Unsupervised domain adaption (UDA), which aims to enhance the segmentation performance of deep models on unlabeled data, has recently drawn much attention. In this paper, we propose a novel UDA method (namely DLaST) for medical image segmentation via disentanglement learning and self-training. Disentanglement learning factorizes an image into domain-invariant anatomy and domain-specific modality components. To make the best of disentanglement learning, we propose a novel shape constraint to boost the adaptation performance. The self-training strategy further adaptively improves the segmentation performance of the model for the target domain through adversarial learning and pseudo label, which implicitly facilitates feature alignment in the anatomy space. Experimental results demonstrate that the proposed method outperforms the state-of-the-art UDA methods for medical image segmentation on three public datasets, i.e., a cardiac dataset, an abdominal dataset and a brain dataset. The code will be released soon.

It has proved that the auscultation of respiratory sound has advantage in early respiratory diagnosis. Various methods have been raised to perform automatic respiratory sound analysis to reduce subjective diagnosis and physicians' workload. However, these methods highly rely on the quality of respiratory sound database. In this work, we have developed the first open-access paediatric respiratory sound database, SPRSound. The database consists of 2,683 records and 9,089 respiratory sound events from 292 participants. Accurate label is important to achieve a good prediction for adventitious respiratory sound classification problem. A custom-made sound label annotation software (SoundAnn) has been developed to perform sound editing, sound annotation, and quality assurance evaluation. A team of 11 experienced paediatric physicians is involved in the entire process to establish golden standard reference for the dataset. To verify the robustness and accuracy of the classification model, we have investigated the effects of different feature extraction methods and machine learning classifiers on the classification performance of our dataset. As such, we have achieved a score of 75.22%, 61.57%, 56.71%, and 37.84% for the four different classification challenges at the event level and record level.

Vigilance decrement in driving tasks has been reported to be a major factor in fatal accidents and could severely endanger public transportation safety. However, efficient approaches for estimating vigilance in real driving environment are still lacking. In this paper, we propose a novel approach for implementing continuous vigilance estimation using forehead electrooculograms (EOGs) acquired by wearable dry electrodes in both simulated and real driving environments. To improve the feasibility of this approach for real-world applications, a forehead EOG-based electrode placement with only four electrodes is designed. Flexible dry electrodes and an acquisition board are integrated as a wearable device for recording EOGs. Twenty and ten subjects participated in the simulated and real-world driving environment experiments, respectively. Accurate eye movement parameters from eye-tracking glasses are extracted to calculate the PERCLOS index for vigilance annotation. This is because the vigilance state is a temporally dynamic process, and a continuous conditional random field and a continuous conditional neural field are introduced to construct more accurate vigilance estimation models. To evaluate the efficiency of our system, systematic experiments are performed in real scenarios under various illumination and weather conditions following laboratory simulations as preliminary studies. The experimental results demonstrate that the wearable dry electrode prototype, which has a relatively comfortable forehead setup, can efficiently capture vigilance dynamics. The best mean correlation coefficients achieved by our proposed approach are 71.18% and 66.20% in laboratory simulations and real-world driving environments, respectively. The cross-environment experiments are performed to evaluate the simulated-to-real generalization and a best mean correlation coefficient of 53.96% is achieved.

Electrocardiogram (ECG) is one of the major physiological vital signs and an effective monitoring method for patients with cardiovascular diseases. However, existing ECG recordings require a galvanic body contact, which is unpractical in daily life. This paper presents the design of an ECG chip that facilitates non-contact ECG recording through capacitive coupling. With the input impedance boosting techniques, as well as an active driven-right-leg (DRL) which boosts common-mode rejection ratio (CMRR) to 70 dB, the single-ended capacitive feedback active electrode (AE) achieves ultra-high input impedance of 400 GΩ (< 0.5 Hz), a large common-mode (CM) interference tolerance (2.8 Vpp), and a high linear-input-range (220 mVpp). Implemented in 0.18 μm 5V CMOS process, the prototype occupies an area of 1.23 mm2, and consumes 18 μA and 13 μA for the AE and DRL respectively. Real life non-contact capacitively coupled ECG acquisition has been demonstrated, obtaining ECG waves and heart rate in the presence of motion artifacts as well as ambient interference.

Embedded arrhythmia classification is the first step towards heart diseases prevention in wearable applications. In this paper, a robust arrhythmia classification algorithm, NEO-CCNN, for wearables that can be implemented on a simple microcontroller is proposed. The NEO-CCNN algorithm not only detects QRS complex but also accurately locates R-peak with the help of the proposed adaptive time-dependent thresholding technique, improving the accuracy and sensitivity in arrhythmia classification. An optimized compact 1D-CNN network (CCNN) with 9,701 parameters is used for classification. A QRS complex augmentation method is introduced in the training process to cater for R-peak location error (RLE). A nested k ₁ k ₂ -fold cross-validation method is utilized to evaluate the robustness of the proposed algorithm. Simulation results show that the proposed algorithm has the ability to detect more than 99.79% of R peaks with an RLE of 7.94 ms for the MIT-BIH database. Implemented on the STM32F407 microcontroller, NEO-CNN attains a classification accuracy of 97.83% and sensitivity of 96.46% using only 8s window size.

An ultra-low power level-crossing analog-to-digital converter (LC-ADC) with on-chip adaptive sampling is presented. Different from conventional ADCs based on Nyquist sampling, LC-ADC utilizes sparsity of signals for low power data acquisition. To save power, the proposed adaptive sampling scheme is implemented with only one scaler and one high-precision comparator, which is in sharp contrast to conventional LC-ADCs that require an n-bit DAC and two comparators. Implemented in 0.18 um CMOS process, the proposed ADC consumes only 61 nW under 0.5V supply, and achieves 5.6 bits equivalent numbers of bits (ENOB) and 35 dB signal to noise and distortion ratio (SNDR) with operating frequency up to 1 kHz.