corrupted color inter-frames, a joint approach of DTMCA and DIECA is suggested. To estimate more additional depth-aided DVs and MVs for recovering the erroneous depth frames, an error recovery EIDD-DAEC technique is suggested. These depth-estimated motion and disparity vectors are then added together with the estimated candidate texture DVs and MVs for reconstructing the corrupted color-plus-depth 3DV frames. Finally, the best color-plus-depth MVs and DVs are chosen by the DMVEA and DIECA. Simulation outcomes on various 3DV+D frames elucidate that the suggested hybrid color-plus-depth techniques achieve high robustness at high PLRs.
In the third part of this thesis, we suggest multi-stage error control algorithms for efficient and reliable communication of 3DV+D over wireless channels with severe losses. We propose multi-level enhanced ER-EC algorithms for intra-compressed and inter-compressed frames in 3DV+D communication through wireless networks. At the encoder, the SSC, EFMO, and CAVLC are utilized. At the decoder, for color intra-frame recovery, a hybrid approach
ABSTRACT ix
comprising CSOIA and PMCA is suggested. For the lossy color inter-frames, a joint approach of DTMCA and DIECA is suggested. For the concealment of the lost depth intra- and inter-frames, an EIDD-DAEC algorithm is suggested. Moreover, the WOBMDC algorithm is exploited to choose the candidate concealment MVs and DVs. Furthermore, an improved recursive BKF algorithm is utilized as a refinement stage to smooth the remnant errors in the previously selected candidate MVs and DVs for achieving better 3DV quality. Simulation results on several 3DV+D sequences show that the proposed multi-stage error control algorithms achieve adequate objective and subjective 3DV+D quality performance at severe PLRs compared to the state-of-the-art algorithms.
In the fourth part of this thesis, we propose a multi-level security framework for transmitting 3DV+D streams. We propose two robust hybrid watermarking techniques for securing the 3D color-and-depth MVC and HEVC content. The first watermarking technique is the homomorphic transform based Singular Value Decomposition (SVD) in the Discrete Wavelet Transform (DWT) domain. The second watermarking technique is the three-level Discrete Stationary Wavelet Transform (DSWT) in Discrete Cosine Transform (DCT) domain. The objective of the two proposed hybrid watermarking techniques is to increase the immunity of the watermarked 3DV streams to attacks. Also, we propose a wavelet-based fusion technique to combine two depth watermark frames into one fused depth watermark frame. Then, the resultant fused depth watermark is encrypted using chaotic Baker map to increase the level of security. After that, the resultant chaotic encrypted fused depth watermark is embedded in the 3DV color frames using the proposed hybrid watermarking techniques to produce the watermarked 3DV+D streams. In addition to achieving multi-level security in the transmitted 3DV+D streams, the proposed hybrid techniques reduce the required bit rate for transmitting the color-plus-depth 3DV data over limited-bandwidth networks. The performance of the proposed hybrid multi-level security techniques is compared with those of the state-of-the-art techniques. Extensive simulation results on standard 3DV+D sequences have been conducted in the presence of attacks. The obtained results confirm that the proposed hybrid fusion-encryption-watermarking techniques achieve not only a good perceptual quality with high PSNR values and less bit rate, but also high correlation coefficient values between the original and extracted watermarks in the presence of attacks. Furthermore, the proposed hybrid multi-level security techniques improve the capacity of the embedded information and the robustness without affecting the perceptual quality of the original 3DV frames. Indeed, the extraction of
ABSTRACT x
the encrypted, fused, primary, and secondary depth watermark frames is possible in the presence of attacks.
In addition, we introduce a new framework for securing the selective sensitive data of the compressed HEVC content transmission. The suggested selective security HEVC framework employs a robust hybrid watermarking technique for securing the transmitted HEVC data, which is the homomorphic transform based SVD in DWT domain. Moreover, the proposed selective security HEVC framework exploits the low complexity overhead chaotic Logistic Map (LM) to encrypt the sign bits of the DCT coefficients and the Motion Vector Difference (MVD) in the entropy stage of the HEVC compression process. The objective of the proposed hybrid watermarking technique is to increase the immunity of the watermarked HEVC streams to attacks. The objective of the proposed chaotic LM encryption technique is to encrypt the sensitive video bits with the features of fast encoding, low complexity overhead, format compliance, and keeping the HEVC constant bit rate. Furthermore, this research proposal presents a comparative study between the proposed selective security HEVC framework and the Glenn HEVC selective Encryption (SE) technique that uses the Advanced Encryption Standard (AES). Extensive simulation results on standard HEVC sequences have been conducted in the existence of attacks. The obtained results confirm that the proposed robust watermarking technique achieves not only a good perceptual quality with high PSNR values, but also high correlation coefficient values between the original and extracted watermarks in the presence of attacks. Furthermore, the experimental comparative results demonstrate the main feature of the proposed chaotic LM SE technique, which turned out to save the time of the video encoding while maintaining the near visual distortion of the encrypted video stream with the state-of-the-art Glenn HEVC SE technique. This feature is due to the low complexity of the chaotic LM SE technique employed in the proposed selective security HEVC framework instead of using the AES in the Glenn HEVC SE technique. A course of security investigation experiments is carried out upon the proposed security framework including the main security performance metrics like encryption quality, key space, statistical, and sensitivity tests. The obtained test results ensured the superiority of the proposed selective security HEVC framework for digital HEVC streams transmission.
In the fifth part of this thesis, we suggest the application of a chaotic Baker interleaving approach with equalization and convolution coding for efficient SVD watermarked 3D-MVC and 3D-HEVC frame transmission over an Orthogonal Frequency Division Multiplexing (OFDM) wireless system. Also, we suggest the application of chaotic Baker map interleaving
ABSTRACT xi
with equalization and convolution coding for efficient DWT-based SVD watermarked 3D-MVC and 3D-HEVC frame transmission over a Multi-Carrier Code Division Multiple Accesses (MC-CDMA) wireless channel. Rayleigh fading and Additive White Gaussian Noise (AWGN) are considered in the real scenarios of watermarked 3DV transmission. To test the performance of the proposed techniques, several simulation experiments on different SVD watermarked 3DV frames and DWT+SVD watermarked 3DV frames have been executed. The experimental results show that the received watermarked 3DV frames still have high PSNRs and watermark extraction is possible in the proposed frameworks.
In addition, we propose an efficient hybrid framework to study the effect of the application of a chaotic Baker interleaving technique with equalization for efficient transmission of composite 3D H.264 and H.265 compressed video frames over an OFDM wireless channel. Firstly, the 3DV content is compressed exploiting the intra- and inter-prediction correlations between frames. After that, a composite frame of luminance is generated from each four consecutive frames using DCT, which represents a second level of compression. The resultant composite frame is converted to binary data format. Then, chaotic interleaving is applied on the binary information prior to the modulation process. This chaotic interleaving is used to mitigate the OFDM induced Peak-to-Average Power Ratio (PAPR) problem and to reduce the wireless channel effects on the transmitted bit streams. It also adds a degree of encryption to the transmitted 3DV compressed frames. Several simulation experiments on different 3DV frames have been executed to evaluate the performance of the proposed hybrid framework. The experimental results show that the received 3DV frames have high average PSNR gains and a reduction of the average PAPR values with the proposed hybrid framework compared to the other traditional techniques.
Furthermore, a novel 3DV hybrid encryption framework based on the Rubik’s cube is suggested to attain simultaneous encryption of a group of 3DV frames. The suggested hybrid encryption framework begins with Cipher Feed Back (CFB) operation mode of chaotic Baker map permutation or AES or Ron’s Code (RC6) technique as a first step for encrypting the multiple 3DV frames, separately. Then, the resulting encrypted 3DV frames are further encrypted in a second stage with Rubik’s cube technique. Chaotic, RC6 or AES encrypted 3DV frames are utilized as the faces of the Rubik’s cube. from the concepts of 3D image encryption, the RC6 or AES technique introduces a degree of diffusion, whilst the chaotic Baker map adds a degree of permutation. Moreover, the Rubik’s cube technique adds more permutation to the encrypted 3DV frames, simultaneously. The encrypted 3DV frames are further transmitted
ABSTRACT xii
through a wireless channel with different versions of the OFDM system and decrypted at the receiver side. The quality evaluation of the decrypted 3DV frames at the receiver side reveals good performance. The simulation results reveal that the suggested hybrid encryption framework is efficient, and it presented strong security and robustness.
In the sixth part of this thesis, an efficient implementation of the Enhanced Predictive Zonal Search (EPZS) Motion Estimation (ME) algorithm for the 3D H.264/MVC standard is introduced. The overall inter-frame and inter-view prediction mechanism including Motion Compensation (MC) and ME has been implemented. For validation and comparative analysis purposes, the outcomes of the suggested design for the EPZS ME algorithm are compared to the Full Search (FS) ME algorithm. The suggested architecture of the EPZS ME algorithm is implemented in Very High-Speed Integrated Circuit (VHSIC) Hardware Description Language (VHDL), synthesized utilizing Xilinx Virtex-6 Field Programmable Gate Array (FPGA) and Xilinx ISE Design Suite 13.3, simulated employing ModelSim SE 6.5, and validated utilizing MATLAB SIMULINK. Experimental results prove that the suggested architecture achieves a low hardware complexity implementation and a high-speed of 3D H.264/MVC compression. This can be exploited for the utilization of the proposed work in real-time applications of the 3D H.264/MVC standard. Moreover, in this part, we present an efficient software design and FPGA implementation of the two proposed watermarking techniques introduced in the fourth part of this thesis. The proposed watermarking techniques are synthesized utilizing Spartan-3E FPGA Kit and Xilinx ISE Design Suite 11.4 and validated utilizing MATLAB SIMULINK. The experimental results prove that the FPGA could provide an excellent platform in implementing 3DV processing applications, since inherent parallelism of the architecture can be exploited, explicitly. Also, both FPGA experimental and MATLAB simulation results for the proposed watermarking techniques are appreciated and acceptable.