Category: IEEE Transactions on Biomedical Engineering
Frontcover
Presents the front cover for this issue of the publication.
Bayesian Inference Identifies Combination Therapeutic Targets in Breast Cancer
Dominant-Current Deep Learning Scheme for Electrical Impedance Tomography
Liver Extraction Using Residual Convolution Neural Networks From Low-Dose CT Images
An efficient and precise liver extraction from computed tomography (CT) images is a crucial step for computer-aided hepatic diseases diagnosis and treatment. Considering the possible risk to patient's health due to X-ray radiation of repetitive CT examination, low-dose CT (LDCT) is an effective solution for medical imaging. However, inhomogeneous appearances and indistinct boundaries due to additional noise and streaks artifacts in LDCT images often make it a challenging task. This study aims to extract a liver model from LDCT images for facilitating medical expert in surgical planning and post-operative assessment along with low radiation risk to the patient. Our method carried out liver extraction by employing residual convolutional neural networks (LER-CN), which is further refined by noise removal and structure preservation components. After patch-based training, our LER-CN shows a competitive performance relative to state-of-the-art methods for both clinical and publicly available MICCAI Sliver07 datasets. We have proposed training and learning algorithms for LER-CN based on back propagation gradient descent. We have evaluated our method on 150 abdominal CT scans for liver extraction. LER-CN achieves dice similarity coefficient up to 96.5$pm text{1.8}%$ , decreased volumetric overlap error up to 4.30$pm text{0.58}%$ , and average symmetric surface distance less than 1.4 $pm text{0.5mm}$ . These findings have shown that LER-CN is a favorable method for medical applications with high efficiency allowing low radiation risk to patients.
Stimulation and Artifact-Suppression Techniques for <italic>In Vitro</italic> High-Density Microelectrode Array Systems
We present novel voltage stimulation buffers with controlled output current, along with recording circuits featuring adjustable high-pass cut-off filtering to perform efficient stimulation while actively suppressing stimulation artifacts in high-density microelectrode arrays. Owing to the dense packing and close proximity of the electrodes in such systems, a stimulation through one electrode can cause large electrical artifacts on neighboring electrodes that easily saturate the corresponding recording amplifiers. To suppress such artifacts, the high-pass corner frequencies of all available 2048 recording channels can be raised from several Hz to several kHz by applying a “soft-reset” or pole-shifting technique. With the implemented artifact suppression technique, the saturation time of the recording circuits, connected to electrodes in immediate vicinity to the stimulation site, could be reduced to less than 150 μs. For the stimulation buffer, we developed a circuit, which can operate in two modes: either control of only the stimulation voltage or control of current and voltage during stimulation. The voltage-only controlled mode employs a local common-mode feedback operational transconductance amplifier with a near rail-to-rail input/output range, suitable for driving high-capacitive loads. The current/voltage controlled mode is based on a positive current conveyor generating adjustable output currents, whereas its upper and lower output voltages are limited by two feedback loops. The current/voltage controlled circuit can generate stimulation pulses up to 30 μA with less than ±0.1% linearity error in the low-current mode and up to 300 μA with less than ±0.2% linearity error in the high-current mode.