Person re-identification (Re-ID) has been widely used in public security and surveillance. Due to the influence of different shooting times and locations, can lead to lighting variations in the images captured by the ...
详细信息
Attitude estimation is a pivotal issue in unmanned aerial vehicle (UAV) control system. Due to its high dynamics, the current attitude estimation methods suffer from drift and instability in dynamic situations. In thi...
详细信息
This research develops an effective and precise collision detection (CD) algorithm for real-time simulation in virtual environments such as computer graphics, realistic and immersive virtual reality (VR), augmented re...
详细信息
This research develops an effective and precise collision detection (CD) algorithm for real-time simulation in virtual environments such as computer graphics, realistic and immersive virtual reality (VR), augmented reality (AR) and physical-based simulation within an enhanced algorithm for object collision detection in 3D geometry. We describe an improved algorithm through a comparison in the application of a central processing unit (CPU) and graphics processing units (GPU). Although leveraging CPU for computational speed improvements has gained significant recognition in recent years, this study distinguishes by tracking 3D geometry bounding volume hierarchy (BVH) constructed in a spatial decomposition structure with a focus on Octree-based Axis-Aligned Bounding Box (AABB) structure in 3D scene to compute collision detection to swiftly reject disjoint objects and minimize the number of triangle primitives that need to be processed and then the M & ouml;ller method is utilized to compute precise triangle primitives, further enhancing the efficiency and precision of the collision detection process. This approach is also designed to implement computation with GPU which utilizes the high-level shader language (HLSL) programming language on the compute shader Unity3D. AABB is structured as the maximum and minimum hexahedron enclosing an object that is parallel to the coordinate axis. Otherwise, GPU computational technique is a crucial method for further enhancing the object's performance. The proposed method utilizes Octree AABB-based GPU parallel processing to reduce the computational load of real-time collision detection simulations and to handle multiple computations simultaneously. Comparative performance evaluations demonstrate that our GPU-accelerated framework consistently reaches the fastest collision detection times from 1.01 to 45.62 times, respectively.
Irony is nowadays a pervasive phenomenon in social networks. The multimodal functionalities of these platforms (i.e., the possibility to attach audio, video, and images to textual information) are increasingly leading...
详细信息
Beyond the text detection and recognition tasks in image text spotting, video text spotting presents an augmented challenge with the inclusion of tracking. While advanced end-to-end trainable methods have shown commen...
Cryptography is used by all organizations to protect the data files and ensures confidentiality mainly at the time of sharing and storing in the cloud data storage. The cloud service providers use a wide range of tool...
详细信息
How to quickly and accurately detect the surface defects of objects has always been the focus of computer vision research. In this paper, a defect detection based on AI method for anode copper plate is proposed. First...
详细信息
Emotion Recognition in Conversations (ERC) facilitates a deeper understanding of the emotions conveyed by speakers in each utterance within a conversation. Recently, Graph Neural Networks (GNNs) have demonstrated thei...
详细信息
The seamless integration of intelligent Internet of Things devices with conventional wireless sensor networks has revolutionized data communication for different applications,such as remote health monitoring,industria...
详细信息
The seamless integration of intelligent Internet of Things devices with conventional wireless sensor networks has revolutionized data communication for different applications,such as remote health monitoring,industrial monitoring,transportation,and smart *** and reliable data routing is one of the major challenges in the Internet of Things network due to the heterogeneity of *** paper presents a traffic-aware,cluster-based,and energy-efficient routing protocol that employs traffic-aware and cluster-based techniques to improve the data delivery in such *** proposed protocol divides the network into clusters where optimal cluster heads are selected among super and normal nodes based on their residual *** protocol considers multi-criteria attributes,i.e.,energy,traffic load,and distance parameters to select the next hop for data delivery towards the base *** performance of the proposed protocol is evaluated through the network simulator *** different traffic rates,number of nodes,and different packet sizes,the proposed protocol outperformed LoRaWAN in terms of end-to-end packet delivery ratio,energy consumption,end-to-end delay,and network *** 100 nodes,the proposed protocol achieved a 13%improvement in packet delivery ratio,10 ms improvement in delay,and 10 mJ improvement in average energy consumption over LoRaWAN.
Long-term forecasting is widely used in meteorology, hydrology, and finance. However, non-stationary time series make it hard to make accurate long-term predictions because of their complicated multi-period local-glob...
ISBN:
(纸本)9798350337020
Long-term forecasting is widely used in meteorology, hydrology, and finance. However, non-stationary time series make it hard to make accurate long-term predictions because of their complicated multi-period local-global temporal dynamic patterns. Currently, state-of-the-art methods use transformers or temporal convolutions to obtain global and local temporal dynamic patterns. Nevertheless, the former suffers from the computational complexity of self-attention mechanisms despite having a global temporal receptive field. Despite being able to catch local temporal patterns, the latter requires additional layers to capture global temporal patterns. Moreover, the present research disregards integrating multi-period patterns into longterm forecasting. In this paper, we propose MLGNet to tackle the mentioned challenges, which integrates local and global temporal dynamic patterns with multiple periods for longterm forecasting. In particular, we suggest using the maximal overlap discrete wavelet transform (MODWT) as a multi-period decoupling method to decompose non-stationary time series and apply it for the first time to long-term forecasting. In addition, we suggest a multi-scale encoder-decoder framework to capture and fuse local-global temporal dynamic patterns in each decomposed period. Inception dilated causal convolutions-based encoder and a lightweight MLP-based decoder in the framework capture local and global temporal dynamic patterns in series while avoiding the high computational complexity of self-attention mechanisms. Lastly, we suggest time-separable convolutions for aggregating information on temporal dynamic patterns among multiple periods. The above method helps MLGNet better balance the representation ability of time series in 1D and 2D space. Evaluation of five benchmark datasets shows that MLGNet outperforms traditional and state-of-the-art methods, with relative improvements of 13.8 % and 21.9% for multivariate and univariate long-term forecasting, resp
暂无评论