A batteryless electrochemical sensing system IC for miniaturized wearable sensor nodes is presented. A battery-powered hub node transmits the power and control signal required for the operation of sensor nodes through a human body. The sensor data acquired by sensor nodes are transmitted to the hub node, again through a body channel, using frequency modulation (FM). The intra-body power and data transfer leads to batteryless sensor node operation, and the direct-FM-based sensor interface using a current-controlled oscillator (26.8 and 44.9 kHz/$\mu\mathrm{A}$) simplifies sensor node hardware, which enables miniaturized wearable sensor nodes. By adding a proper multiplexing scheme among sensor nodes, the proposed system can be extended to a further advanced system consisting of distributed wearable sensor nodes.

Bootstrapping has been a primary tool for ensemble and uncertainty quantification in machine learning and statistics. However, due to its nature of multiple training and resampling, bootstrapping deep neural networks is computationally burdensome; hence it has difficulties in practical application to the uncertainty estimation and related tasks. To overcome this computational bottleneck, we propose a novel approach called Neural Bootstrapper (NeuBoots), which learns to generate bootstrapped neural networks through single model training. NeuBoots injects the bootstrap weights into the high-level feature layers of the backbone network and outputs the bootstrapped predictions of the target, without additional parameters and the repetitive computations from scratch. We apply NeuBoots to various machine learning tasks related to uncertainty quantification, including prediction calibrations in image classification and semantic segmentation, active learning, and detection of out-of-distribution samples. Our empirical results show that NeuBoots outperforms other bagging based methods under a much lower computational cost without losing the validity of bootstrapping.

Optical diffraction tomography measures the three-dimensional refractive index map of a specimen and visualizes biochemical phenomena at the nanoscale in a non-destructive manner. One major drawback of optical diffraction tomography is poor axial resolution due to limited access to the three-dimensional optical transfer function. This missing cone problem has been addressed through regularization algorithms that use a priori information, such as non-negativity and sample smoothness. However, the iterative nature of these algorithms and their parameter dependency make real-time visualization impossible. In this article, we propose and experimentally demonstrate a deep neural network, which we term DeepRegularizer, that rapidly improves the resolution of a three-dimensional refractive index map. Trained with pairs of datasets (a raw refractive index tomogram and a resolution-enhanced refractive index tomogram via the iterative total variation algorithm), the three-dimensional U-net-based convolutional neural network learns a transformation between the two tomogram domains. The feasibility and generalizability of our network are demonstrated using bacterial cells and a human leukaemic cell line, and by validating the model across different samples. DeepRegularizer offers more than an order of magnitude faster regularization performance compared to the conventional iterative method. We envision that the proposed data-driven approach can bypass the high time complexity of various image reconstructions in other imaging modalities.

In this paper, the feasibility of the intra-body power transfer is verified and an area-efficient rectifier with threshold voltage cancellation for intra-body power transfer is proposed. Through the measurement, the optimum power transmission frequency is determined as 1 MHz. Since the transmission voltage gain is sufficiently high (about 0.3) and nearly constant over a wide range of the load resistance, the intra-body power transfer works well even with the nonlinear and varying input impedance of the rectifier that is placed right after the receiving electrode. The proposed area-efficient rectifier cancels the threshold voltage by using a diode-connected device with a dynamically biased current source. It achieves a considerably smaller area of 65 μm × 30 μm than the area consumed by other designs reported previously. The overall power conversion efficiency (PCE) and voltage conversion ratio (VCR) are also improved. It achieves 84.4% of PCE and 85.6% of VCR with 1-V input voltage and 40-kΩ load.