Compressed video bitstream transmissions over wireless networks are addressed in this work. We first consider error control and power allocation for transmitting wireless video over CDMA networks in conjunction with multiuser detection. We map a layered video bitstream to several CDMA fading channels and inject multiple source/parity layers into each of these channels at the transmitter. We formulate a combined optimization problem and give the optimal joint rate and power allocation for each of linear minimum mean-square error (MMSE) multiuser detector in the uplink and two types of blind linear MMSE detectors, i.e., the direct-matrix-inversion (DMI) blind detector and the subspace blind detector, in the downlink. We then present a multiple-channel video transmission scheme in wireless CDMA networks over multipath fading channels. For a given budget on the available bandwidth and total transmit power, the transmitter determines the optimal power allocations and the optimal transmission rates among multiple CDMA channels, as well as the optimal product channel code rate allocation. We also make use of results on the large-system CDMA performance for various multiuser receivers in multipath fading channels. We employ a fast joint source-channel coding algorithm to obtain the optimal product channel code structure. Finally, we propose an end-to-end architecture for multi-layer progressive video delivery over space-time differentially coded orthogonal frequency division multiplexing (STDC-OFDM) systems. We propose to use progressive joint source-channel coding to generate operational transmission distortion-power-rate (TD-PR) surfaces. By extending the rate-distortion function in source coding to the TD-PR surface in joint source-channel coding, our work can use the ??equal slope?? argument to effectively solve the transmission rate allocation problem as well as the transmission power allocation problem for multi-layer video transmission. It is demonstrated through simulations that as the wireless channel conditions change, these proposed schemes can scale the video streams and transport the scaled video streams to receivers with a smooth change of perceptual quality.

While there has always been a tremendous demand for streaming video over TCP/IP networks, the nature of the application still presents some challenging issues. These applications that transmit multimedia data over best-effort networks like the Internet must cope with the changing network behavior; specifically, the source encoder rate should be controlled based on feedback from a channel estimator that probes the network periodically. First, one such Multimedia Streaming TCP-Friendly Protocol (MSTFP) is considered, which iteratively integrates forward estimation of network status with feedback control to closely track the varying network characteristics. Second, a network-adaptive embedded bit stream is generated using a r-domain rate controller. The conceptual elegance of this r-domain framework stems from the fact that the coding bit rate ) (R is approximately linear in the percentage of zeros among the quantized spatial transform coefficients ) ( r , as opposed to the more traditional, complex and highly nonlinear ) ( Q R characterization. Though the r-model has been successfully implemented on a few other video codecs, its application to the emerging video coding standard H.264 is considered. The extensive experimental results show thatrobust rate control, similar or improved Peak Signal to Noise Ratio (PSNR), and a faster implementation.

Video coding standards such as H.264 AVC (Advanced Video Coding) rely on predictive coding to achieve high compression efficiency. Predictive coding consists of predicting each frame using preceding frames. However, predictive coding incurs a cost when transmitting over unreliable networks: frames are no longer independent and the loss of data in one frame may affect future frames. In this thesis, we study the effectiveness of Flexible Macroblock Ordering (FMO) in mitigating the effect of errors on the decoded video and propose solutions to improve the error concealment on H.264 decoders. After introducing the subject matter, we present the H.264 profiles and briefly determine their intended applications. Then we describe FMO and justify its usefulness for transmission over lossy networks. More precisely, we study the cost in terms of overheads and the improvements it offers in visual quality for damaged video frames. The unavailability of FMO in most H.264 profiles leads us to design a lossless FMO removal scheme, which allows the playback of FMO-encoded video on non FMO-compliant decoders. Then, we describe the process of removing the FMO structure but also underline some limitations that prevent the application of the scheme. Finally, we assess the induced overheads and propose a model to predict these overheads when FMO Type 1 is employed. Eventually, we develop a new error concealment method to enhance video quality without relying on channel feedback. This method is shown to be superior to existing methods, including those from the JM reference software and can be applied to compensate for the limitations of the scheme proposed FMO-removal scheme. After introducing our new method, we evaluate its performance and compare it to some classical algorithms.

Efficient and effective vision systems are proposed in this work for object detection for ground&aerial robots venturing into unknown environments with minimum vision aids, i.e. a single camera. The first problem attempted is that of object search and identification in a situation similar to a disaster site. Based on image analysis, typical pixel-based characteristics of a visual marker have been established to search for, using a block based search algorithm, along with a noise and interference filter. The proposed algorithm has been successfully utilized for the International Aerial Robotics competition 2009. The second problem deals with object detection for collision avoidance in 3D environments. It has been shown that a 3D model of the scene can be generated from 2D image information from a single camera flying through a very small arc of lateral flight around the object, without the need of capturing images from all sides. The forward flight simulations show that the depth extracted from forward motion is usable for large part of the image. After analyzing various constraints associated with this and other existing approaches, Motion Estimation has been proposed. Implementation of motion estimation on videos from onboard cameras resulted in various undesirable and noisy vectors. An in depth analysis of such vectors is presented and solutions are proposed and implemented, demonstrating desirable motion estimation for collision avoidance task.

Thesis (Ph. D.)--University of Missouri-Columbia, 2000. / Typescript. Vita. Includes bibliographical references (leaves 202-221). Also available on the Internet.

Includes bibliographical references (p. 31-34).

Persistics is an advanced framework for processing wide-area aerial surveillance video. This framework handles the tasks of data collection, stitching of multi-sensor imagery, image registration and stabilization, motion tracking, and compression. As the technology for image sensor sizes improves, significant improvements in compression techniques are necessary in order to make full use of the data. Because the information of interest in such video is naturally moving, point-like targets, the applicability of foveated coding to the compression problem is an interesting question. Foveated coding, a compression technique that was designed to be perceptually optimal for the human visual system, has several components that are appropriate to the persistics compression problem. Foveation is applied in several different scenarios and methods to persistics data. As foveation can make good use of the persistics tracker data, a problem affecting tracker performance is explored as well. The multi-sensor stitching component of persistics can generate artifacts that reduce the effectiveness of the tracker. A method for characterizing, detecting, and correcting such artifacts is desirable. These three concepts are explored, and a method for detection is developed. Components of these algorithms were absorbed into a more general framework for artifact correction. / text

With the rapid growth of video services on smartphones such as video conferencing, video telephone and WebTV, implementation of video compression on mobile terminal becomes extremely important. However, the low computation capability of mobile devices becomes a bottleneck which calls for low complexity techniques for video coding. This work presents two set of algorithms for reducing the complexity of motion estimation. Binary motion estimation techniques using one-bit and two-bit transforms reduce the computational complexity of matching error criterion, however sometimes generate inaccurate motion vectors. The first set includes two neighborhood matching based algorithms which attempt to reduce computations to only a fraction of other methods. Simulation results demonstrate that full search local binary pattern (FS-LBP) algorithm reconstruct visually more accurate frames compared to full search algorithm (FSA). Its reduced complexity LBP (RC-LBP) version decreases computations significantly to only a fraction of the other methods while maintaining acceptable performance. The second set introduces edge detection approach for partial distortion elimination based on binary patterns. Spiral partial distortion elimination (SpiralPDE) has been proposed in literature which matches the pixel-to-pixel distortion in a predefined manner. Since, the contribution of all the pixels to the distortion function is different, therefore, it is important to analyze and extract these cardinal pixels. The proposed algorithms are called lossless fast full search partial distortion elimination ME based on local binary patterns (PLBP) and lossy edge-detection pixel decimation technique based on local binary patterns (ELBP). PLBP reduces the matching complexity by matching more contributable pixels early by identifying the most diverse pixels in a local neighborhood. ELBP captures the most representative pixels in a block in order of contribution to the distortion function by evaluating whether the individual pixels belong to the edge or background. Experimental results demonstrate substantial reduction in computational complexity of ELBP with only a marginal loss in prediction quality.

With the rapid growth of video services on smartphones such as video conferencing, video telephone and WebTV, implementation of video compression on mobile terminal becomes extremely important. However, the low computation capability of mobile devices becomes a bottleneck which calls for low complexity techniques for video coding. This work presents two set of algorithms for reducing the complexity of motion estimation. Binary motion estimation techniques using one-bit and two-bit transforms reduce the computational complexity of matching error criterion, however sometimes generate inaccurate motion vectors. The first set includes two neighborhood matching based algorithms which attempt to reduce computations to only a fraction of other methods. Simulation results demonstrate that full search local binary pattern (FS-LBP) algorithm reconstruct visually more accurate frames compared to full search algorithm (FSA). Its reduced complexity LBP (RC-LBP) version decreases computations significantly to only a fraction of the other methods while maintaining acceptable performance. The second set introduces edge detection approach for partial distortion elimination based on binary patterns. Spiral partial distortion elimination (SpiralPDE) has been proposed in literature which matches the pixel-to-pixel distortion in a predefined manner. Since, the contribution of all the pixels to the distortion function is different, therefore, it is important to analyze and extract these cardinal pixels. The proposed algorithms are called lossless fast full search partial distortion elimination ME based on local binary patterns (PLBP) and lossy edge-detection pixel decimation technique based on local binary patterns (ELBP). PLBP reduces the matching complexity by matching more contributable pixels early by identifying the most diverse pixels in a local neighborhood. ELBP captures the most representative pixels in a block in order of contribution to the distortion function by evaluating whether the individual pixels belong to the edge or background. Experimental results demonstrate substantial reduction in computational complexity of ELBP with only a marginal loss in prediction quality.

The development of multimedia technology, along with the emergence of parallel architectures, has revived the interest on dataflow programming for designing embedded systems. Indeed, dataflow programming offers a flexible development approach in order to build complex applications while expressing concurrency and parallelism explicitly. Paradoxically, most of the studies focus on static dataflow models of computation, even if a pragmatic development process requires the expressiveness and the practicality of a programming language based on dynamic dataflow models, such as the language included in the Reconfigurable Video Coding framework. In this thesis, we describe a complete development environment for dataflow programming that eases multimedia development for embedded multi-core platforms. This development environment is built upon a modular software architecture that benefits from modern software engineering techniques such as meta modeling and aspect-oriented programming. Then, we develop an optimized software implementation of dataflow programs targeting desktop and embedded multi-core platforms. Our implementation aims to bridge the gap between the practicality of the programming language and the efficiency of the execution. Finally, we present a set of runtime actors mapping/scheduling algorithms that enable the execution of dynamic dataflow programs over multi-core platforms with scalable performance.