用脉冲神经网络实时分类微流控中高速流动的微粒，精度超98%。
Neuromorphic Imaging Flow Cytometry combined with Adaptive Recurrent Spiking Neural Networks
- 用事件相机捕捉微粒流动，脉冲网络分析时间依赖性
- 递归脉冲网络达98.4%准确率，轻量前馈网络提升4.3%性能
- 适合需要低延迟、高能效的生物医学检测场景
我们展示了一种基于1 μs时间分辨率事件型CMOS相机的实验性成像流式细胞仪，数据由自适应前馈与递归脉冲神经网络处理。研究对以0.7 m/s速度在微流控通道中流动的PMMA微粒（12、16、20 μm）进行分类。实验数据分析表明，利用时间依赖性的脉冲递归网络（包括LSTM和GRU模型）实现了98.4%的分类准确率。此外，轻量级前馈脉冲网络中的自适应机制使准确率提升了4.3%。该工作为类脑神经形态技术在生物医学应用中的发展提供了技术路线图，在保持低延迟与稀疏性的同时提升了分类性能。
原文摘要 · Abstract (English)
We present an experimental imaging flow cytometer using a 1 μs temporal resolution event-based CMOS camera, with data processed by adaptive feedforward and recurrent spiking neural networks. Our study classifies PMMA particles (12, 16, 20 μm) flowing at 0.7 m/s in a microfluidic channel. Processing of experimental data highlighted that spiking recurrent networks, including LSTM and GRU models, achieved 98.4% accuracy by leveraging temporal dependencies. Additionally, adaptation mechanisms in lightweight feedforward spiking networks improved accuracy by 4.3%. This work outlines a technological roadmap for neuromorphic-assisted biomedical applications, enhancing classification performance while maintaining low latency and sparsity.
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。