Logic gate networks enable low power ECG arrhythmia classification
Logic Gate Networks and Lookup Table Networks as Lightweight Hardware Classifiers for Inter-patient ECG Arrhythmia Classification
Machine LearningArtificial Intelligence
Summary
Classifying irregular heartbeats from ECG signals usually requires heavy computation, which consumes a lot of power. The authors studied two lightweight kinds of neural networks based on simple logic gates and lookup tables that use very little energy. They tested these networks on a common ECG dataset and found they achieved accuracy comparable to more complex methods while using thousands of times less computation and energy. They also built a hardware prototype that showed the approach works efficiently for real devices. This could help in building battery-friendly heart monitors.
What this means in practice
- •For medical device engineers: Develop ultra-low power ECG monitors that classify arrhythmias in real time using lightweight hardware classifiers.$Commercial implications: Enables design of energy-efficient wearable heart monitors that extend battery life and maintain classification accuracy, suitable for commercial health gadgets.
- •For embedded systems developers: Implement minimal-computation classifiers on resource-constrained hardware for real-time signal processing tasks like ECG arrhythmia detection.
Authors
Wout Mommen, Lars Keuninckx, Siddharth Patil, Paul Detterer, Achiel Colpaert, Piet Wambacq
Abstract
Deep Differentiable Logic Gate Networks (LGNs) and Lookup Table Networks (LUTNs) offer a promising approach for very low power inference due to their use of simple binary logic operations instead of arithmetic. In this work, we generalize the logic gates of LGNs to more than two input pins, naturally arriving at networks consisting of $N$-input LUTs. To obtain a differentiable expression for training the $N$-LUT entries, we adopt the Boolean equation of a $2^N$:1 multiplexer (MUX) and optimize its input parameters during training. We investigate the applicability of LGNs and LUTNs to inter-patient ECG arrhythmia classification using the MIT-BIH data set. The proposed models achieve up to 94.41\% accuracy and a $jκ$ index of 0.683 on a four-class task, showing a competitive performance compared to existing CNN-, SVM- and SNN-based methods. Our LGNs and LUTNs only require an estimated 2.89k to 6.17k FLOPs, including preprocessing and readout, which is three to six orders of magnitude less than state-of-the-art methods. We verified our design, which consists of the preprocessing pipeline and a 6-LUTN classifier, by implementing it on a Xilinx Zynq-7000 ZedBoard. The complete system consumes a dynamic energy of 8.25 $μ$J/inference, of which only 0.46 nJ is utilized by the LUTN classifier. These results show that both LGNs and LUTNs can be employed as lightweight hardware-based classifiers for inter-patient ECG arrhythmia classification.