Proactively detecting falls and preventing injuries are among the primary keys to a healthy life for the elderly. Near-fall remote monitoring in daily life could provide key information to prevent ...future falls and obtain quantitative rehabilitation status for patients with weak balance ability. In this study, we developed a deep learning-based novel classification algorithm to precisely categorize three classes (falls, near-falls, and activities of daily living (ADLs)) using a single inertial measurement unit (IMU) device attached to the waist. A total of 34 young participants were included in this study. An IMU containing accelerometer and gyroscope sensors was fabricated to acquire acceleration and angular velocity signals. A comprehensive experiment including thirty-six types of activities (10 types of falls, 10 types of near-falls, and 16 types of ADLs) was designed based on previous studies. A modified directed acyclic graph-convolution neural network (DAG-CNN) architecture with hyperparameter optimization was proposed to predict fall, near-fall, and ADLs. Prediction results of the modified DAG-CNN structure were found to be approximately 7% more accurate than the traditional CNN structure. For the case of near-falls, the modified DAG-CNN demonstrated excellent prediction performance with accuracy of over 98% by combining gyroscope and accelerometer features. Additionally, by combining acceleration and angular velocity the trained model showed better performance than each model of acceleration and angular velocity. It is believed that information to preemptively handle the risk of falls and quantitatively evaluate the rehabilitation status of the elderly with weak balance will be provided by monitoring near-falls.
In recent years, convolutional neural networks (CNNs) have been used as an alternative to recurrent neural networks (RNNs) in text processing with promising results. In this paper, we investigated ...the newly introduced capsule networks (CapsNets), which are getting a lot of attention due to their great performance gains on image analysis more than CNNs, for sentence classification or sentiment analysis in some cases. The results of our experiment show that the proposed well-tuned CapsNet model can be a good, sometimes better and cheaper, substitute of models based on CNNs and RNNs used in sentence classification. In order to investigate whether CapsNets can learn the sequential order of words or not, we performed a number of experiments by reshuffling the test data. Our CapsNet model shows an overall better classification performance and better resistance to adversarial attacks than CNN and RNN models.
Sampling-based PLLs have become a new research trend due to the possibility of removing the frequency divider (FDIV) from the feedback path, where the FDIV increases the contribution of in-band noise ...by the factor of dividing ratio square (N2). Between two possible sampling methods, sub-sampling and reference-sampling, the latter provides a relatively wide locking range, as the slower input reference signal is sampled with the faster VCO output signal. However, removal of FDIV makes the PLL not feasible to implement fractional-N operation based on varying divider ratios through random sequence generators, such as a Delta-Sigma-Modulator (DSM). To address the above design challenges, we propose a reference-sampling-based calibration-free fractional-N PLL (RSFPLL) with a phase-interpolator-linked sampling clock generator (PSCG). The proposed RSFPLL achieves fractional-N operations through phase-interpolator (PI)-based multi-phase generation instead of a typical frequency divider or digital-to-time converter (DTC). In addition, to alleviate the power burden arising from VCO-rated sampling, a flexible mask window generation method has been used that only passes a few sampling clocks near the point of interest. The prototype PLL system is designed with a 65 nm CMOS process with a chip size of 0.42 mm2. It achieves 322 fs rms jitter, −240.7 dB figure-of-merit (FoM), and −44.06 dBc fractional spurs with 8.17 mW power consumption.
Pre-impact fall detection can detect a fall before a body segment hits the ground. When it is integrated with a protective system, it can directly prevent an injury due to hitting the ground. An ...impact acceleration peak magnitude is one of key measurement factors that can affect the severity of an injury. It can be used as a design parameter for wearable protective devices to prevent injuries. In our study, a novel method is proposed to predict an impact acceleration magnitude after loss of balance using a single inertial measurement unit (IMU) sensor and a sequential-based deep learning model. Twenty-four healthy participants participated in this study for fall experiments. Each participant worn a single IMU sensor on the waist to collect tri-axial accelerometer and angular velocity data. A deep learning method, bi-directional long short-term memory (LSTM) regression, is applied to predict a fall's impact acceleration magnitude prior to fall impact (a fall in five directions). To improve prediction performance, a data augmentation technique with increment of dataset is applied. Our proposed model showed a mean absolute percentage error (MAPE) of 6.69 ± 0.33% with r value of 0.93 when all three different types of data augmentation techniques are applied. Additionally, there was a significant reduction of MAPE by 45.2% when the number of training datasets was increased by 4-fold. These results show that impact acceleration magnitude can be used as an activation parameter for fall prevention such as in a wearable airbag system by optimizing deployment process to minimize fall injury in real time.
We investigated the usefulness of quantitative 99mTc-white blood cell (WBC) single photon emission computed tomography (SPECT)/computed tomography (CT) for predicting lower extremity amputation in ...diabetic foot infection (DFI). A total of 93 feet of 83 consecutive patients with DFI who underwent WBC SPECT/CT for treatment planning were retrospectively analysed. The clinical and SPECT/CT parameters were collected along with the measurements of the maximum standardized uptake value (SUVmax) at DFI. Statistical logistic regression analysis was performed to explore the predictors of LEA and receiver operating characteristic (ROC) curve was analysed to assess the predictive value of SPECT/CT. The independent predictors of amputation were previous amputation (OR 11.9), numbers of SPECT/CT lesions (OR 2.1), and SUVmax of DFI; either continuous SUVmax (1-increase) (OR 1.3) or categorical SUVmax > 1.1 (OR 21.6). However, the conventional SPECT/CT interpretation failed to predict amputation. In ROC analysis, the SUVmax yielded a fair predictor (area under the curve (AUC) 0.782) of amputation. The model developed from these independent predictors yielded an excellent performance for predicting amputation (AUC 0.873). Quantitative WBC SPECT/CT can provide new information useful for predicting the outcomes and guiding treatment for patients with DFI.
The study aimed to elucidate simple and effective risk factors for scabies infection in the medical environment for early detection and prevention of exposure to other patients and medical staff. We ...conducted a case-control study of patients who were diagnosed with scabies among hospitalized patients between 2008 and 2019 in an acute-care teaching hospital. Each case was matched according to sex and age with two randomly selected controls without scabies during hospitalization. Clinical characteristics of cases at the time of hospital admission were compared to those of the control group. The scabies group included 34 patients and 68 patients who were included in the control group. After adjusting for confounding factors, previous long-term care facility admission was only associated with scabies (adjusted odds ratio, 5.44; 95% confidence interval, 1.46-20.27;
= 0.012). Careful examination, particularly for patients with previous long-term care facility admission, might be useful for the early detection of scabies.
As a developing technology for flexible electronic device fabrication, ultra-violet (UV) photodetectors (PDs) based on a ZnO nanostructure are an effective approach for large-area integration of ...sensors on nonconventional substrates, such as plastic or paper. However, photoconductive ZnO nanorods grown on flexible substrates have slow responses or recovery as well as low spectral responsivity
because of the native defects and inferior crystallinity of hydrothermally grown ZnO nanorods at low temperatures. In this study, ZnO nanorod crystallites are doped with Cu or Ni/Cu when grown on polyethylene terephthalate (PET) substrates in an attempt to improve the performance of flexible PDs. The doping with Ni/Cu or Cu not only improves the crystalline quality but also significantly suppresses the density of deep-level emission defects in as-grown ZnO nanorods, as demonstrated by X-ray diffraction and photoluminescence. Furthermore, the X-ray photoelectron spectroscopy analysis shows that doping with the transition metals significantly increases the oxygen bonding with metal ions with enhanced O/Zn stoichiometry in as-grown nanorods. The fabricated flexible PD devices based on an interdigitated electrode structure demonstrates a very high
of ~123 A/W, a high on-off current ratio of ~130, and a significant improvement in transient response speed exhibiting rise and fall time of ~8 and ~3 s, respectively, by using the ZnO nanorods codoped by Ni/Cu.
Maintaining and monitoring the quality of eggs is a major concern during cold chain storage and transportation due to the variation of external environments, such as temperature or humidity. In this ...study, we proposed a deep learning-based Haugh unit (HU) prediction model which is a universal parameter to determine egg freshness using a non-destructively measured weight loss by transfer learning technique. The temperature and weight loss of eggs from a laboratory and real-time cold chain environment conditions are collected from ten different types of room temperature conditions. The data augmentation technique is applied to increase the number of the collected dataset. The convolutional neural network (CNN) and long short-term memory (LSTM) algorithm are stacked to make one deep learning model with hyperparameter optimization to increase HU value prediction performance. In addition, the general machine learning algorithms are applied to compare HU prediction results with the CNN-LSTM model. The source and target model for stacked CNN-LSTM used temperature and weight loss data, respectively. Predicting HU using only weight loss data, the target transfer learning CNN-LSTM showed RMSE value decreased from 6.62 to 2.02 compared to a random forest regressor, respectively. In addition, the MAE of HU prediction results for the target model decreased when the data augmentation technique was applied from 3.16 to 1.39. It is believed that monitoring egg freshness by predicting HU in a real-time cold chain environment can be implemented in real-life by using non-destructive weight loss parameters along with deep learning.
Background Thoracic actinomycosis is an uncommon, chronic, and progressive infection, especially in patients with HIV. We report a case of thoracic actinomycosis presenting as an isolated pleural ...effusion in a patient with an HIV infection. Case presentation A 68-year-old patient with progressive dyspnea and fever was admitted. On the right side, an ipsilateral massive pleural effusion was confirmed on the chest radiograph, and an HIV infection was newly diagnosed. A pleural biopsy was performed for the further differential diagnosis of potential opportunistic infections and malignancies. The pathology findings were consistent with actinomycosis. Conclusions Active diagnostic approaches such as a pleural biopsy should be considered for indeterminate pleural effusions in immunocompromised patients. Keywords: Human immunodeficiency virus, Actinomycosis, Pleural effusion