High-fidelity real-world lighting is a complex and rapidly expanding field of study in computer graphics. Rendering with real-world lighting plays a crucial part in motion pictures, computer games, Augmented Reality (AR) and Virtual Reality (VR) applications. There are, however, many constraints when capturing and representing real-world lights for rendering. In particular, dimensionality plays a significant role although existing industry-standard methods are inadequate to capture light throughout the three spatial, two angular and a temporal dimension simultaneously. Image Based Lighting (IBL) techniques addresses temporality by capturing two angular and the temporal dimension simultaneously. The Incident Light Fields (ILF) technique, on the other hand, can capture complex spatially varying real-world light incident on a static scene covering five angular and spatial dimensions. However, any changes in the positions or the radiometric properties of direct light sources in the scene over time invalidates the captured ILF due to the subsequent changes in the indirect lighting. In a dynamically varying lighting condition, ILF needs to be recaptured with each change in the lighting which is infeasible in most real-world situations. This thesis proposes a novel technique called “Dynamic Change Propagation” (DCP) that can simulate any changes made in the direct light and propagate the effects to the indirect lighting recorded in a captured ILF. Evaluations show average RMSE errors of 0.034 with absolute percentage errors of 6.8% for light source movement simulation and 0.013 (RMSE) for 3.4% for intensity change simulations. In addition to the DCP technique, this thesis proposes a novel “Temporal Incident Light Field” (Temporal ILF) technique which records the changes in the light sources over time and utilizes the DCP technique to simulate those changes into the originally recorded static ILF thus, capturing six (spatial, angular and temporal) dimensions. To the best of our knowledge, Temporal ILF is the first method which can record high-fidelity real-world light over all six spatial, angular and temporal dimensions simultaneously. The introduction of the DCP and Temporal ILF techniques in this thesis offers new ways of rendering with spatio-temporally variant high-fidelity real-world light.

High dynamic range (HDR) imaging techniques enable the full range of light in a scene to be captured, transmitted and displayed, eliminating under- and over-exposed regions. Storing the full range of light typically requires 32 bits of information per colour channel, four times larger than the 8 bits required for low dynamic range (LDR) data. If HDR is to fulfil its potential in a wide variety of applications such as live broadcast and interactive remote gaming, fast, efficient compression is required to provide HDR video without major changes to existing communications infrastructure. A number of methods have so far been proposed for HDR video compression, however they rely on computationally expensive transformations to either split the video into multiple 8-bit streams or convert to a perceptually uniform domain. This thesis will address the question of whether high-quality HDR video compression can be achieved in a computationally efficient manner by introducing a number of novel techniques. Initially, the power-law functions used by LDR video are extended to HDR data to provide a straightforward and efficient solution to HDR compression. The Power Transfer Function (PTF) is computationally inexpensive and an objective evaluation shows that it provides improved compression quality at a thirtieth of the computational cost of other leading methods while maintaining equivalent perceived quality. Subsequently, information about the content and ambient environment at the display are used to adaptively transform the compression method. A subjective evaluation involving 40 participants demonstrates that the necessary peak luminance of content is dependent on the ambient illumination of the display, and an objective evaluation confirms that optimal compression is affected by the peak luminance. An adaptive extension to PTF, Adaptive PTF (PTFa) is proposed using iterative optimisation to calculate the ideal compression curve, gaining 2.88 VDP over non-adaptive PTF. Finally, the computational performance of PTFa is improved by four orders of magnitude to enable real-time compression with little decrease in quality through deep learning. Predictive PTF (PTFp) utilises a model to predict the compression curved based on the content, display and ambient environment. This thesis demonstrates a fast, efficient method of general HDR video compression which is extended to provide a fast, adaptive solution for content-specific compression.

Post-production facilities deal with large amounts of digital video, which presents difficulties when tracking, managing and searching this material. Recent research work in image and video analysis promises to offer help in these tasks, but there is a gap between what these systems can provide and what users actually need. In particular the popular research models for indexing and retrieving visual data do not fit well with how users actually work. In this thesis we explore how image and video analysis can be applied to an online video collection to assist users in reviewing and searching for material faster, rather than purporting to do it for them. We introduce a framework for automatically generating static 2-dimen- sional storyboards from video sequences. The storyboard consists of a series of frames, one for each shot in the sequence, showing the principal objects and motions of the shot. The storyboards are rendered as vector images in a familiar comic book style, allowing them to be quickly viewed and understood. The process consists of three distinct steps: shot-change detection, object segmentation, and presentation. The nature of the video material encountered in a post-production fa- cility is quite different from other material such as television programmes. Video sequences such as commercials and music videos are highly dy- namic with very short shots, rapid transitions and ambiguous edits. Video is often heavily manipulated, causing difficulties for many video processing techniques. We study the performance of a variety of published shot-change de- tection algorithms on the type of highly dynamic video typically encoun- tered in post-production work. Finding their performance disappointing, we develop a novel algorithm for detecting cuts and fades that operates directly on Motion-JPEG compressed video, exploiting the DCT coeffi- cients to save computation. The algorithm shows superior performance on highly dynamic material while performing comparably to previous algorithms on other material.

Conventional 3D imaging techniques such as laser scanning, focus-stacking and confocal microscopy either require scanning in all or a subset of the spatial dimensions, or else are limited by their depth of field (DOF). Scanning increases the acquisition time, therefore techniques which rely on it cannot be used to image moving scenes. In order to acquire both the intensity of the scene and its depth, extending the DOF without scanning is therefore necessary. This is traditionally achieved by stopping the system down (reducing the f/#). This, however, has the highly undesirable effect of lowering both the throughput and the lateral resolution of the system. In microscopy in particular, both these parameters are critical, therefore there is scope in breaking this trade-off. The objective of this work, therefore, is to develop a practical and simple 3D imaging technique which is capable of acquiring both the irradiance of the scene and its depth in a single snapshot over an extended DOF without incurring a reduction in optical throughput and lateral resolution. To this end, a new imaging technique, referred to as complementary Kernel Matching (CKM), is proposed in this thesis. To extend the DOF, in CKM a hybrid imaging technique known as wavefront coding (WC) has been used. WC permits the DOF to be extended by an order of magnitude typically without reducing the efficiency and the resolution of the system. Moreover, WC only requires the introduction of a phase mask in the aperture of the system, hence it also has the benefit of simplicity and practicality. Unfortunately, in practice, WC systems are found to suffer from post-recovery artefacts and distortion, which substantially degrade the quality of the acquired image. To date, this long-standing problem has found no solution and is probably the cause for the lack of exploitation of this imaging technique by the industry. In CKM, use of a largely ignored phenomenon associated with WC was made to measure the depth of the sample. This is the lateral translation of the scene in proportion to its depth. Furthermore, once the depth of the scene is known, the ensuing artefacts and distortion due to the introduction of the WC element can be compensated for. As a result, a high quality intensity image of the scene and its depth profile (referred to in stereo vision parlance as a depth map) is obtained over a DOF which is typically an order of magnitude larger than that of an equivalent clear-aperture system. This implies that, besides being a 3D imaging technique, CKM is also a solution to one of the longest standing problem in WC itself. By means of WC, therefore, the DOF was extended without scanning and without reducing the throughput and the optical resolution, allowing both an intensity image of the scene to be acquired and its depth map. In addition, CKM is inherently monocular, therefore it does not suffer from occlusion, which is a major problem affecting triangulation-based 3D imaging techniques such as the popular stereo vision. One therefore concludes that CKM fulfils the objectives set for this project. In this work, various ways of implementing CKM were explored and compared; and the theory associated with them was developed. An experimental prototype was then built and the technique was demonstrated experimentally in microscopy. The results show that CKM eliminates WC artefacts and thus gives high quality images of the scene over an extended DOF. A DOF of ∼ 20μm was achieved on a 40×, 0.5NA system experimentally, however this can be increased if required. The experimental depth reconstructions of real samples (such as pollen grains and a silicon die) imaged in various modalities (reflection, transmission and fluorescence) were comparable to those given by a focus-stack. However, as with all other passive techniques, the performance of CKM depends on the texture and features in the scene itself. On a binary systematic scene consisting of regularly spaced dots with a linear depth gradient, an RMS error of ±0.15μm was obtained from an image signal-to-noise ratio of 60dB. Finally, owing to its simplicity and large DOF, there is scope in investigating the possibility of using the same CKM setup for 3D point localisation applications such as super resolution. An initial investigation was therefore conducted by localising sub-resolution fluorescent beads. On a 40×, 0.5NA system, a mean precision of 148nm in depth and < 30nm in the lateral dimensions was observed experimentally from 4, 000 photons per localisation over a DOF of 26μm. From these experimental values, a mean localisation precision of < 34nm in depth and < 13nm in the lateral dimensions from 2, 000 photons per localisation over a DOF of 3μm is expected on a more typical 100×, 1.4NA system. This compares favourably to the competition, therefore we conclude that there is scope in investigating this technique for 3D point localisation applications further.

Separate landscape is a research that combines a theory and practice through the examination of 'non-place'. Non-places such as airports, waiting lounges, car parks, shopping malls have been defined as places which lack a sense of history, social relations, and identity.

Facial information can reveal the emotional status of individuals. Although traditional cameras can capture this information, such cameras struggle to acquire the necessary information in extreme lighting conditions. This thesis aim to investigate whether High Dynamic Range (HDR) imaging can capture human facial expression under complex lighting conditions, and in doing so, enhance Facial Expression Recognition (FER) performance. The techniques presented in this thesis focus on developing a baseline for images captured in scenes with harsh lighting conditions, where Low Dynamic Range (LDR) images have difficulty capturing the full range of light in a single exposure. The thesis considers unprocessed images and a variety of pre-processing methods to examine whether reducing the impact of large lighting variations could improve the quality of an input image. In addition, realistic facial data plays a key role in validating facial expression analysis systems. Today, the majority of FER algorithms are evaluated only on images generated in highly controlled laboratory environments. The variability of a facial appearance in an image could be dominated by changes in head pose and illumination conditions. This can effectively hide features that are necessary to discriminate different subjects or different facial articulations. New HDR imaging techniques are thus introduced to help ensure that all the details in a scene is captured no matter what the lighting conditions present, and all this detail is then available to the FER algorithms. This is also investigated on Face recognition algorithms.

Human computation can be described as outsourcing part of a computational process to humans. This technique might be used when a problem can be solved better by humans than computers or it may require a level of adaptation that computers are not yet capable of handling. This can be particularly important in changeable settings which require a greater level of adaptation to the surrounding environment. In most cases, human computation has been used to gather data that computers struggle to create. Games with by-products can provide an incentive for people to carry out such tasks by rewarding them with entertainment. These are games which are designed to create a by-product during the course of regular play. However, such games have traditionally been unable to deal with requests for specific data, relying instead on a broad capture of data in the hope that it will cover specific needs. A new method is needed to focus the efforts of human computation and produce specifically requested results. This would make human computation a more valuable and versatile technique. Mutually reinforcing systems are a new approach to human computation that tries to attain this focus. Ordinary human computation systems tend to work in isolation and do not work directly with each other. Mutually reinforcing systems are an attempt to allow multiple human computation systems to work together so that each can benefit from the other's strengths. For example, a non-game system can request specific data from a game. The game can then tailor its game-play to deliver the required by-products from the players. This is also beneficial to the game because the requests become game content, creating variety in the game-play which helps to prevent players getting bored of the game. Mobile systems provide a particularly good test of human computation because they allow users to react to their environment. Real world environments are changeable and require higher levels of adaptation from the users. This means that, in addition to the human computation required by other systems, mobile systems can also take advantage of a user's ability to apply environmental context to the computational task. This research explores the effects of mutually reinforcing systems on mobile games with by-products. These effects will be explored by building and testing mutually reinforcing systems, including mobile games. A review of existing literature, human computation systems and games with by-products will set out problems which exist in outsourcing parts of a computational process to humans. Mutually reinforcing systems are presented as one approach of addressing some of these problems. Example systems have been created to demonstrate the successes and failures of this approach and their evolving designs have been documented. The evaluation of these systems will be presented along with a discussion of the outcomes and possible future work. A conclusion will summarize the findings of the work carried out. This dissertation shows that extending human computation techniques to allow the collection and classification of useful contextual information in mobile environments is possible and can be extended to allow the by-products to match the specific needs of another system.

The thesis examines the work of photographer Nick Hedges (b. 1953) who made photographs for the housing charity Shelter between 1968 and 1972. It concentrates on Hedges’ methodology, his representation of the homeless child, and how this was deployed in Shelter’s campaign strategy. Moreover, it examines the wider political, sociological and cultural debates surrounding the conception, production, dissemination and reception of the Shelter photographs. The thesis argues that Hedges’ photographs, although contextualised by an ostensibly radical charity agenda, were shaped by an established photographic and art historical tradition reaching back to the nineteenth century. This is examined in the light of a shifting conception of what constituted an ethically sound representation of homelessness amongst leftist critics in Britain from the 1970s onwards. The thesis equally discusses the archive as a site of photographic accession, interpretation and display, and outlines the issues that face archive professionals charged with the presentation of the Shelter photographs to a contemporary audience. By combining art historical analysis of Hedges’ photographs with research into their current framing in the archive, the thesis offers a distinctive contribution to scholarship, exploring how photographic meaning is shaped, subverted and disseminated by individuals, organisations and institutions alike.

To date, the study of Chinese propaganda photography has been limited. While some research has been made on post-1949 photography, the photographic production of the pre-1949 period has not been sufficiently explored. Focusing on the years of the Second Sino-Japanese War (1937-45), this thesis aims at addressing this gap in the literature and at providing an analysis of how the Chinese Communist Party exploited photography for propaganda purposes during the war. Through the images taken by Party-affiliated photographers and printed on the Jin Cha Ji Pictorial, the first Communist photographic propaganda magazine, this study aims to show how this type of visual propaganda aimed not only at narrating the events of the war against Japan, but also at creating a new idea of the Chinese nation. This thesis is divided into four chapters. The first, The Jin Cha Ji Pictorial: A Brief History presents the history of the magazine and the work of the CCP affiliated photographers who contributed to its creation and popularity. Chapter two, The Geography of a Revolution, explores how a new cultural landscape was visually constructed to create the basis of the political legitimation that the CCP needed during wartime. Chapter three, Becoming Modern Women, investigates the symbolic and ideological value of the spinning wheel in 1943 in relation to women’s contribution to the war effort and the thorny issue of women empowerment. Lastly, chapter four, Moulding the Future looks at the visual representation of childhood and discusses the issue of militarisation and masculinisation of childhood during wartime. This study ends with few considerations on the propagandistic, historical and artistic value of Communist propaganda photography during the Second Sino-Japanese War as well as a reflection on how the symbolic and ideological significance of some of the photographs presented here are still recognisable in contemporary Chinese propaganda.

The Berlin photographer Marta Astfalck-Vietz was active in the Weimar Republic, between 1924 and 1936. Her work was re-discovered in 1989 by the curator Janos Frecot, who staged a solo exhibition of her work in the Berlinische Galerie in 1991, and her archive, comprising just over 400 prints, is now held there. Since this show, the work of Marta Astfalck-Vietz has rarely been exhibited. In both German and Anglo-American scholarship it has received sporadic attention; the only text devoted to her work is still the catalogue to Frecot’s 1991 show. This thesis presents her photography to an Anglo-American audience, not as a compensatory exercise in retrieving a forgotten woman artist, but as a contribution to the existing body of literature on women in the Weimar era. Like other female artists of the period, Marta Astfalck-Vietz negotiated between her private experiences of life in the Republic and representations of its social and political tensions, which were circulated in novels, in films, on the stage, and in magazines. Her photographs address the issues of sex gender and race that preoccupy historians of Weimar Germany, confirming, confounding, and expanding our knowledge of the era.