Point cloud registration is a fundamental problem in 3D computer vision. Outdoor LiDAR point clouds are typically large-scale and complexly distributed, which makes the challenging. In this paper, we propose an efficient hierarchical network named HRegNet for out-door registration. Instead of using all points clouds, performs on hierarchically extracted keypoints descriptors. The overall framework combines reliable features deeper layer precise position information shallower layers to...

Abstract This study investigates the influence of curing defects on leakage behavior CFRP. A macroscopic homogeneous model incorporating was developed to account for their impact leakage. By combining simulation and experimental methods, relationship between rates analyzed. The results indicate a positive correlation rate porosity. Furthermore, is also influenced by void spacing, where an initial increase in spacing leads rise rate, followed stabilization at later stages. In addition, shows...

Abstract BACKGROUND Integrating non‐invasive measures to estimate abnormal amyloid beta accumulation (Aβ+) is key developing a screening tool for preclinical Alzheimer's disease (AD). The predictive capability of standard neuropsychological tests in estimating Aβ+ has not been quantified. METHODS We constructed machine learning models using six cognitive measurements alongside demographic and genetic risk factors predict Aβ status. Data were drawn from three cohorts: Anti‐Amyloid Treatment...

Point cloud registration is a fundamental problem in 3D computer vision. Outdoor LiDAR point clouds are typically large-scale and complexly distributed, which makes the challenging. In this paper, we propose an efficient hierarchical network named HRegNet for outdoor registration. Instead of using all points clouds, performs on hierarchically extracted keypoints descriptors. The overall framework combines reliable features deeper layer precise position information shallower layers to achieve...

In this article, it is demonstrated the generation of high pulse energy, beam quality and brightness mode-locked picosecond pulses from a compact Nd:YVO4 master oscillator power amplifier system. This system mainly consisted SESAM seed generator four-stage multi-pass amplifiers. A picker was adopted prior to amplifiers efficiently reduce repetition rate. The maximum average output 65.5 W obtained with rate 496.85 kHz duration 16.9 ps, corresponding energy 131.83 µJ peak 7.8 MW. While...

A air quality monitoring system based on the NB-IoT was designed for wireless of special field environment. The STM32FI03RCT6 microprocessor unit core processor and data gas pollutants collected by electrochemical sensor through serial port. switching mechanism network storage disconnected designed. communication TCP/IP protocol realized module or GPRS module. experimental results show that transmission is stable. can automatically switch to as abnormal. In absence network, stored in SPI...

Integrating scores from multiple cognitive tests into a single composite has been shown to improve sensitivity detect AD-related impairment. However, existing composites have little amyloid-β status (Aβ +/-) in preclinical AD.

Driver status monitoring systems are a vital component of smart cars in the future, especially era when an increasing amount time is spent vehicle. The heart rate (HR) one most important physiological signals driver status. To infer HR drivers, mainstream existing research focused on capturing subtle heartbeat-induced vibration torso or leveraged photoplethysmography (PPG) that detects cardiac cycle-related blood volume changes microvascular. However, approaches rely dedicated sensors...

In recent years, driver behavior analysis has led to countless assistance systems. these systems, earlier detection of a driver's maneuver intentions offers opportunities improve driving experience and safety. Especially brake maneuvers are fundamental importance because they directly related the avoidance potential hazards.Current state-of-the-art systems rely on release speed accelerator pedal as an indicator whether event is planned. However, this simple practical algorithm, fails capture...

Abstract The characteristics of underwater acoustic radiation generated by a single raindrop impacting the water surface play fundamental role in study rainfall noise. However, sound signal has great randomness amplitude pressure, frequency range and energy size, so that uncertainty spectrum curve is relatively large. In this paper, generating artificial containing only one size raindrop, composed about 136,000 raindrops with an equivalent diameter 0.9 mm was measured reverberation method. A...

Abstract Background For large genomic studies of middle‐aged individuals, the prevalence Alzheimer’s disease (AD) is extremely low, making it difficult to conduct analysis condition. To enable genome‐wide association AD in such datasets, an approach called Genome‐wide by proxy (GWAX) uses family history as a for status. Borrowing from machine learning (ML) literature, we treat development phenotypes pseudo‐labelling task, where ideal label accurately predicts lifetime risk AD. Given this,...

Abstract Background For large genomic studies of middle‐aged individuals, the prevalence Alzheimer’s disease (AD) is extremely low, making it difficult to conduct analysis condition. To enable genome‐wide association AD in such datasets, an approach called Genome‐wide by proxy (GWAX) uses family history as a for status. Borrowing from machine learning (ML) literature, we treat development phenotypes pseudo‐labelling task, where ideal label accurately predicts lifetime risk AD. Given this,...

For emergent situation such as fire rescuing, effective localization and real-time status monitoring are very important because firefighters always at risks. Generally, strive the disaster in indoor environment where GPS signals cannot be acquired. In this paper, we proposed an intelligent system of information acquisition for firefighters. Real-time position firefighter is acquired by dead-reckoning sensors based on step-size measurement. Various physiological integrated to achieve physical...

Urban information is a kind of multi-sources data, the paper focused on data indexes construction, classified for non-spatial R-trees index spatial and puts forward an area hiberarchy tree structure to build up direct relationship between combined queries, experiment shows that much validated something useful obtained.

에코뮤지엄 개념이 중국에 도입된 이후, 중국 소수민족의 지역 문화 유실과 경제 개발의 후퇴 등 현상이 개선되었다. 그러나 소수민족 지역을 위한 관련 선행연구에서 알 수 있듯, 에코뮤지엄에 대한 명확한 평가 지표가 부족하다. 지표의 수립은 에코뮤지엄이 문화를 보호하는 능력을 표준화하며, 지도 및 향상시키는 데에 필수적이다. 이러한 배경하에, 본 연구는 지역의 평가지표를 수립하고자 하였다. 첫째, 초기 연구를 기반으로 지역과 에코뮤지엄의 특성을 두 차례의 델파이 조사를 통해 진행하였으며, 중복 부적합한 그룹화한 후 7개의 핵심 지표와 18개의 상세 지표를 도출하였다. 둘째, 유효성을 검증하기 위해 켄달(Kendall) 상관계수를 조사의 일관성을 검증하였다. 위와 같은 방법을 객관적으로 도출된 지표에는 지역성, 활동성, 다양성, 지속성, 개방성, 공유성, 자발성이 포함된다. 지표는 목표 계획 수립과 정책에 도움을 줄 있다. 또한 에코뮤지엄을 지속적으로 개선하는 필요한 기초 자료로 활용될

Abstract Background Computerized cognitive tests are critical in the screening of preclinical Alzheimer’s Disease (AD) given their ability to be conducted remotely at low cost. Here, we explore whether machine‐learning derived composites remote can improve sensitivity when detecting brain amyloid (Aβ) accumulation cognitively normal (CN) cohorts than traditional pen‐and‐paper alternatives. Method Based on CogState Brief Battery CN individuals from A4 dataset (n = 4230, Aβ‐ 3094), construct a...