Human Activity Prediction Based on Forecasted IMU Activity Signals by Sequence-to-Sequence Deep Neural Networks.

Human Activity Recognition (HAR) has gained significant attention due to its broad range of applications, such as healthcare, industrial work safety, activity assistance, and driver monitoring. Most prior HAR systems are based on recorded sensor data (i.e., past information) recognizing human activities. In fact, HAR works based on future sensor data to predict human activities are rare. Human Activity Prediction (HAP) can benefit in multiple applications, such as fall detection or exercise routines, to prevent injuries. This work presents a novel HAP system based on forecasted activity data of Inertial Measurement Units (IMU). Our HAP system consists of a deep learning forecaster of IMU activity signals and a deep learning classifier to recognize future activities. Our deep learning forecaster model is based on a Sequence-to-Sequence structure with attention and positional encoding layers. Then, a pre-trained deep learning Bi-LSTM classifier is used to classify future activities based on the forecasted IMU data. We have tested our HAP system for five daily activities with two tri-axial IMU sensors. The forecasted signals show an average correlation of 91.6% to the actual measured signals of the five activities. The proposed HAP system achieves an average accuracy of 97.96% in predicting future activities.

Real-Time Human Activity Recognition with IMU and Encoder Sensors in Wearable Exoskeleton Robot via Deep Learning Networks.

Wearable exoskeleton robots have become a promising technology for supporting human motions in multiple tasks. Activity recognition in real-time provides useful information to enhance the robot's control assistance for daily tasks. This work implements a real-time activity recognition system based on the activity signals of an inertial measurement unit (IMU) and a pair of rotary encoders integrated into the exoskeleton robot. Five deep learning models have been trained and evaluated for activity recognition. As a result, a subset of optimized deep learning models was transferred to an edge device for real-time evaluation in a continuous action environment using eight common human tasks: stand, bend, crouch, walk, sit-down, sit-up, and ascend and descend stairs. These eight robot wearer's activities are recognized with an average accuracy of 97.35% in real-time tests, with an inference time under 10 ms and an overall latency of 0.506 s per recognition using the selected edge device.

Oxide Thin-Film Transistor-Based Vertically Stacked Complementary Inverter for Logic and Photo-Sensor Operations.

Numerous studies have addressed the utilization of oxide thin-film transistor (TFT)-based complementary logic circuits that are based on two-dimensional (2D) planar structures. However, there are fundamental limits to the 2D planar structured complementary logic circuits, such as a large dimension and a large parasitic resistance. This work demonstrated a vertically stacked three-dimensional complementary inverter composed of a p-channel tin monoxide (SnO) TFT and an n-channel indium-gallium-zinc oxide (IGZO) TFT. A bottom-gate p-channel SnO TFT was formed on the top-gate n-channel IGZO TFT with a shared common gate electrode. The fabricated vertically stacked complementary inverter exhibited full swing characteristics with a voltage gain of ~33.6, a high noise margin of 3.13 V, and a low noise margin of 3.16 V at a supplied voltage of 10 V. The achieved voltage gain of the fabricated complementary inverter was higher than that of the vertically stacked complementary inverters composed of other oxide TFTs in previous works. In addition, we showed that the vertically stacked complementary inverter exhibited excellent visible-light photoresponse. This indicates that the oxide TFT-based vertically stacked complementary inverter can be used as a sensitive photo-sensor operating in the visible spectral range with the voltage read-out scheme.

Decreased muscle mass in Korean subjects with intracranial arterial stenosis: The Kangbuk Samsung Health Study.

BACKGROUNDS AND AIMS: Intracranial arterial stenosis (ICAS) is a common cause of ischemic stroke in Asians. Decreased muscle mass is one of the major causes of chronic disease in adults. The purpose of this study was to analyze the relationship between muscle mass and ICAS in Korean adults. METHODS: For this study, we selected a total of 10,530 participants (mean age, 43.3 years; 8558 men) in a health screening program, for whom transcranial Doppler (TCD) ultrasound was used to detect >50% ICAS based on criteria modified from the stroke outcomes and neuroimaging of intracranial atherosclerosis trial. Body composition was evaluated by bioelectrical impedance analysis (BIA). Skeletal muscle index (SMI) was calculated with muscle mass/weight (kg) * 100. RESULTS: Among the total patient population, 322 (3.1%) subjects had ICAS. Subjects with ICAS were older, and had higher mean values for fasting glucose, body mass index and blood pressure compared with those without ICAS. Subjects with ICAS had significantly lower muscle mass, SMI and higher percent body fat compared with those without ICAS. In logistic regression analysis, the subjects in the highest tertile of muscle mass had the lowest odds ratio for ICAS with the lowest tertile group of muscle mass as the reference group even after adjusting for age, systolic blood pressure, fasting blood glucose, sex, smoking and exercise (OR 0.650, 95% CI 0.442-0.955). CONCLUSIONS: Subjects with ICAS had significantly decreased muscle mass compared with those without ICAS in Korean adults. The risk for ICAS was lower in subjects with higher muscle mass.

Associations between amount of smoking and alcohol intake and risk of colorectal neoplasm.

BACKGROUND AND AIM: Although smoking and alcohol has been linked to an increased risk of colorectal neoplasm (CRN), large-scale studies to identify dose-dependent relationship between amount of smoking and alcohol consumption and risk of CRN are rare. We aimed to investigate the risk for CRN according to the amount of smoking and alcohol intake in a large sample of Korean adults. METHODS: A cross-sectional study was performed on 31,714 examinees aged ≥30 years undergoing their first colonoscopy as part of routine preventive health care between 2010 and 2011. Never smokers were compared with six groups of smokers according to smoking amount, and individuals with alcohol intake of ≤ 6.25 g ethanol per day were compared with three groups according to alcohol amount. RESULTS: In adjusted models, the risk of overall CRN increased with increasing amount of smoking (P for trend < 0.001). The adjusted odds ratios for overall CRN comparing never smokers with six smoker groups according to smoking amount (≤2.50, 2.51-5.60, 5.61-9.00, 9.01-13.00, 13.01-19.50, and ≥19.51 pack-years) were 1.02, 1.19, 1.35, 1.53, 1.63, and 2.03, respectively. In addition, the risk of both non-advanced and advanced CRN increased with increasing amount of smoking (both P for trend < 0.001). However, the amount of alcohol consumption was not correlated with the risk of CRN. CONCLUSIONS: The prevalence of CRN was associated with increasing amount of smoking in a dose-response manner, whereas it was not associated with the amount of alcohol consumption. Our study suggests that smoking amount as well as smoking status should be considered for CRN risk stratification.