Fusion-3D: Integrated Acceleration for Instant 3D Reconstruction and Real-Time Rendering
Sixu Li, Yang Zhao, Chaojian Li, Bowei Guo, Jingqun Zhang, Wenbo Zhu, Zhifan Ye, Cheng Wan, Yingyan Celine Lin
摘要
Recent breakthroughs in Neural Radiance Field (NeRF) based 3D reconstruction and rendering have spurred the possibility of immersive experiences in augmented and virtual reality (AR/VR). However, current NeRF acceleration techniques are still inadequate for real-world AR/VR applications due to: 1) the lack of end-to-end pipeline acceleration support, which causes impractical off-chip bandwidth demands for edge devices, and 2) limited scalability in handling large-scale scenes. To tackle these limitations, we have developed an end-to-end, scalable 3D acceleration framework called Fusion-3D, capable of instant scene reconstruction and real-time rendering. Fusion-3D achieves these goals through two key innovations: 1) an optimized end-to-end processor for all three stages of the NeRF pipeline, featuring dynamic scheduling and hardware-aware sampling in the first stage, and a shared, reconfigurable pipeline with mixed-precision arithmetic in the second and third stages; 2) a multi-chip architecture for handling large-scale scenes, integrating a three-level hierarchical tiling scheme that minimizes inter-chip communication and balances workloads across chips. Extensive experiments validate the effectiveness of Fusion-3D in facilitating real-time, energy-efficient 3D reconstruction and rendering. Specifically, we tape out a prototype chip in 28nm CMOS to evaluate the effectiveness of the proposed end-to-end processor. Extensive simulation based on the on-silicon measurements demonstrates aandthroughput improvement in training and inference, respectively, compared to state-of-the-art accelerators. Furthermore, to assess the multi-chip architecture, we integrate four chips into a single PCB as a prototype. Further simulation results show that the multi-chip system achieves aandthroughput improvement in training and inference, respectively, over the Nvidia 2080Ti GPU. To the best of our knowledge, Fusion-3D is the first to achieve both instant (≤ 2 seconds) 3D reconstruction and real-time (≥ 30 FPS) rendering, while only requiring the bandwidth of the most commonly used USB port (0.625 GB/s, 5 Gbps) in edge devices for off-chip communication.
问问这篇 Paper
问问你的智能体。
Lune 读过与它相关的顶会 Paper,每个回答都会注明依据哪几篇。
引用它的顶会 Paper5
- GCC: A 3DGS Inference Architecture with Gaussian-Wise and Cross-Stage Conditional ProcessingMinnan Pei, Gang Li, Junwen Si, Zeyu Zhu 等MICRO 2025 · 被引用 7 次
- Neo: Real-Time On-Device 3D Gaussian Splatting with Reuse-and-Update Sorting AccelerationChanghun Oh, Seongryong Oh, Jinwoo Hwang, Yoonsung Kim 等ASPLOS 2026 · 被引用 6 次
- FlexNeRFer: A Multi-Dataflow, Adaptive Sparsity-Aware Accelerator for On-Device NeRF RenderingSeock-Hwan Noh, Banseok Shin, Jeik Choi, Seungpyo Lee 等ISCA 2025 · 被引用 3 次
- GRTX: Efficient Ray Tracing for 3D Gaussian-Based RenderingJunseo Lee, Sangyun Jeon, Jungi Lee, Junyong Park 等HPCA 2026 · 被引用 2 次
- FractalCloud: A Fractal-Inspired Architecture for Efficient Large-Scale Point Cloud ProcessingYuzhe Fu, Changchun Zhou, Hancheng Ye, Bowen Duan 等HPCA 2026 · 被引用 1 次
相关 Paper
- Instant-3D: Instant Neural Radiance Field Training Towards On-Device AR/VR 3D ReconstructionSixu Li, Chaojian Li, Wenbo Zhu, Boyang Tony Yu 等ISCA 2023 · 被引用 79 次
- FastNeRF: High-Fidelity Neural Rendering at 200FPSStephan J. Garbin, Marek Kowalski, Matthew Johnson, Jamie Shotton 等ICCV 2021 · 被引用 778 次
- Re-ReND: Real-time Rendering of NeRFs across DevicesSara Rojas, Jesus Zarzar, Juan C. Pérez, Artsiom Sanakoyeu 等ICCV 2023 · 被引用 27 次
- Uni-Render: A Unified Accelerator for Real-Time Rendering Across Diverse Neural RenderersChaojian Li, Sixu Li, Linrui Jiang, Jingqun Zhang 等HPCA 2025 · 被引用 6 次
- UE4-NeRF: Neural Radiance Field for Real-Time Rendering of Large-Scale SceneJiaming Gu, Minchao Jiang, Hongsheng Li, Xiaoyuan Lu 等NeurIPS 2023 · 被引用 37 次
