Article
Agriculture, Multidisciplinary
Shisong Zhu, Wanli Ma, Jiangwen Lu, Bo Ren, Chunyang Wang, Jianlong Wang
Summary: To address the difficulties in leaf edge identification and the imbalance of pixel ratios between the background area and the target area, a novel two-stage DeepLabv3+ model with adaptive loss is proposed for apple leaf disease image segmentation. The adaptive loss reduces the weight of losses generated by easily classified pixels, allowing the model to focus more on hard-to-classify pixels and improve segmentation accuracy. The experimental results show that the proposed LD-DeepLabv3+ model with adaptive loss achieves high IoU scores of 98.70% for leaf segmentation and 86.56% for spot extraction. It also outperforms the two-stage model DUNet in terms of segmentation accuracy and computational costs.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2023)
Article
Automation & Control Systems
Yazhou Yuan, Xu Cao, Zhixin Liu, Cailian Chen, Xinping Guan
Summary: With the popularity of smart terminal equipment, the interaction between industrial field information systems and production equipment has become more intense. In order to achieve real-time and coordinated flow transmission, time-sensitive network-related technologies are used for flow queue forwarding. The proposed algorithm improves scheduling success rate, reduces delays, and achieves optimal scheduling method.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Automation & Control Systems
Saquib Mazhar, Nadeem Atif, M. K. Bhuyan, Shaik Rafi Ahamed
Summary: This study proposes a deep-learning-based semantic segmentation network that incorporates attention-based context guidance to recover context information. The network runs efficiently on resource-constrained devices and achieves high accuracy with a lightweight design and minimal parameters.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Zhan Shi, Zhanshan Wang
Summary: This paper investigates an adaptive output-feedback optimal control problem for a class of continuous time (CT) linear systems with dynamic uncertainties. An algorithm based on adaptive dynamic programming (ADP) technique is proposed for data-based controller design, which only uses measured input and output information to learn optimal control gain without requiring exact system knowledge. The adaptive controllers learned by the algorithm exhibit robustness to dynamic uncertainties, as demonstrated through three examples.
Article
Computer Science, Information Systems
Shiva Raj Pokhrel, Jinho Choi
Summary: In this study, we investigate the coexistence of delay-sensitive and delay-tolerant devices in machine-type communication (MTC) in future cellular networks. We propose an extension of the fast retrial idea to reduce access delay for delay-sensitive devices. Using a control-theoretic approach, we analyze the stability dynamics and develop an adaptive algorithm to allocate the number of preambles for delay-sensitive devices. Our findings are validated through extensive simulations. Furthermore, we introduce a novel framework that applies control theory to address buffer tracking and coexistence in diverse network environments under realistic application constraints. The extension of the control-theoretic idea predicts system stability and promotes desynchronization with smaller delay-sensitive buffers.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Aiman Almas, Waseem Iqbal, Ayesha Altaf, Kashif Saleem, Shynar Mussiraliyeva, Muhammad Wajahat Iqbal
Summary: Fog computing is suitable for scenarios with a large number of decentralized devices that require real-time communication and data analysis. It provides dependability and security for time-critical smart healthcare systems. However, trust solutions for fog computing in healthcare are lacking and this research proposes a context-based adaptive trust model using a Bayesian approach and similarity measures.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Neurosciences
Manuel Schottdorf, Barry B. Lee
Summary: Research shows that primate ganglion cell responses to natural scenes are primarily driven by temporal variations in color and luminance caused by eye movements, with little influence from interaction with receptive field structure. Model predictions suggest that responses derive from the temporal pattern of stimulation from eye movements, reducing redundancy in the retinal signal. The magnocellular pathway is better suited to transmit detailed structure of natural scenes than the parvocellular pathway.
JOURNAL OF PHYSIOLOGY-LONDON
(2021)
Article
Computer Science, Software Engineering
Ke Li, Hantao Zhao, Qian Zhang, Jinyuan Jia
Summary: With the continuous growth and evolution of mobile technology, presenting BIM with an online platform has become an important application in civil engineering, architecture, and computer visualization. This article proposes CEBOW, an online architecture that combines transmission scheduling, cache management, and optimal initial loading to effectively improve the visualization of large-scale BIM scenes and reduce loading time and network delay.
COMPUTER ANIMATION AND VIRTUAL WORLDS
(2022)
Article
Neurosciences
Danping Liu, Dong Zhang, Lei Wang, Jun Wang
Summary: This paper presents a novel semantic segmentation approach, named Multi-Scale Adaptive Mechanism (MSAAM), for autonomous driving scenes. The proposed method effectively addresses the challenges associated with complex driving environments, and achieves precise segmentation by integrating multiple scale features and adaptively selecting the most relevant features.
FRONTIERS IN NEUROSCIENCE
(2023)
Article
Computer Science, Information Systems
Mingxue Zheng, Xiangcheng Shen, Zhiqing Luo, Pingting Chen, Bo Guan, Jicheng Yi, Hairong Ma
Summary: In this paper, a knowledge-based multi-scale adaptive classification approach is proposed for efficiently classifying objects in mobile laser scanning point clouds. By exploring the multi-scale spatial local relation and distinguishable features of objects, the method establishes a match between the feature descriptions and actual objects' point clouds based on human knowledge. Experimental results demonstrate the promising performance of the proposed approach for classifying unlabeled objects in urban scenes.
Article
Physics, Multidisciplinary
Shangwang Liu, Tongbo Cai, Xiufang Tang, Yangyang Zhang, Changgeng Wang
Summary: This paper proposes a traffic sign detection method based on RetinaNet-NeXt, aiming at recognizing small proportion, blurred and complex traffic signs in natural scenes. The method improves the detection accuracy and effection of RetinaNet by utilizing the ResNeXt backbone network, transfer learning, and group normalization. Experimental results show significant improvements in precision, recall, and mAP compared to the original RetinaNet, indicating the effectiveness of the proposed method for traffic sign detection.
Article
Engineering, Environmental
Jonathan Tenorio Vinhal, Rafael Piumatti de Oliveira, Jorge Luis Coleti, Denise Crocce Romano Espinosa
Summary: LED waste contains valuable metals, making the recycling of LED waste important. This study characterized LED models and found variations in metal content and distribution among different models, providing opportunities for future recycling and treatment of LED waste.
Article
Chemistry, Analytical
Jun Dai, Xiangyang Hao, Songlin Liu, Zongbin Ren
Summary: This paper proposes a robust adaptive positioning algorithm for UAVs, which can improve the accuracy and reliability of autonomous navigation and positioning by using a multi-source fusion model and adaptively assigning information sharing coefficients. It is applicable to various complex scenes and sensor combinations.
Article
Geography
Changfeng Jing, Yanli Zhu, Mingyi Du, Xintao Liu
Summary: This study utilized comprehensive city service data along with novel visual analytical methods to examine city service demands in Xicheng District, Beijing. The findings revealed that emerging data and visualization approaches have more explanatory power for overall trends and micro-scale details of city services compared to other methods. High-incidence locations were found to be associated with built environment and population density.
TRANSACTIONS IN GIS
(2021)
Article
Automation & Control Systems
Roberto Franco, Hector Rios, Alejandra Ferreira de Loza
Summary: This article presents a finite-time model reference adaptive control approach to address the robust tracking problem for a class of disturbed scalar linear systems. The proposed method, based on nonlinear adaptive gains, ensures a finite-time convergence rate. The convergence proofs and robustness analysis are established using Lyapunov function approach, input-to-state stability theory, and homogeneity theory. Simulation and experimental results demonstrate the feasibility of the proposed scheme.
INTERNATIONAL JOURNAL OF ADAPTIVE CONTROL AND SIGNAL PROCESSING
(2022)
