提出渐进剪枝方法，让立体匹配模型在边缘设备上实时运行且保持高精度。
Pip-Stereo: Progressive Iterations Pruner for Iterative Optimization based Stereo Matching
- 通过渐进迭代剪枝，减少冗余更新步骤，实现近单次遍历推理。
- 在2K分辨率下，速度提升7.28倍，内存峰值降低76.6%，全局内存请求减少80.9%。
- 适合部署在边缘设备的实时高保真立体匹配，尤其适用于自动驾驶与机器人导航。
尽管迭代式立体匹配能实现高精度，但其依赖循环神经网络（RNN）限制了边缘部署，这一挑战在现有研究中尚未充分探索。我们分析了迭代优化过程，发现视差更新具有空间稀疏性和时间冗余性。首先，提出渐进迭代剪枝策略，抑制冗余更新步骤，将递归计算有效压缩为近单次遍历推理。其次，设计协同单目先验迁移框架，无需专用单目编码器即可隐式嵌入深度先验，消除其计算开销。第三，开发FlashGRU——一种面向硬件的RNN算子，利用结构化稀疏与输入输出感知设计，在2K分辨率下相较原生ConvGRU实现7.28×加速、76.6%内存峰值降低和80.9%全局内存请求减少。PipStereo在边缘硬件上实现实时高保真立体匹配：在NVIDIA Jetson Orin NX（FP16）上处理320×640图像仅需75ms，RTX 4090上仅需19ms，精度媲美大型迭代模型，泛化能力与精度远超现有实时方法。
原文摘要 · Abstract (English)
While iterative stereo matching achieves high accuracy, its dependence on Recurrent Neural Networks (RNN) hinders edge deployment, a challenge underexplored in existing researches. We analyze iterative refinement and reveal that disparity updates are spatially sparse and temporally redundant. First, we introduce a progressive iteration pruning strategy that suppresses redundant update steps, effectively collapsing the recursive computation into a near-single-pass inference. Second, we propose a collaborative monocular prior transfer framework that implicitly embeds depth priors without requiring a dedicated monocular encoder, thereby eliminating its associated computational burden. Third, we develop FlashGRU, a hardware-aware RNN operator leveraging structured sparsity and I/O-conscious design, achieving a 7.28$\times$ speedup, 76.6\% memory peak reduction and 80.9\% global memory requests reduction over natvie ConvGRUs under 2K resolution. Our PipStereo enables real-time, high-fidelity stereo matching on edge hardware: it processes 320$\times$640 frames in just 75ms on an NVIDIA Jetson Orin NX (FP16) and 19ms on RTX 4090, matching the accuracy of large iterative based models, and our generalization ability and accuracy far exceeds that of existing real-time methods. Our embedded AI projects will be updated at: https://github.com/XPENG-Aridge-AI.
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。