Energy-efficient high-fidelity image reconstruction with memristor arrays for medical diagnosis

• Published in: Nature communications Vol. 14; no. 1; pp. 2276 - 10
• Main Authors: Zhao, Han, Liu, Zhengwu, Tang, Jianshi, Gao, Bin, Qin, Qi, Li, Jiaming, Zhou, Ying, Yao, Peng, Xi, Yue, Lin, Yudeng, Qian, He, Wu, Huaqiang
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 20.04.2023; Nature Publishing Group; Nature Portfolio
• Subjects: 59/57; 639/166/985; 639/166/987; 639/925/927/1007; Accuracy; Algorithms; Arrays; Artificial neural networks; Computed tomography; Computer Graphics; Data processing; Diagnosis; Energy efficiency; Fourier transforms; Graphics processing units; Humanities and Social Sciences; Image processing; Image Processing, Computer-Assisted - methods; Image reconstruction; Image segmentation; Magnetic resonance imaging; Mapping; Medical diagnosis; Medical imaging; Memristors; multidisciplinary; Neural networks; Science; Science (multidisciplinary); Software; Tomography, X-Ray Computed
• ISSN: 2041-1723; 2041-1723
• DOI: 10.1038/s41467-023-38021-7

Medical imaging is an important tool for accurate medical diagnosis, while state-of-the-art image reconstruction algorithms raise critical challenges in massive data processing for high-speed and high-quality imaging. Here, we present a memristive image reconstructor (MIR) to greatly accelerate image reconstruction with discrete Fourier transformation (DFT) by computing-in-memory (CIM) with memristor arrays. A high-accuracy quasi-analogue mapping (QAM) method and generic complex matrix transfer (CMT) scheme was proposed to improve the mapping precision and transfer efficiency, respectively. High-fidelity magnetic resonance imaging (MRI) and computed tomography (CT) image reconstructions were demonstrated, achieving software-equivalent qualities and DICE scores after segmentation with nnU-Net algorithm. Remarkably, our MIR exhibited 153× and 79× improvements in energy efficiency and normalized image reconstruction speed, respectively, compared to graphics processing unit (GPU). This work demonstrates MIR as a promising high-fidelity image reconstruction platform for future medical diagnosis, and also largely extends the application of memristor-based CIM beyond artificial neural networks. Image reconstruction algorithms raise critical challenges in massive data processing for medical diagnosis. Here, the authors propose a solution to significantly accelerate medical image reconstruction on memristor arrays, showing 79× faster speed and 153× higher energy efficiency than state-of-the-art graphics processing unit.