Physics-informed CNN convolution for emulating scattering in quantum-dot color conversion displays
Abstract
We propose a physics-informed convolutional framework for simulating optical scattering in quantum-dot color conversion films (QDCFs). Scattering intensity distributions, computed using LightTools (LT), are spatially discretized into 3 × 3, 5 × 5, 7 × 7, and 9 × 9 grids. The integrated energy within each region is normalized to construct convolution kernels, where each element represents the proportion of scattered energy from a central source to neighboring positions. This kernel structure conceptually mirrors the convolution mechanism in convolutional neural networks (CNNs) and enables spatial energy redistribution through iterative application. When applied to input images, the LT-derived kernels yield scattering patterns that closely match those generated by full LT simulations. In contrast, random kernels produce degraded image quality. Using a 5 × 5 kernel and 7 convolution iterations, the simulated result achieves an SSIM of 0.95, a standard deviation of 7.51, and an MSE of 57.02. Furthermore, we observe that larger kernels such as 9 × 9 achieve optimal performance with fewer iterations, with the 9 × 9 kernel applied twice yielding the closest match to the LT reference in both SSIM and edge preservation metrics. Evaluation based on Laplacian of Gaussian (LoG) edge analysis and structural similarity confirms the robustness of the method. The proposed framework provides an interpretable and efficient tool for predicting scattering behavior in QDCF-based display systems…
Read the paper
Related studies
- Quantum dots: synthesis, bioapplications, and toxicity
- A Guide to In Silico Drug Design
- Quantum Computing for Healthcare: A Review
- Automated in Silico Design of Homogeneous Catalysts
- Identification of Novel 3-Hydroxy-pyran-4-One Derivatives as Potent HIV-1 Integrase Inhibitors Using in silico Structure-Based Combinatorial Library Design Approach
- In silico structural elucidation of RNA-dependent RNA polymerase towards the identification of potential Crimean-Congo Hemorrhagic Fever Virus inhibitors
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.