用8位整数量化神经网络，让全息图生成更快更省电。
Quantized neural network for complex hologram generation
- 将全息生成模型从32位浮点转为8位整数，大幅压缩体积。
- 模型大小减少70%，速度提升4倍，画质基本不变。
- 已在嵌入式平台验证，适合移动AR设备部署。
计算机生成全息图（CGH）是增强现实显示（如头戴或抬头显示）的有前景技术，但其高计算需求使其难以实用化。近年来将神经网络引入CGH已显著加速计算，有望突破计算成本与图像质量的权衡。然而，在计算资源受限的嵌入式系统上部署基于神经网络的CGH算法，仍需更高效、低功耗、低内存占用的模型。本研究通过引入神经网络量化，构建了一种轻量级复杂全息图生成模型。具体而言，基于张量全息架构，将模型从32位浮点精度（FP32）量化至8位整数精度（INT8）。性能评估表明，所提出的INT8模型在保持与FP32模型相当的全息图质量的同时，模型尺寸缩减约70%，推理速度提升四倍。此外，我们已在系统级模组上实现该模型，验证了其在嵌入式平台上的可部署性与高能效特性。
原文摘要 · Abstract (English)
Computer-generated holography (CGH) is a promising technology for augmented reality displays, such as head-mounted or head-up displays. However, its high computational demand makes it impractical for implementation. Recent efforts to integrate neural networks into CGH have successfully accelerated computing speed, demonstrating the potential to overcome the trade-off between computational cost and image quality. Nevertheless, deploying neural network-based CGH algorithms on computationally limited embedded systems requires more efficient models with lower computational cost, memory footprint, and power consumption. In this study, we developed a lightweight model for complex hologram generation by introducing neural network quantization. Specifically, we built a model based on tensor holography and quantized it from 32-bit floating-point precision (FP32) to 8-bit integer precision (INT8). Our performance evaluation shows that the proposed INT8 model achieves hologram quality comparable to that of the FP32 model while reducing the model size by approximately 70% and increasing the speed fourfold. Additionally, we implemented the INT8 model on a system-on-module to demonstrate its deployability on embedded platforms and high power efficiency.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。