Patent Pending, UK Patent Application No. 2411463.9, 2024
Winner and Special Mention Awarded at Innovation RCA Pitch Deck 2024
Awarded IRCA Enterprise Ltd's Design & Impact S/EIS Research Seed Fund (£250k) 2024
Note: This research is patent pending and details cannot be disclosed at this time; please contact for further information.
Figure 1: Our Delta-E Color Shifting algorithm aims to reduce light stimulation in VR scenes while preserving color fidelity. The top-left showcases a VR scene comparison between the original, the widely-used "Night Mode", and our method. The bottom-left illustrates their color gamut distribution in the sRGB color space, highlighting Night Mode's significant reduction in blue color range, leading to diminished contrast and detail recognition challenges. Our approach, however, maintains the color gamut distribution. The right chart shows the spectrum comparison in Meta Quest 2, showing our method significantly reduces blue light peak and M-EDI value while maintaining color fidelity, achieving a level comparable to Night Mode.
Vision serves as our primary biological channel for receiving external information. Virtual reality (VR) headsets, utilizing electronic screens to synthesize light, enable us to immerse in artificially created virtual environments. Recent advancements in artificial intelligence (AI) and chip technology have facilitated VR to achieve higher resolutions and refresh rates, significantly enhancing the visual immersion experience. However, a crucial issue remains under-addressed: the light stimulation and resultant physiological discomfort experienced while wearing VR headsets.
In VR, the pixel structure of screens plays a crucial role in shaping the emitted light spectrum, thereby profoundly impacting the visual experience. Predominantly, VR headsets utilize RGB (Red, Green, Blue) color synthesis techniques, which, in contrast to the continuous spectrum of natural daylight, tend to produce elevated levels of blue light. Such heightened blue light emissions are recognized as one of the contributors to visual fatigue. Concurrently, emerging research has shed light on the non-image-forming (NIF) visual system, pivotal in regulating the circadian rhythm through its sensitivity to external light cues, particularly at the 479 nm wavelength (blue-green). Substantial evidence indicates that light from screens can disrupt this critical synchrony, leading to sleeping disorders and hormonal imbalances. However, this issue assumes even greater significance in VR, where the proximity of screens to the eyes intensifies the potential impact, a concern that has been largely overlooked in current VR headset development.
To alleviate discomfort, VR systems often incorporate strategies including a "Night Mode", which involves applying an orange filter on the screen to reduce blue light intensity. This method leverages the color adaptation capability of the human eye to maintain relative color perception. While effective in mitigating light stimulation without hardware modifications, it compresses the original color gamut, reducing image contrast and negatively affecting object recognition and immersive experience.
While "Night Mode" may temporarily be adequate for gaming by offering brief visual adjustment, it falls short in professional fields where rigorous color transmission and extended immersion periods are essential. For instance, the da Vinci Robotic tele-surgery system relies on VR screens for accurate display of human tissues, and earthquake and fire rescue training in VR require precise identification of hazards and vital signs. However, despite significant efforts by developers to create virtual content, existing methods for reducing stimulation have yet to find a balance between color perception and immersion time. This raises a critical question: Is it feasible to develop a re-shading algorithm that diminishes light stimulation without altering hardware, while still preserving color perception?
In this research, we propose a unique re-coloring algorithm - the Delta-E Color Shift. This method, based on the human eye's insensitivity threshold to color changes, adjusts the color of each pixel on the screen. It is meticulously designed to reduce blue or green light stimulation while maintaining similar color perception, especially catering to different requirements such as melanopsin suppression or release during day and night VR usage.
In the algorithm's architecture, we conceptualize the issue through a tripartite framework: Firstly, we devise a scalar function within the color space based on the activation of human photoreceptors. Secondly, we address an optimization problem subject to nonlinear constraints, utilizing the CIEDE2000 color metric, to ensure that color alterations after shifting remain within the perceptual threshold of human vision. Thirdly, we employ polynomial regression to facilitate rapid computation.
To validate the effectiveness of our approach, we designed three evaluation tasks: spectrum and gamut comparison for the same scene across multiple VR headsets, comparisons across different scenes, and a user experiment on comfortable VR usage duration involving 57 participants. In all these assessments, we compared the original scenes with versions modified by our method and the widely-used "Night Mode". Findings from comparisons reveal that our method effectively maintained the color gamut while reducing the energy of blue/green light. Furthermore, in user experiments, our approach not only achieved the longest comfortable immersion time but also scored comparably to the original scene in the subjective scale assessing color perception.
The main contributions of our work are as follows:
-
Delta-E Color Shift Algorithm: Our method effectively reduces light stimulation while maintaining visual color perception, serving as a viable alternative to the "Night Mode". Its universal principle allows for its potential extension to other electronic devices.
-
VR NIF Stimulation Consideration: Our research extends to the NIF visual pathway, addressing the impact of VR screens beyond traditional vision. The DeltaE color shift demonstrates the potential of VR to emulate natural light exposure, aligning with physiological needs.
-
Regression of Delta-E Color Shift Vector Field: Our computation of the Delta-E color shift within both sRGB and CIELAB spaces is expedited through the application of regression outcomes. We supply code that is both Python-based and optimized with Numba for swift execution.