BLINK
BLINK
复制标题
眨
DOI：
10.1145/3370748.3406552
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Chen Z
中科院分区：
文献类型：
--
作者：
Chen Z
微型荧光钙成像显微镜广泛应用于在体监测自由行为动物体内大量神经元的活动，传统的钙图像分析方法通过迭代和批量图像处理来提取钙的踪迹，难以满足神经反馈装置对功耗和延迟的要求.在本文中，我们提出了基于位稀疏长短期记忆（LSTM）推理内核（BLINK）的钙图像处理流水线，用于有效的钙痕量提取。它大大降低了功耗和延迟，同时保持跟踪提取的准确性。我们在Ultra 96平台上实现了定制的流水线。它可以在单个FPGA设备上以亚毫秒的延迟从多达1024个细胞中提取钙痕量。我们采用28 nm工艺设计了BLINK电路。评估表明，提出的比特稀疏表示可以减少38.7%的电路面积，节省38.4%的功耗，而不损失精度。BLINK电路实现了410 pJ/推理，与高性能CPU和GPU的评估相比，能效分别提高了6293倍和52.4倍。
Miniaturized fluorescent calcium imaging microscopes are widely used for monitoring the activity of a large population of neurons in freely behaving animalsin vivo.Conventional calcium image analyses extract calcium traces by iterative and bulk image processing and they are hard to meet the power and latency requirements for neurofeedback devices. In this paper, we propose the calcium image processing pipeline based on a bit-sparse long short-term memory (LSTM) inference kernel (BLINK) for efficient calcium trace extraction. It largely reduces the power and latency while remaining the trace extraction accuracy. We implemented the customized pipeline on the Ultra96 platform. It can extract calcium traces from up to 1024 cells with sub-ms latency on a single FPGA device. We designed the BLINK circuits in a 28-nm technology. Evaluation shows that the proposed bit-sparse representation can reduce the circuit area by 38.7% and save the power consumption by 38.4% without accuracy loss. The BLINK circuits achieve 410 pJ/inference, which has 6293x and 52.4x gains in energy efficiency compared to the evaluation on the high performance CPU and GPU, respectively.