A computer-vision-driven dynamic color rendering algorithm for virtual reality and its application in remote art creation systems
Year: 2026
Type: article
Abstract: Remote art creation in virtual reality continues to be undermined by three coupled problems: color distortion under transient illumination, tail latency during rapid brushwork, and chromatic divergence across heterogeneous headsets. This study develops a computer-vision-driven dynamic color rendering algorithm and embeds it within a working remote art creation platform. The algorithm consists of three loosely coupled stages: a chroma-preserving feature encoder that reads scene context, a hypernetwork-conditioned non-linear mapping module weighted by spatial and channel attention, and a saliency-driven scheduler that applies per-terminal correction within a sub-frame budget. Color is treated as a first-class replicated quantity through a CRDT-based synchronization layer, with bandwidth-adaptive quantization preserving consistency under degraded networks. Experiments on a mixed HTC Vive Pro and Oculus Quest platform across 240 painted scenes show cross-terminal CIEDE2000 values near 2.1 under high-bandwidth conditions, against five to twelve for current platforms, with 95th-percentile latency held below 20 ms. A user study with thirty practicing artists across paired painting, mural construction, and instructor-led demonstration scenarios returned composite experience scores of 6.18 against 4.79 for the strongest baseline, with the gap concentrated in the color consistency and coordination dimensions. Robustness tests under cycling illumination, HDR scenes, and concurrent collaboration up to thirty-two peers confirmed graceful scaling, with logarithmic convergence growth and material-class generality across painting, sculpture, and installation work. Limitations remain at sub-5 Mbps bandwidths and on 8K-class panels, pointing toward multimodal sensing and aesthetic-quality optimization as next steps.
Source: Scientific Reports
Language: English
Subfield: Atomic and Molecular Physics, and Optics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7235881752
