Image and video fusion is the process of combining information from multiple sensors that a single fused image or video is produced containing the useful information from all sources. The data from multiple sensors is often complimentary, either because the sensors are of different modalities, or because the data was recorded from different ositions or at different times. Image fusion has wide ranging applications, including surveillance, navigation, target classification and medicine.

Technological advances in telecommunications and computing power have driven a massive increase in the availability of digital video, both in the traditional broadcast environment and in on-demand scenarios such as the internet and IPTV. In order to support this increase video compression technologies have developed both to reduce bitrates but also to utilise the additional computing power available. The next step for video compression is commonly thought to be in exploitation of the Human Visual System (HVS) in perceptual coding, although this continues to suffer from the difficulty of evaluating the quality of compressed video. This thesis proposes a perceptual video compression framework - from quality estimation, through gaze estimation to variable quality coding.

Video has become a key part of our daily digital lives. An increasing number of people are using video to document their lives and share their experiences through , sites such as YouTube and Facebook. Due to this ever increasing quantity of video content, it is clear that more efficient and robust video compression algorithms are needed. In this thesis video compression and transmission are investigated, specifically focused on High Definition (HD) video coding and wireless transmission. Unfor- tunately, the techniques relied upon for efficient compression are the same tech- niques that hinder robustness. This is because conventional video compression relies on inter-frame prediction (motion estimation), intra frame prediction and variable- length entropy encoding to achieve high compression ratios but, as a consequence, produces an encoded bitstream that is inherently sensitive to channel errors. There- fore, in order to ensure reliable and robust delivery over lossy channels, it is necessary to invoke various additional error detection and correction methods. In this thesis a technique called Pyramid Vector Quantisation (PVQ) is investig- ated. In contrast to the other methods, Pyramid Vector Quantisation has the ability to prevent error propagation through the use of fixed length codewords. This thesis introduces an efficient Rate Distortion Optimisation (RDO) algorithm for intra- mode PVQ which offers similar compression performance and visual quality to intra H.264/ AVC, intra VC-I, VC-3 and Motion JPEG 2000 while offering outstanding inherent error resilience. The results are reported for a range of representative 1080p HD video sequences. The error resilience performance of the enhanced PVQ codec is then evaluated for HD content in the context of a realistic (IEEE 802.11n) wireless environment. The impact of both uncorrelated and correlated channel errors is investigated. It is shown that PVQ provides high tolerance to corrupted data compared to the state-of-the- art H.264/ AVC while obviating the need for complex encoding tools. PVQ (without explicit error resilience tools and without error concealment) is able to significantly outperform H.264/ AVC (with error resilience tools and with error concealment) by up to I3.3dB PSNR for the same wireless simulation scenario. The benefit obtained by employing a Gaussian Markov Random Field (GMRF) based wavelet domain error concealment algorithm to post process erroneously re- ceived PVQ bitstreams is also investigated. It is shown that this algorithm leads to a further 2.5dB performance gain thus outperforming H.264/ AVC by up to 15.8dB PSNR. Overall, it is hoped that the results of this thesis demonstrate PVQ's feasib- ility for in-home HD video coding and transmission.

In order to service heterogenous, variable bandwidth networks, low complexity transcoding architectures have been developed to allow modification of video bitrate. SVC, the scalable extension to H.264, was also developed for this purpose. When encrypted video is transmitted, it cannot be modified prior to decryption. Any decryption carried out at a transcoder would use security weaknesses in the network. In the above context, there is a need to develop transparent encryption techniques that allow bitrate reduction without decryption. For SVC bitstreams, an architecture called SVE is presented in this thesis to allow the full range of scalabilities to be supported without compromising quality or security. This approach is shown to outperform all existing techniques presented in the literature and shows significant potential for commercial deployment.

Advances in surface performance capture have enabled the reconstruction of real world scenes such as people and animals with a realism hard to achieve by an animator. The work presented in this thesis aims to develop techniques for interactive editing and manipulation of captured mesh sequences with the flexibility associated with conventional computer animation techniques. In particular, the application of Laplacian deformation for animation and compression of surface motion capture data is investigated. Laplacian deformation enables the manipulation of a mesh at a vertex level while maintaining its local geometric properties but lacks a mechanism for ensuring the preservation of its underlying physical structure. Motivated by this limitation, a learnt surface deformation basis constructed in the space of differential cqordinates is introduced. The incorporation of this basis into the Laplacian framework constrains the solution to the space of plausible deformations built from a set of examples, therefore preserving the structure of the mesh. The successful application of this approach to space-time editing together with a set of novel non-linear edit propagation techniques are presented. Representations for efficient storage of surface motion capture sequences, generally comprised of hundreds of frames with thousands of vertices, are investigated. A novel layered representation that exploits the articulated nature of the data is presented and compared with other compression techniques based on PCA and Laplacian deformation, with and without using the aforementioned surface deformation basis. The proposed layered representation achieves consistently high compression ratio with low maximum reconstruction errors in three test sequences from different characters.

No description available.

No description available.

No description available.

No description available.

No description available.