As we use our hands frequently in daily activities, the analysis of hand-object interactions plays a critical role to many multimedia understanding and interaction applications. Different from conventional 3D hand-only object-only pose estimation, estimating is more challenging due mutual occlusions between hand object, as well physical constraints them. To overcome these issues, propose fully utilize structural correlations among joints object corners order obtain reliable poses. Our work...

The field of vision-based human hand three-dimensional (3D) shape and pose estimation has attracted significant attention recently owing to its key role in various applications, such as natural humancomputer interactions. With the availability large-scale annotated datasets rapid developments deep neural networks (DNNs), numerous DNN-based data-driven methods have been proposed for accurate estimation. Nonetheless, existence complicated articulation, depth scale ambiguities, occlusions,...

传统地理学/GIS对人群时空行为的研究集中在中宏观尺度,对微观时空行为的研究相对不足。针对这一问题,提出了微观人群数字孪生的概念及其基础逻辑框架,从人群行为实验方法、基础行为/场景行为建模、典型应用及案例分析等方面总结了微观人群数字孪生的技术基础及其研究现状,并从地理学的视角,对基于系统论的人群实验思路和可信人群实验方法、以及利己-利他、个体-群体、确定-随机统一的人群行为建模方法、微观行为地理学和微观人地关系理论体系构建等未来研究进行探讨,希望推动地理学、GIS和虚拟地理环境对微观人群时空行为理论、方法和技术的研究。

Radial pulse signals are produced by the periodic ejection of blood from heart, and physiological pathological information human body can be analyzed extracting time-domain characteristics waves. However, since weak on surface complex, acquisition using traditional curvature method will produce a large error, which cannot meet needs wave analysis in current clinical practice. To solve this problem, multi-morphological signal feature recognition algorithm based one-dimensional deep...

Micro-electro-mechanical system (MEMS) pressure sensors play a significant role in pulse wave acquisition. However, existing MEMS bound with flexible substrate by gold wire are vulnerable to crush fractures, leading sensor failure. Additionally, establishing an effective mapping between the array signal and width remains challenge. To solve above problems, we propose 24-channel acquisition based on novel through-silicon-via (TSV) structure, which connects directly without bonding. Firstly,...

We present a unified framework for camera-space 3D hand pose estimation from single RGB image based on implicit representation. As opposed to recent works, most of which first adopt holistic or pixel-level dense regression obtain relative and then follow with complex second-stage operations global root scale recovery, we propose novel scheme estimate via point-wise voting in camera frustum. Through direct modeling domain inspired by Pixel-aligned Implicit Functions detailed reconstruction,...

The pulse carries important physiological and pathological information about the human body. piezoresistive sensor used to capture vascular pulsation has transitioned from a single-point array. However, interference signal between channels become key bottleneck restricting development of array diagnosis equipment. in contact with obtains signal. When some sensors are displaced due pulsation, other will be driven move, which produce signals. Signal is common problem for arrays, but few people...

Segmentation of pigmented lesions is often affected by factors such as hair around the skin lesions, artificial markings, etc., and complexity lesion itself, boundaries not clear, internal color variable, resulting in segmentation difficulties. Aiming at problem that method using only random forests accurate, a for combination forest fully convolutional neural networks (FCN) proposed. This firstly classifies recognizes images based on to obtain probability distribution background. Then,...

Purpose. Single-period segmentation is one of the important steps in time-domain analysis pulse signals, which basis feature extraction. The existing single-period methods have disadvantages generalization, reliability, and robustness. Method. This paper proposed a period method signals based on long short-term memory (LSTM) network. preprocessing was performed to remove noises baseline drift signals. Thus, LabelMe used label each into two parts according location starting point main wave...

2024年5月30日中国科学院空天信息创新研究院 国家遥感应用工程技术研究中心的黄琳、李文航团队在《遥感学报》发文,介绍了其在数字孪生人群行为研究领域的新进展,建立了数字孪生人群行为实验理论框架,为地铁站疏散瓶颈区域和群组行为规律研究提供新思路。

The load forecasting of host resources is great significance to the operation and maintenance work. traditional methods usually use linear model fit data. In actual equipment, affected by complex internal external environment where many nonlinear factors are included. time series can not well characterize law order improve accuracy, idea decomposing information into part parts proposed, autoregressive integrated moving average (ARIMA) classification regression tree (CART) combined for...

The paper presents a general computer simulation platform CASIA-MR. Designed for simulating experiments on distributed multi-robotic systems, the bears virtues of generality, preciseness, facility as well various functions such real-time display, recording-replay and capability supporting algorithm developing. Moreover, is easy to be modified operated. Many algorithms have been developed this platform.

Due to their strong immersion and real-time interactivity, helmet-mounted VR devices are becoming increasingly popular. Based on these devices, an immersive virtual geographic environment (VGE) provides a promising method for research into crowd behavior in emergency. However, the current cheaper not popular enough will continue coexist with PC-based systems long time. Therefore, heterogeneous distributed (HDVGE) could be feasible solution solve problems caused by various types of clients,...