Head-wearable Holographic Head-mounted Display with 6 Degrees of Freedom

Taichi Sakakihara, Teppei Jodo, Seok Kang, Yuji Sakamoto · 2026

Read Head-wearable Holographic Head-mounted Display with 6 Degrees of Freedom by Taichi Sakakihara, Teppei Jodo, Seok Kang, Yuji Sakamoto (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-15581.txt.

Introduction

Taichi Sakakihara, Teppei Jodo, Seok Kang, and Yuji Sakamoto build a holographic head-mounted display in a 2026 arXiv paper. Holographic displays promise a next step past current headsets. They may reduce the three-dimensional sickness that plagues virtual reality. A good display must update images as the head moves, in real time. For holography, that is hard. Computing hologram data takes heavy resources.

Their device targets six degrees of freedom. That means it tracks both rotation and position of the head. Full tracking makes virtual objects stay put as the wearer moves. The paper's challenge is the compute load behind that freedom. Readers in optics and display engineering are the audience. Working holography at head speed would mark real progress toward comfortable headsets.

Catalog record

CREATORTaichi Sakakihara, Teppei Jodo, Seok Kang, Yuji Sakamoto
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE40 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-15581

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Head-wearable Holographic Head-mounted Display with 6 Degrees of Freedom Taichi Sakakihara, Teppei Jodo, Seok Kang, Yuji Sakamoto Abstract A head-mounted display (HMD) using holography technology (holo-HMD) is expected to be the next generation of HMDs capable of reducing three-dimensional sickness. In HMDs, it is important to generate images that respond to head movement in real time. However, in holo-HMDs, generation of hologram data in real time is difficult due to the large computational resources required. This paper proposes a fast calculation algorithm for generating hologram data for holo-HMDs, which requires low computational power. A holo-HMD supporting six degrees of freedom was also developed using this algorithm and it was confirmed that it obtained reconstructed images with six degrees of freedom in real time (30 fps or more). 1 Introduction Head-mounted displays (HMDs) are commonly used as three-dimensional (3D) image-display devices for augmented reality (AR) and virtual reality (VR). An HMD for VR generates parallax images called stereogram by displaying parallax images in the left and right displays, which enables a viewer to see a 3D image. In the stereogram, however, the perceptual depth of a viewer is different from the actual focal depth of the image, because an HMD displays images at the fixed depth determined with the optical system. This mismatch in the two depths causes vergence accommodation conflict (VAC), which is one of the causes of eye fatigue and 3D sickness when using an HMD for long periods [ 1 ] . Even with AR systems using monocular HMDs that do not display 3D images, frequent eye refocusing is required when the depth of the object in the real world differs from that of the optical image, which also causes eye fatigue. Holography is considered an ideal 3D imaging technology because it can display images at any depth [ 2 , 3 ] . Electro-holography is an electronic implementation of the principles of holography, enabling video, electronic communication, and virtual scene display. An HMD utilizing electro-holography is called a holographic HMD (holo-HMD), which is expected to reduce VAC and eye fatigue caused by focus adjustment of the eyes. There have been many studies on holo-HMDs [ 4 ] . Most involved benchtop-system prototypes for research into optical components or for principal purposes [ 5 , 6 , 7 , 8 , 9 , 10 ] , but there have been investigated holo-HMDs that can actually be worn on the head [ 11 , 12 , 13 ] . For VR and AR applications, to achieve a high sense of immersion with an HMD, it is important that the image changes in accordance with the user’s head movement. There are two types of degrees of freedom in the HMD movement: three degrees of freedom, which corresponds to rotational movements of the head, and six degrees of freedom, which accounts for full motion including both rotation and translation. To achieve this, a holo-HMD must sense head movement and generate hologram data in real time. The hologram data are calculated using a computer using a computer-generated hologram (CGH) algorithm and displayed on spatial light modulators (SLMs). However, it is difficult to generate hologram data in a short time, because it requires a huge amount of calculation. The computers attached to HMDs also generally have low computing power. Generating holograms that produce realistic reconstructed images requires even more computation time. It is thus difficult to generate hologram data in real time for holo-HMDs with three degrees of freedom or six degrees of freedom. The final aim of our line of research is to explore the feasibility of practical holo-HMDs. Their most critical challenge is achieving real-time hologram data generation for six degrees of freedom. To address this, we propose a fast calculation algorithm for generating hologram data for holo-HMDs and that requires low computational power. We confirmed that our algorithm can be used to enable a holo-HMD to produce six degrees of freedom reconstructed images in real time. 2 Related work There are various methods for calculating hologram data in CGH. We explain the point-light source method. 2.1 Point-light method Figure 1: point-light method. An object is represented as a collection of point light sources, and object light on hologram plane is obtained as synthesis of spherical waves from each point light source. We used the point-light method [ 14 ] to calculate light waves propagating from an object. This method represents a virtual object as a group of point light sources, and calculates the light-wave propagation from each point light source to the hologram. Finally, these contributions are summed to calculate the complex amplitude distribution of the object light. Consider the object light wave distribution to a hologram of a virtual object defined by point light sources. When the coordinates of an arbitrary point light source are (x_{i},y_{i},z_{i}) and an arbitrary pixel on the hologram is (x,y,0) , as shown in Fig. 1 , the object light wave u_{i} propagating from the point light source to the hologram is expressed as \displaystyle u_{i}(x,y) \displaystyle=\frac{a_{i}}{r_{i}}\exp\{-j(kr_{i}+\phi_{i})\} (1) \displaystyle k \displaystyle=\frac{2\pi}{\lambda} (2) \displaystyle r_{i} \displaystyle=\sqrt{(x-x_{i})^{2}+(y-y_{i})^{2}+z_{i}^{2}}, (3) where a_{i} is the amplitude of the point source, r_{i} is the distance from the point light source to the pixel (x,y,0) on the hologram, k is the wavenumber, and \phi_{i} is the initial phase of the point source. When the total number of point sources of the virtual object is N , the light wave distribution O(x,y) of the object light on the hologram is expressed as the sum of u_{i} from each point source, as in the following equation. O(x,y)=\sum^{N}_{i=1}u_{i}(x,y) (4) This equation shows that the computational complexity of the point-light method is proportional to N and the number of pixels in the hologram P . As a result, the amount of calculation becomes extremely large, and a computer with high computational power is required. The hologram data I(x,y) are obtained by interfering the object light O(x,y) from the reference light R(x,y) , as shown in I(x,y)=\left|O(x,y)+R(x,y)\right|^{2}. (5) 2.2 Fast calculation method Various methods have been proposed to speed up calculations with the point-source method. One important method is the table-look-up method [ 15 , 16 ] , with which the object light from a single light source (Eq. 1) is calculated in advance then used to calculate Eq. 2, which is the superposition of the pre-calculated object light. These methods achieved extremely fast calculations, but real-time calculations were not possible with the CPU of an ordinary PC. By using general-purpose computing on graphics processing Units (GPGPUs) [ 17 , 18 ] , fast calculations can be executed by taking advantage of the massive parallelism of the computing cores. This makes it possible to conduct real-time calculations even with a small number of point-light sources. Currently, GPUs are commonly employed for the point source method, and the proposed algorithm was implemented on a GPU in this study. However, calculating realistic rendering, which can represent hidden surface removal and object textures (shading) [ 19 , 20 ] , requires an even greater amount of calculation, making real-time calculations impossible. Methods for fast realistic rendering have been proposed using the ray tracing cores built into modern GPGPUs [ 21 , 22 ] , but real-time calculations have not

Read the complete text