Title Page
Abstract
Contents
Chapter 1. Introduction 14
1.1. Overview of holographic displays 14
1.2. Motivation and purpose of this dissertation 16
1.3. Scope and organization 18
Chapter 2. Visual perception 21
2.1. Introduction 21
2.2. Image perception 21
2.2.1. Traditional evaluation metrics of image contents 22
2.2.2. Perceptual image quality metrics 23
2.2.3. Metrics from pairwise comparison 24
2.3. Depth perception 26
2.3.1. Measurements of depth perception 27
2.4. Three-dimensional perceptual realism 29
Chapter 3. Depth of field and speckle noise in 2D holographic displays and their effects on accommodation response 31
3.1. Introduction 31
3.1.1. Holographic wave propagation 32
3.1.2. CGH encoding 33
3.1.3. Phase uniformity and depth of field 34
3.1.4. Accommodation measurement with holographic displays 36
3.2. Motivation 37
3.2.1. Limited depth of field in high-quality 2D CGH 37
3.2.2. Low image quality in classical 2D CGH 38
3.3. Improving accommodation response 39
3.3.1. Contrast ratio of holographic stimuli 39
3.3.2. Principle 43
3.3.3. Speckle reduction through temporal multiplexing 43
3.3.4. CGH optimization with contrast ratio regularization 47
3.4. Implementation 49
3.4.1. Hardware 49
3.4.2. Software 52
3.4.3. Camera-based assessments 53
3.5. Accommodation experiments 58
3.5.1. Methods 58
3.5.2. Results 62
3.6. Subjective image quality assessments 65
3.6.1. Methods 65
3.6.2. Results 68
3.7. Discussion 71
3.8. Conclusion 74
Chapter 4. Holographic parallax and its effects on 3D perceptual realism 75
4.1. Introduction 75
4.1.1. Parallax and occlusion as depth perception cues 76
4.1.2. Gaze-contingent VR rendering 77
4.2. Holographic near-eye displays with 3D CGH 78
4.2.1. Various 3D target formats 78
4.2.2. System scheme for 3D holographic near-eye displays 80
4.2.3. CGH supervision with various 3D target formats 82
4.3. Optimal CGH supervision targets for 3D perceptual realism 88
4.3.1. Simulation 89
4.3.2. Implementation 95
4.3.3. User validation 102
4.4. Number of view requirements in light field-based CGH 112
4.4.1. User validation 112
4.4.2. Analysis 118
4.5. Discussion 120
4.6. Conclusion 121
Chapter 5. Conclusion 123
Bibliography 125
Appendix 136
초록 137
Figure 1.1. Recording and reconstruction procedure of analog, digital holograms. 14
Figure 1.2. Block diagram for illustration of dissertation organization. 18
Figure 3.1. Concise schematic of conventional holographic near-eye displays. 32
Figure 3.2. Illustration of differentiable 2D CGH optimization pipeline. 34
Figure 3.3. Relationship between the phase uniformity and the depth of field. 35
Figure 3.4. Reconstructed images of an incoherent display with two different focal states: focused and defocused with 0.6 diopter (D, reciprocal of the metric... 37
Figure 3.5. Reconstructed holographic images of 2D CGH acquired with SGD algorithm and the magnitude of the field's angular spectrum. The dashed line in... 38
Figure 3.6. Reconstructed holographic images of 2D CGH acquired with classical GS algorithm and the magnitude of the field's angular spectrum. 38
Figure 3.7. Comparison of contrast ratio of holographic images realized with SGD CGH, GS CGH with ground truth case. 41
Figure 3.8. Speckle noise present in both focused image (left) and defocused image (right). 42
Figure 3.9. Concept of holographic near-eye displays with narrowed effective DOF to improve accommodation response. 44
Figure 3.10. Reconstructed holographic images of B-SGD CGHs and corresponding contrast curves with different TM conditions. 45
Figure 3.11. Reconstructed single-frame holographic images and corresponding contrast curves with different contrast ratio regularizers. 46
Figure 3.12. Schematic of the apparatus built for the experimental assessments. (LD: laser diode, CL: collimating lens, BS: beam splitter, LP: linear polarizer,... 49
Figure 3.13. Overview of the benchtop prototype utilized in the experimental evaluations. 50
Figure 3.14. Experimental results of holographic images acquired with various algorithms. 53
Figure 3.15. Experimental results of castle scene of various binary CGH algorithms and TM conditions are provided with PSNR. 55
Figure 3.16. Experimental results of lion scene of various binary CGH algorithms and TM conditions are provided with PSNR. 56
Figure 3.17. Experimental results of market scene of various binary CGH algorithms and TM conditions are provided with PSNR. 57
Figure 3.18. Accommodation response measurement experiments with a holographic near-eye display. 59
Figure 3.19. Results of accommodation experiments. 63
Figure 3.20. Subjective image quality evaluation on holographic contents through pairwise comparisons. 66
Figure 3.21. Results of subjective image quality evaluation. 68
Figure 3.22. Effects of artifacts such as dust in the captured holographic images with different conditions. 69
Figure 4.1. Depth perception cues and its sensitivity. 76
Figure 4.2. Motion parallax and dynamic occlusion. 77
Figure 4.3. Various 3D target formats and the reconstructed epipolar plane image (EPI) corresponding to the target type. 78
Figure 4.4. Illustration demonstrating (a) schematic of holographic near-eye display and the (b) relationship between wavefront recording plane (WRP) domain... 80
Figure 4.5. A direct comparison of the surrogate gradient approaches for 4D supervision. (a) convergence graph of different binary CGH optimization ap-... 85
Figure 4.6. Illustration of the eyebox domain of holographic near-eye display. 89
Figure 4.7. Comparison of 4D CGH supervision with 2.5D (left), 3D (right) CGH supervision in various pupil states. 90
Figure 4.8. Near-depth focused holographic images with different CGH supervision targets reconstructed under three different pupil settings 92
Figure 4.9. Extensive assessments on the reconstructed holographic images with various 3D CGH supervision targets. 94
Figure 4.10. Light field map, RGB-D image, and epipolar plane image (EPI) of the scenes. 97
Figure 4.11. The photograph depicts the testbed of a holographic near-eye display prototype used for user validation. 99
Figure 4.12. The apparatus for the user study. 102
Figure 4.13. Experimental results with different pupil positions. 104
Figure 4.14. 3D perceptual realism user evaluation results. 107
Figure 4.15. Tracked eye trajectory of fourteen subjects in a single session. 110
Figure 4.16. Reconstruction results of landscape night, village scene. 113
Figure 4.17. Reconstruction results of village mirror, dragon bunny scene. 114
Figure 4.18. Experimental results with different number of views (3x1, 5x2, 7x3, and 9x4) used in 4D CGH supervision. 116
Figure 4.19. Scene-dependent user experiment results with a different number of views (3x1, 5x2, 7x3, and 9x4) used in 4D CGH supervision. 117
Figure 4.20. Required number of views in horizontal for 4D CGH supervision based on light field sampling theorem. 119