P‐16.4: Research on Deep Learning‐Based Color Calibration Method for Laser Projection Display
SID Symposium Digest of Technical Papers · 2026-08 · 57 (S1) · 2064-2067
DOI: 10.1002/sdtp.70003_555 ↗Kurz gefasst
Fachartikel in der Zeitschrift SID Symposium Digest of Technical Papers, online erschienen am 11.09.2026. Verfasst von Ranwei Chen und 3 weiteren Autor:innen. Die Arbeit nennt 12 Literaturangaben.
Abstract (Originalsprache)
With the increasing adoption of wide‐gamut displays, inherent device variability in laser display hardware has led to significant color inconsistency. Traditional 3D lookup table (3D‐LUT) methods face inherent limitations in calibration accuracy due to sparse sampling and interpolation errors, while requiring substantial memory for high‐precision mapping. To address these challenges, we propose a novel calibration scheme that integrates a residual Multilayer Perceptron (MLP) network with a color difference‐driven partitioning algorithm. The process begins by applying the CAT16 chromatic adaptation transform to normalize the white point to D65. A residual neural network is then trained to establish a high‐precision base mapping from the RGB color space to CIELAB. To further refine accuracy, the color gamut is adaptively partitioned into continuous, non‐overlapping sub‐regions based on prediction error and spatial connectivity. A fusion model balances the global mapping with these localized calibrations. Finally, an efficient batch lookup strategy is implemented for practical application. Experimental results demonstrate that the proposed method significantly reduces the average color difference ΔE from 3.81 to 0.97, outperforming conventional 3D‐LUT and standalone neural network approaches in calibration performance.
Zitieren
APA
Chen, R., Guo, D., Liu, H., & Yuan, G. (2026). P‐16.4: Research on Deep Learning‐Based Color Calibration Method for Laser Projection Display. SID Symposium Digest of Technical Papers, 57(S1), 2064-2067. https://doi.org/10.1002/sdtp.70003_555
Vancouver
Chen R, Guo D, Liu H, Yuan G. P‐16.4: Research on Deep Learning‐Based Color Calibration Method for Laser Projection Display. SID Symposium Digest of Technical Papers. 2026;57(S1):2064-2067. doi:10.1002/sdtp.70003_555
BibTeX
@article{Chen2026,
author = {Chen, Ranwei and Guo, Dabo and Liu, Huiping and Yuan, Guang},
title = {P‐16.4: Research on Deep Learning‐Based Color Calibration Method for Laser Projection Display},
journal = {SID Symposium Digest of Technical Papers},
year = {2026},
volume = {57},
number = {S1},
pages = {2064--2067},
doi = {10.1002/sdtp.70003_555}
}
EUVIMED