Accurate and timely estimation of above-ground biomass is crucial for understanding crop growth dynamics,optimizing agricultural input management,and assessing productivity in sustainable farming practices.However,con...Accurate and timely estimation of above-ground biomass is crucial for understanding crop growth dynamics,optimizing agricultural input management,and assessing productivity in sustainable farming practices.However,conventional biomass assessments are destructive and resource-intensive.In contrast,remote sensing techniques,particularly those utilizing low-altitude unmanned aerial vehicles,provide a non-destructive approach to collect imagery data on plant canopy features,including spectral reflectance and structural details at any stage of the crop life cycle.This study explores the potential visible-light-derived vegetative indices to improve biomass prediction during the flowering period of buckwheat(Fagopyrum tataricum).Red,green,and blue(RGB)images of buckwheat were acquired during peak flowering,using a DJI P4 multispectral Drone.From the analysis of those images,four vegetative indices were calculated.Aboveground fresh biomass was harvested and measured on 14 September 2024.The results showed negative correlations between the green-band based excess green(ExG),excess green minus excess red(ExGR),and green leaf index(GLI)indices and the fresh above-ground biomass of buckwheat,while the red band-based excess red(ExR)index showed an insignificant positive correlation at p<0.10.An investigation into greenband-based vegetation indices(VIs)for estimating fresh biomass revealed significant negative correlations during the experimental period.This unexpected inverse relationship is attributed to spectral interference from abundant white flowers during the flowering stage,where the high reflectance of white petals masked the green vegetation signal.Consequently,these green-band VIs demonstrated limited predictive power for biomass under such conditions,indicating that their utility is compromised when floral reflectance is dominant.Therefore,we suggest that further experiments are required to validate this relationship and improve the estimation of fresh above-ground biomass in white-flowered buckwheat plants.展开更多
In this paper we propose a novel method for video quality prediction using video classification. In essence, our ap- proach can serve two goals: (1) To measure the video quality of compressed video sequences without r...In this paper we propose a novel method for video quality prediction using video classification. In essence, our ap- proach can serve two goals: (1) To measure the video quality of compressed video sequences without referencing to the original uncompressed videos, i.e., to realize No-Reference (NR) video quality evaluation; (2) To predict quality scores for uncompressed video sequences at various bitrates without actually encoding them. The use of our approach can help realize video streaming with ideal Quality of Service (QoS). Our approach is a low complexity solution, which is specially suitable for application to mobile video streaming where the resources at the handsets are scarce.展开更多
The measurement of solar irradiation is still a necessary basis for planning the installation of photovoltaic parks and concentrating solar power systems. The meteorological stations for the measurement of the solar f...The measurement of solar irradiation is still a necessary basis for planning the installation of photovoltaic parks and concentrating solar power systems. The meteorological stations for the measurement of the solar flux at any point of the earth’s surface are still insufficient worldwide;moreover, these measurements on the ground are expensive, and rare. To overcome this shortcoming, the exploitation of images from the European meteorological satellites of the second generation MSG is a reliable solution to estimate the global horizontal irradiance GHI on the ground with a good spatial and temporal coverage. Since 2004, the new generation MSG satellites provide images of Africa and Europe every 15 minutes with a spatial resolution of about 1 km × 1 km at the sub-satellite point. The objective of this work was to apply the Brazil-SR method to evaluate the global horizontal GHI irradiance for the entire Moroccan national territory from the European Meteosat Second Generation MSG satellite images. This bibliographic review also exposed the standard model of calculation of GHI in clear sky by exploiting the terrestrial meteorological measurements.展开更多
Key challenges for 5G and Beyond networks relate with the requirements for exceptionally low latency, high reliability, and extremely high data rates. The Ultra-Reliable Low Latency Communication (URLLC) use case is t...Key challenges for 5G and Beyond networks relate with the requirements for exceptionally low latency, high reliability, and extremely high data rates. The Ultra-Reliable Low Latency Communication (URLLC) use case is the trickiest to support and current research is focused on physical or MAC layer solutions, while proposals focused on the network layer using Machine Learning (ML) and Artificial Intelligence (AI) algorithms running on base stations and User Equipment (UE) or Internet of Things (IoT) devices are in early stages. In this paper, we describe the operation rationale of the most recent relevant ML algorithms and techniques, and we propose and validate ML algorithms running on both cells (base stations/gNBs) and UEs or IoT devices to handle URLLC service control. One ML algorithm runs on base stations to evaluate latency demands and offload traffic in case of need, while another lightweight algorithm runs on UEs and IoT devices to rank cells with the best URLLC service in real-time to indicate the best one cell for a UE or IoT device to camp. We show that the interplay of these algorithms leads to good service control and eventually optimal load allocation, under slow load mobility. .展开更多
This paper presents the evaluation performance of a hybrid satellite constellation network which provides internet access based Transport Control Protocol (TCP). The evaluated satellite network uses the constellation ...This paper presents the evaluation performance of a hybrid satellite constellation network which provides internet access based Transport Control Protocol (TCP). The evaluated satellite network uses the constellation topology operating on Low Earth Orbit (LEO). The COMMStellationTM satellite system implemented using network simulator 2 (NS-2). In this paper, we modified a congestion window control based on TCP Westwood. The results show that the proposed technique reduces end-to-end (E2E) delay of transmitting packets. Hence the number of retransmission packets reduced.展开更多
The rapid development of biological and medical examination methods has vastly expanded personal biomedical information,including molecular,cel-lular,image,and electronic health record datasets.Integrating this wealth...The rapid development of biological and medical examination methods has vastly expanded personal biomedical information,including molecular,cel-lular,image,and electronic health record datasets.Integrating this wealth of information enables precise disease diagnosis,biomarker identification,and treatment design in clinical settings.Artificial intelligence(Al)techniques,particularly deep learning models,have been extensively employed in biomedical applications,demonstrating increased precision,efficiency,and generalization.The success of the large language and vision models fur-ther significantly extends their biomedical applications.However,challenges remain in learning these multimodal biomedical datasets,such as data privacy,fusion,and model interpretation.In this review,we provide a comprehensive overview of various biomedical data modalities,multimodal rep-resentation learning methods,and the applications of Al in biomedical data integrative analysis.Additionally,we discuss the challenges in applying these deep learning methods and how to better integrate them into biomedical scenarios.We then propose future directions for adapting deep learn-ing methods with model pretraining and knowledge integration to advance biomedical research and benefit their clinical applications.展开更多
Three-dimensional(3D)reconstruction of shapes is an important research topic in the fields of computer vision,computer graphics,pattern recognition,and virtual reality.Existing 3D reconstruction methods usually suffer...Three-dimensional(3D)reconstruction of shapes is an important research topic in the fields of computer vision,computer graphics,pattern recognition,and virtual reality.Existing 3D reconstruction methods usually suffer from two bottlenecks:(1)they involve multiple manually designed states which can lead to cumulative errors,but can hardly learn semantic features of 3D shapes automatically;(2)they depend heavily on the content and quality of images,as well as precisely calibrated cameras.As a result,it is difficult to improve the reconstruction accuracy of those methods.3D reconstruction methods based on deep learning overcome both of these bottlenecks by automatically learning semantic features of 3D shapes from low-quality images using deep networks.However,while these methods have various architectures,in-depth analysis and comparisons of them are unavailable so far.We present a comprehensive survey of 3D reconstruction methods based on deep learning.First,based on different deep learning model architectures,we divide 3D reconstruction methods based on deep learning into four types,recurrent neural network,deep autoencoder,generative adversarial network,and convolutional neural network based methods,and analyze the corresponding methodologies carefully.Second,we investigate four representative databases that are commonly used by the above methods in detail.Third,we give a comprehensive comparison of 3D reconstruction methods based on deep learning,which consists of the results of different methods with respect to the same database,the results of each method with respect to different databases,and the robustness of each method with respect to the number of views.Finally,we discuss future development of 3D reconstruction methods based on deep learning.展开更多
Service providers usually require detailed statistics in order to improve their services.On the other hand,privacy concerns are intensifying and sensitive data is protected by legislation,such as GDPR(General Data Pro...Service providers usually require detailed statistics in order to improve their services.On the other hand,privacy concerns are intensifying and sensitive data is protected by legislation,such as GDPR(General Data Protection Regulation).In this paper,we present the design,implementation,and evaluation of a marketplace that allows“data consumers”to buy information from“data providers”,which can then be used for generating meaningful statistics.Additionally,our system enables“system operators”that can select which data providers are allowed to provide data,based on filtering criteria specified by the data consumer.We leverage local differential privacy to protect the data provider's privacy against data consumers,as well as against system operators,and we build a blockchain-based solution for ensuring fair exchange,and immutable data logs.Our design targets use cases that involve hundreds or even thousands of data providers.We prove the feasibility of our approach through a proof-of concept implementation of a measurement sharing application for smart-grid systems.展开更多
AVS stands for the Audio Video coding Standard Workgroup of China, who develops audio/video coding standards as well as system and digital right management standards. AVS-M is the AVS video coding standard targeting f...AVS stands for the Audio Video coding Standard Workgroup of China, who develops audio/video coding standards as well as system and digital right management standards. AVS-M is the AVS video coding standard targeting for mobile multimedia applications. Besides the coding specification, AVS also developed the file format and Real-time Transport Protocol （RTP） payload format specifications to enable the application of AVS-M video in various services. This paper reviews the high-level coding tools and features of the AVS-M coding standard as well as the file format and payload format standards. In particular, sixteen AVS-M high-level coding tools and features, which cover most of the high-level topics during AVS-M standardization, are discussed in some detail. After that, the error resilience tools are briefly reviewed before the file format and RTP payload format discussions. The coding efficiency and error resiliency performances of AVS-M are provided finally. H.264/AVC has been extensively used as a comparison in many of the discussions and the simulation results.展开更多
The proliferation of Internet of Things(IoT)devices that operate unattended providing a multitude of important and often sensitive services highlights the need for seamless interoperability and increased security.We a...The proliferation of Internet of Things(IoT)devices that operate unattended providing a multitude of important and often sensitive services highlights the need for seamless interoperability and increased security.We argue that digital twins of IoT devices,with the right design,can enhance the security,reliability,auditability,and interoperability of IoT systems.The salient features of digital twins have made them key elements for the IoT and Industry 4.0.In this paper,we leverage advances in W3C’s Web of Things(WoT)standards and distributed ledger technologies(DLTs)to present a novel design of the smart contract-based digital twins with enhanced security,transparency,interoperability,and reliability.We provide two different variations of that general design using two different blockchains(one public and one private,permissioned blockchain),and we present design trade-offs.Furthermore,we introduce an architecture for accessing and controlling IoT devices securely and reliably,providing full auditability,while at the same time using the proposed digital twins as an indirection mechanism(proxy).The proposed architecture leverages the blockchain to offer notable properties,namely,decentralization,immutability,auditability,non-repudiation,availability,and reliability.Moreover,it introduces mass actuation,easier management of IoT devices,and enhanced security to the IoT gateways,enables new business models,and makes consumer devices(vendor-)agnostic.展开更多
Dynamic programming algorithms based on Lagrange multiplier method is often used for obtaining an optimal bit allocation strategy to minimize the total distortion given a constrained rate budget in both source and cha...Dynamic programming algorithms based on Lagrange multiplier method is often used for obtaining an optimal bit allocation strategy to minimize the total distortion given a constrained rate budget in both source and channel coding applications. Due to possible large quantizer set and improper initialization, the algorithm often suffers from heavy computational complexity. There have been many solutions in recent years to the above question. In this paper,a simple but efficient algorithm is presented to further speed up the convergence of the algorithm.This algorithm can be easily realized and get the final solution much faster. The experimental result shows that our new algorithm can figure out the optimal solution with a speed 5-7 times faster than the original algorithm.展开更多
Recently,there has been an upsurge of activity in image-based non-photorealistic rendering(NPR),and in particular portrait image stylisation,due to the advent of neural style transfer(NST).However,the state of perform...Recently,there has been an upsurge of activity in image-based non-photorealistic rendering(NPR),and in particular portrait image stylisation,due to the advent of neural style transfer(NST).However,the state of performance evaluation in this field is poor,especially compared to the norms in the computer vision and machine learning communities.Unfortunately,the task of evaluating image stylisation is thus far not well defined,since it involves subjective,perceptual,and aesthetic aspects.To make progress towards a solution,this paper proposes a new structured,threelevel,benchmark dataset for the evaluation of stylised portrait images.Rigorous criteria were used for its construction,and its consistency was validated by user studies.Moreover,a new methodology has been developed for evaluating portrait stylisation algorithms,which makes use of the different benchmark levels as well as annotations provided by user studies regarding the characteristics of the faces.We perform evaluation for a wide variety of image stylisation methods(both portrait-specific and general purpose,and also both traditional NPR approaches and NST)using the new benchmark dataset.展开更多
基金supported by the 2025 scientific promotion program funded by Jeju National University.
文摘Accurate and timely estimation of above-ground biomass is crucial for understanding crop growth dynamics,optimizing agricultural input management,and assessing productivity in sustainable farming practices.However,conventional biomass assessments are destructive and resource-intensive.In contrast,remote sensing techniques,particularly those utilizing low-altitude unmanned aerial vehicles,provide a non-destructive approach to collect imagery data on plant canopy features,including spectral reflectance and structural details at any stage of the crop life cycle.This study explores the potential visible-light-derived vegetative indices to improve biomass prediction during the flowering period of buckwheat(Fagopyrum tataricum).Red,green,and blue(RGB)images of buckwheat were acquired during peak flowering,using a DJI P4 multispectral Drone.From the analysis of those images,four vegetative indices were calculated.Aboveground fresh biomass was harvested and measured on 14 September 2024.The results showed negative correlations between the green-band based excess green(ExG),excess green minus excess red(ExGR),and green leaf index(GLI)indices and the fresh above-ground biomass of buckwheat,while the red band-based excess red(ExR)index showed an insignificant positive correlation at p<0.10.An investigation into greenband-based vegetation indices(VIs)for estimating fresh biomass revealed significant negative correlations during the experimental period.This unexpected inverse relationship is attributed to spectral interference from abundant white flowers during the flowering stage,where the high reflectance of white petals masked the green vegetation signal.Consequently,these green-band VIs demonstrated limited predictive power for biomass under such conditions,indicating that their utility is compromised when floral reflectance is dominant.Therefore,we suggest that further experiments are required to validate this relationship and improve the estimation of fresh above-ground biomass in white-flowered buckwheat plants.
文摘In this paper we propose a novel method for video quality prediction using video classification. In essence, our ap- proach can serve two goals: (1) To measure the video quality of compressed video sequences without referencing to the original uncompressed videos, i.e., to realize No-Reference (NR) video quality evaluation; (2) To predict quality scores for uncompressed video sequences at various bitrates without actually encoding them. The use of our approach can help realize video streaming with ideal Quality of Service (QoS). Our approach is a low complexity solution, which is specially suitable for application to mobile video streaming where the resources at the handsets are scarce.
文摘The measurement of solar irradiation is still a necessary basis for planning the installation of photovoltaic parks and concentrating solar power systems. The meteorological stations for the measurement of the solar flux at any point of the earth’s surface are still insufficient worldwide;moreover, these measurements on the ground are expensive, and rare. To overcome this shortcoming, the exploitation of images from the European meteorological satellites of the second generation MSG is a reliable solution to estimate the global horizontal irradiance GHI on the ground with a good spatial and temporal coverage. Since 2004, the new generation MSG satellites provide images of Africa and Europe every 15 minutes with a spatial resolution of about 1 km × 1 km at the sub-satellite point. The objective of this work was to apply the Brazil-SR method to evaluate the global horizontal GHI irradiance for the entire Moroccan national territory from the European Meteosat Second Generation MSG satellite images. This bibliographic review also exposed the standard model of calculation of GHI in clear sky by exploiting the terrestrial meteorological measurements.
文摘Key challenges for 5G and Beyond networks relate with the requirements for exceptionally low latency, high reliability, and extremely high data rates. The Ultra-Reliable Low Latency Communication (URLLC) use case is the trickiest to support and current research is focused on physical or MAC layer solutions, while proposals focused on the network layer using Machine Learning (ML) and Artificial Intelligence (AI) algorithms running on base stations and User Equipment (UE) or Internet of Things (IoT) devices are in early stages. In this paper, we describe the operation rationale of the most recent relevant ML algorithms and techniques, and we propose and validate ML algorithms running on both cells (base stations/gNBs) and UEs or IoT devices to handle URLLC service control. One ML algorithm runs on base stations to evaluate latency demands and offload traffic in case of need, while another lightweight algorithm runs on UEs and IoT devices to rank cells with the best URLLC service in real-time to indicate the best one cell for a UE or IoT device to camp. We show that the interplay of these algorithms leads to good service control and eventually optimal load allocation, under slow load mobility. .
文摘This paper presents the evaluation performance of a hybrid satellite constellation network which provides internet access based Transport Control Protocol (TCP). The evaluated satellite network uses the constellation topology operating on Low Earth Orbit (LEO). The COMMStellationTM satellite system implemented using network simulator 2 (NS-2). In this paper, we modified a congestion window control based on TCP Westwood. The results show that the proposed technique reduces end-to-end (E2E) delay of transmitting packets. Hence the number of retransmission packets reduced.
基金supported by the National Key R&D Program(Grant Nos.2023YFF1204701 and 2022YFF1202101)the Self-supporting Program of Guangzhou Laboratory(Grant No.SRPG22007)+1 种基金the CAS Research Fund(Grant No.XDB38050200)Guangdong Basic and Applied Basic Research Foundation(Grant No.2023B1515130008),China.
文摘The rapid development of biological and medical examination methods has vastly expanded personal biomedical information,including molecular,cel-lular,image,and electronic health record datasets.Integrating this wealth of information enables precise disease diagnosis,biomarker identification,and treatment design in clinical settings.Artificial intelligence(Al)techniques,particularly deep learning models,have been extensively employed in biomedical applications,demonstrating increased precision,efficiency,and generalization.The success of the large language and vision models fur-ther significantly extends their biomedical applications.However,challenges remain in learning these multimodal biomedical datasets,such as data privacy,fusion,and model interpretation.In this review,we provide a comprehensive overview of various biomedical data modalities,multimodal rep-resentation learning methods,and the applications of Al in biomedical data integrative analysis.Additionally,we discuss the challenges in applying these deep learning methods and how to better integrate them into biomedical scenarios.We then propose future directions for adapting deep learn-ing methods with model pretraining and knowledge integration to advance biomedical research and benefit their clinical applications.
基金Project supported by the National Natural Science Foundation of China(Nos.61772049,61632006,61876012,U19B2039,and 61906011)the Beijing Natural Science Foundation of China(No.4202003)。
文摘Three-dimensional(3D)reconstruction of shapes is an important research topic in the fields of computer vision,computer graphics,pattern recognition,and virtual reality.Existing 3D reconstruction methods usually suffer from two bottlenecks:(1)they involve multiple manually designed states which can lead to cumulative errors,but can hardly learn semantic features of 3D shapes automatically;(2)they depend heavily on the content and quality of images,as well as precisely calibrated cameras.As a result,it is difficult to improve the reconstruction accuracy of those methods.3D reconstruction methods based on deep learning overcome both of these bottlenecks by automatically learning semantic features of 3D shapes from low-quality images using deep networks.However,while these methods have various architectures,in-depth analysis and comparisons of them are unavailable so far.We present a comprehensive survey of 3D reconstruction methods based on deep learning.First,based on different deep learning model architectures,we divide 3D reconstruction methods based on deep learning into four types,recurrent neural network,deep autoencoder,generative adversarial network,and convolutional neural network based methods,and analyze the corresponding methodologies carefully.Second,we investigate four representative databases that are commonly used by the above methods in detail.Third,we give a comprehensive comparison of 3D reconstruction methods based on deep learning,which consists of the results of different methods with respect to the same database,the results of each method with respect to different databases,and the robustness of each method with respect to the number of views.Finally,we discuss future development of 3D reconstruction methods based on deep learning.
基金supported by the EU funded Horizon 2020 project SOFIE(Secure Open Federation for Internet Everywhere),under grant agreement No.779984.
文摘Service providers usually require detailed statistics in order to improve their services.On the other hand,privacy concerns are intensifying and sensitive data is protected by legislation,such as GDPR(General Data Protection Regulation).In this paper,we present the design,implementation,and evaluation of a marketplace that allows“data consumers”to buy information from“data providers”,which can then be used for generating meaningful statistics.Additionally,our system enables“system operators”that can select which data providers are allowed to provide data,based on filtering criteria specified by the data consumer.We leverage local differential privacy to protect the data provider's privacy against data consumers,as well as against system operators,and we build a blockchain-based solution for ensuring fair exchange,and immutable data logs.Our design targets use cases that involve hundreds or even thousands of data providers.We prove the feasibility of our approach through a proof-of concept implementation of a measurement sharing application for smart-grid systems.
文摘AVS stands for the Audio Video coding Standard Workgroup of China, who develops audio/video coding standards as well as system and digital right management standards. AVS-M is the AVS video coding standard targeting for mobile multimedia applications. Besides the coding specification, AVS also developed the file format and Real-time Transport Protocol （RTP） payload format specifications to enable the application of AVS-M video in various services. This paper reviews the high-level coding tools and features of the AVS-M coding standard as well as the file format and payload format standards. In particular, sixteen AVS-M high-level coding tools and features, which cover most of the high-level topics during AVS-M standardization, are discussed in some detail. After that, the error resilience tools are briefly reviewed before the file format and RTP payload format discussions. The coding efficiency and error resiliency performances of AVS-M are provided finally. H.264/AVC has been extensively used as a comparison in many of the discussions and the simulation results.
文摘The proliferation of Internet of Things(IoT)devices that operate unattended providing a multitude of important and often sensitive services highlights the need for seamless interoperability and increased security.We argue that digital twins of IoT devices,with the right design,can enhance the security,reliability,auditability,and interoperability of IoT systems.The salient features of digital twins have made them key elements for the IoT and Industry 4.0.In this paper,we leverage advances in W3C’s Web of Things(WoT)standards and distributed ledger technologies(DLTs)to present a novel design of the smart contract-based digital twins with enhanced security,transparency,interoperability,and reliability.We provide two different variations of that general design using two different blockchains(one public and one private,permissioned blockchain),and we present design trade-offs.Furthermore,we introduce an architecture for accessing and controlling IoT devices securely and reliably,providing full auditability,while at the same time using the proposed digital twins as an indirection mechanism(proxy).The proposed architecture leverages the blockchain to offer notable properties,namely,decentralization,immutability,auditability,non-repudiation,availability,and reliability.Moreover,it introduces mass actuation,easier management of IoT devices,and enhanced security to the IoT gateways,enables new business models,and makes consumer devices(vendor-)agnostic.
文摘Dynamic programming algorithms based on Lagrange multiplier method is often used for obtaining an optimal bit allocation strategy to minimize the total distortion given a constrained rate budget in both source and channel coding applications. Due to possible large quantizer set and improper initialization, the algorithm often suffers from heavy computational complexity. There have been many solutions in recent years to the above question. In this paper,a simple but efficient algorithm is presented to further speed up the convergence of the algorithm.This algorithm can be easily realized and get the final solution much faster. The experimental result shows that our new algorithm can figure out the optimal solution with a speed 5-7 times faster than the original algorithm.
文摘Recently,there has been an upsurge of activity in image-based non-photorealistic rendering(NPR),and in particular portrait image stylisation,due to the advent of neural style transfer(NST).However,the state of performance evaluation in this field is poor,especially compared to the norms in the computer vision and machine learning communities.Unfortunately,the task of evaluating image stylisation is thus far not well defined,since it involves subjective,perceptual,and aesthetic aspects.To make progress towards a solution,this paper proposes a new structured,threelevel,benchmark dataset for the evaluation of stylised portrait images.Rigorous criteria were used for its construction,and its consistency was validated by user studies.Moreover,a new methodology has been developed for evaluating portrait stylisation algorithms,which makes use of the different benchmark levels as well as annotations provided by user studies regarding the characteristics of the faces.We perform evaluation for a wide variety of image stylisation methods(both portrait-specific and general purpose,and also both traditional NPR approaches and NST)using the new benchmark dataset.
