Segmenting lung lesions of COVID-19 from CT images via pyramid pooling improved Unet.

Segmenting lesion regions of Coronavirus Disease 2019 (COVID-19) from computed tomography (CT) images is a challenge owing to COVID-19 lesions characterized by high variation, low contrast between infection lesions and around normal tissues, and blurred boundaries of infections. Moreover, a shortage of available CT dataset hinders deep learning techniques applying to tackling COVID-19. To address these issues, we propose a deep learning-based approach known as PPM-Unet to segmenting COVID-19 lesions from CT images. Our method improves an Unet by adopting pyramid pooling modules instead of the conventional skip connection and then enhances the representation of the neural network by aiding the global attention mechanism. We first pre-train PPM-Unet on COVID-19 dataset of pseudo labels containing1600 samples producing a coarse model. Then we fine-tune the coarse PPM-Unet on the standard COVID-19 dataset consisting of 100 pairs of samples to achieve a fine PPM-Unet. Qualitative and quantitative results demonstrate that our method can accurately segment COVID-19 infection regions from CT images, and achieve higher performance than other state-of-the-art segmentation models in this study. It offers a promising tool to lay a foundation for quantitatively detecting COVID-19 lesions.

Prediction of Freezing of Gait in Parkinson's Disease Using a Random Forest Model Based on an Orthogonal Experimental Design: A Pilot Study.

PURPOSE: The purpose of this study was to introduce an orthogonal experimental design (OED) to improve the efficiency of building and optimizing models for freezing of gait (FOG) prediction. METHODS: A random forest (RF) model was developed to predict FOG by using acceleration signals and angular velocity signals to recognize possible precursor signs of FOG (preFOG). An OED was introduced to optimize the feature extraction parameters. RESULTS: The main effects and interaction among the feature extraction hyperparameters were analyzed. The false-positive rate, hit rate, and mean prediction time (MPT) were 27%, 68%, and 2.99 s, respectively. CONCLUSION: The OED was an effective method for analyzing the main effects and interactions among the feature extraction parameters. It was also beneficial for optimizing the feature extraction parameters of the FOG prediction model.

Experiment and hydro-mechanical coupling simulation study on the human periodontal ligament.

In this paper, a new method involving an experiment in vivo and hydro-mechanical coupling simulations was proposed to investigate the biomechanical property of human periodontal ligament (PDL). Teeth were loaded and their displacements were measured in vivo. The finite element model of the experiment was built and hydro-mechanical coupling simulations were conducted to test some PDL's constitutive models. In the simulations, the linear elastic model, the hyperfoam model, and the Ogden model were assumed for the solid phase of the PDL coupled with a model of the fluid phase of the PDL. The displacements of the teeth derived from the simulations were compared with the experimental data to validate these constitutive models. The study shows that a proposed constitutive model of the PDL can be reliably tested by this method. Furthermore, the influence of species, areas, and the fluid volume ratio on PDL's mechanical property should be considered in the modeling and simulation of the mechanical property of the PDL.

Pullulan production and physiological characteristics of Aureobasidium pullulans under acid stress.

In this study, batch processes of pullulan production by Aureobasidium pullulans CCTCC M 2012259 under different pH environments were evaluated. The pH of the medium decreased quickly to an acid stress condition under batch fermentation without pH control. A higher pullulan production was always obtained with a lower biomass under a given glucose concentration with constant pH control, and vice versa. Based on the nonlinear regression analysis of the results obtained from diverse pH control modes, a constant controlled pH of 3.8 was predicted as an optimum pH for efficient pullulan production using a one-element cubic equation. A maximum pullulan concentration of 26.8 g/L and a minimum biomass of 8.1 g/L were achieved under the optimal pH of 3.8, which were in good agreement with the results predicted by the mathematical model. Further information on the physiological characteristics of A. pullulans CCTCC M 2012259 such as intracellular pH, NADH/NAD(+), ATP/ADP, and glutathione generation under moderate or severe acidic conditions were investigated, and the results presented more evidence on why pullulan biosynthesized with high efficiency under moderate acid stress (e.g., pH 3.8), which would also help us to better understand the response of the cells to acid stress.

Media optimization for elevated molecular weight and mass production of pigment-free pullulan.

In this study, an Aureobasidium pullulans SZU 1001 mutant, deficient in pigment production, was screened by complex UV and γ-ray mutagenesis. Medium composition optimization for increased pullulan molecular weight and production was conducted using this mutant. Six nutrients: yeast extract, (NH4)2SO4, K2HPO4, NaCl, MgSO4·7H2O and CaCl2 were detected within pullulan production in flasks. It is shown that NaCl and K2HPO4 have significant influences on molecular weight of pullulan, while yeast extract and (NH4)2SO4 significantly affect pullulan yield. To achieve a higher molecular weight and more efficient pullulan production, a response surface methodology approach was introduced to predict an optimal nutrient combination. A molecular weight of 5.74 × 10(6) and pullulan yield on glucose of 51.30% were obtained under batch pullulan fermentation with the optimized media, which increased molecular weight and pullulan production by 97.25% and 11.04%, respectively compared with the control media.