Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817258
D. Chiper, A. Cracan
A new Very Large-Scale Integration (VLSI) algorithm for an integer based discrete cosine transform (IntDCT) allowing an effective VLSI implementation based on systolic array is proposed. While maintaining a good approximation of the floating-point transform coefficients, the proposed algorithm has the advantages of an integer transform which also allows a restructuring into a regular and modular, systolic array-based computation structure with an effective VLSI implementation. The resulting VLSI architecture is flexible enough to allow an efficient incorporation of an obfuscation technique that considerably improves the hardware security and offers high-speed performances due to concurrent computation and use of the pipelining technique with a low hardware complexity.
{"title":"A New Integer Algorithm for a VLSI Implementation of DCT Using Obfuscation Technique","authors":"D. Chiper, A. Cracan","doi":"10.1109/comm54429.2022.9817258","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817258","url":null,"abstract":"A new Very Large-Scale Integration (VLSI) algorithm for an integer based discrete cosine transform (IntDCT) allowing an effective VLSI implementation based on systolic array is proposed. While maintaining a good approximation of the floating-point transform coefficients, the proposed algorithm has the advantages of an integer transform which also allows a restructuring into a regular and modular, systolic array-based computation structure with an effective VLSI implementation. The resulting VLSI architecture is flexible enough to allow an efficient incorporation of an obfuscation technique that considerably improves the hardware security and offers high-speed performances due to concurrent computation and use of the pipelining technique with a low hardware complexity.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131715512","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817183
G. Popescu, Călin Bîră
This paper presents a low-maintenance, short development-cycle programming framework (MRAFW) which allows writing and running software for a custom heterogeneous pseudo-reconfigurable computing system integrating a MapReduce Accelerator. The target system implementation is based on AMD/Xilinx's Zynq SoC hardware platform. The programming framework uses the PYNQ software package to enable access to the system resources and to manage CPU-FPGA program and data transfers. Furthermore, we provide a library of optimized low-level functions that offer support for executing some basic linear algebra operations.
{"title":"Python-Based Programming Framework for a Heterogeneous MapReduce Architecture","authors":"G. Popescu, Călin Bîră","doi":"10.1109/comm54429.2022.9817183","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817183","url":null,"abstract":"This paper presents a low-maintenance, short development-cycle programming framework (MRAFW) which allows writing and running software for a custom heterogeneous pseudo-reconfigurable computing system integrating a MapReduce Accelerator. The target system implementation is based on AMD/Xilinx's Zynq SoC hardware platform. The programming framework uses the PYNQ software package to enable access to the system resources and to manage CPU-FPGA program and data transfers. Furthermore, we provide a library of optimized low-level functions that offer support for executing some basic linear algebra operations.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126857549","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817266
A. Jitaru, Cosmina-Elena Barbu, B. Ionescu
Multi-class aircraft recognition is important in aerial surveillance applications to make consistent proposals for decision makers. Motivated by the state-of-art object detection methods for aerial sensing images, we explored and achieved satisfactory results based on standard YOLO archi-tectures by providing an analysis of certain aerial scenarios, from chained to scattered or yin-yang pairs objects and model's performance comparison on a typical wild emergency situation. First, we tackle the scarcity of multi-class aircraft detection data by developing of a 50 cm GSD dataset. Then, applying different training strategies and using a series of good practices, we achieved a F1-score of 0.820 and a con-siderable mAP@50 of 0.809 for the more difficult detection class. Experiments have been conducted over a dataset from Google Earth platform. Thus, our final proposal for aerial surveillance systems contains the pre-trained YOLOv5×6 architecture with attention on performance maximization for multiple specific aerial scenarios, processing 50km2 in no more than 6 seconds.
多类飞机识别在空中监视应用中具有重要意义，可以为决策者提供一致的建议。在当前航空遥感图像目标检测方法的推动下，我们在标准YOLO架构的基础上进行了探索，并取得了令人满意的结果。我们对特定的空中场景进行了分析，从链状到分散或阴阳对的目标，并在典型的野外紧急情况下对模型进行了性能比较。首先，我们通过开发一个50 cm GSD数据集来解决多类飞机检测数据的稀缺性问题。然后，应用不同的训练策略并使用一系列良好的实践，我们获得了f1得分0.820，对于更困难的检测类，我们获得了相当可观的mAP@50得分0.809。在谷歌地球平台的数据集上进行了实验。因此，我们对空中监视系统的最终建议包含预先训练的YOLOv5×6架构，关注多个特定空中场景的性能最大化，在不超过6秒的时间内处理50平方公里。
{"title":"Deep Learning-based Object Searching and Reporting for Aerial Surveillance Systems","authors":"A. Jitaru, Cosmina-Elena Barbu, B. Ionescu","doi":"10.1109/comm54429.2022.9817266","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817266","url":null,"abstract":"Multi-class aircraft recognition is important in aerial surveillance applications to make consistent proposals for decision makers. Motivated by the state-of-art object detection methods for aerial sensing images, we explored and achieved satisfactory results based on standard YOLO archi-tectures by providing an analysis of certain aerial scenarios, from chained to scattered or yin-yang pairs objects and model's performance comparison on a typical wild emergency situation. First, we tackle the scarcity of multi-class aircraft detection data by developing of a 50 cm GSD dataset. Then, applying different training strategies and using a series of good practices, we achieved a F1-score of 0.820 and a con-siderable mAP@50 of 0.809 for the more difficult detection class. Experiments have been conducted over a dataset from Google Earth platform. Thus, our final proposal for aerial surveillance systems contains the pre-trained YOLOv5×6 architecture with attention on performance maximization for multiple specific aerial scenarios, processing 50km2 in no more than 6 seconds.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"84 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116101830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817215
Nicolae Valentin Rusu, Silvia Dumitrache, M. Alexandru
Today's applications focus on connecting multiple devices which communicate using radio waves. The SDR (Soft-ware Defined Radio) is designed to offer a flexible platform which meets today's applications requirements in various fields such as medical monitoring, agriculture processes or public safety. In SDR the physical components are virtualized using software algorithms and concepts. This offers the advantage of using the same hardware equipment for multiple purposes as well as reuse software defined components throughout the project development process. SDR integrates configurable RF (Radio Frequency) devices with the FPGA (Field Programmable Gate Array) platform enabling them to work together and perform digital operations. The suitable software tools for a top-down approach on SDR are Matlab and Simulink which offers a variety of dedicated libraries for radio communications. This paper describes a system dedicated for transmitting and receiving FM (Frequency Modulation) radio communications using the Zedboard re-configurable platform and the Analog Devices AD-FMCOMMS2-EBZ (AD9361) communication module.
{"title":"Software Defined Communications Platform","authors":"Nicolae Valentin Rusu, Silvia Dumitrache, M. Alexandru","doi":"10.1109/comm54429.2022.9817215","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817215","url":null,"abstract":"Today's applications focus on connecting multiple devices which communicate using radio waves. The SDR (Soft-ware Defined Radio) is designed to offer a flexible platform which meets today's applications requirements in various fields such as medical monitoring, agriculture processes or public safety. In SDR the physical components are virtualized using software algorithms and concepts. This offers the advantage of using the same hardware equipment for multiple purposes as well as reuse software defined components throughout the project development process. SDR integrates configurable RF (Radio Frequency) devices with the FPGA (Field Programmable Gate Array) platform enabling them to work together and perform digital operations. The suitable software tools for a top-down approach on SDR are Matlab and Simulink which offers a variety of dedicated libraries for radio communications. This paper describes a system dedicated for transmitting and receiving FM (Frequency Modulation) radio communications using the Zedboard re-configurable platform and the Analog Devices AD-FMCOMMS2-EBZ (AD9361) communication module.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125193649","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817150
Rebecca Acheampong, T. Balan, D. Popovici, Alexandre Rekeraho
Cyber security is a 21st-century challenge doubled by the risk introduced by the lack of cyber security workforce. To bridge this workforce gap, cyber security learning environment are necessary. The Cyber range is cyber virtual environment for security training for cyber security professionals. This study presents the next-generation tools for developing Cyber ranges in the cloud and integrates automation with Ansible hence, orchestrating the virtual environment for the efficacy of security testsbeds, equipping the user with cyber security readiness to defend against real-life cyber-attacks. The proposed automation methodology can be also incorporated in continuous integration queues, defining the Infrastructure as Code (IaC) DevSecOps strategy. The developed scalable and reusable platform allows malicious cyber activities to be simulated through automation. The exercises and the tools employed gives security professionals a good understanding of adversary's operations. while the vulnerabilities are mapped to MITRE ATT &CK knowledge base. Our results show that, the developed Cyber range platform is suitable and effective for cyber security training. Notwithstanding, we propose future studies be carried out to look into cyber security problematic in virtual reality (VR) platforms.
网络安全是一项21世纪的挑战，缺乏网络安全人员带来的风险使其成倍增加。为了弥补这一劳动力缺口，网络安全学习环境是必要的。Cyber靶场是为网络安全专业人员提供安全培训的网络虚拟环境。本研究提出了用于在云中开发网络范围的下一代工具，并将自动化与Ansible集成，从而为安全测试平台的有效性编排虚拟环境，为用户提供网络安全准备，以抵御现实生活中的网络攻击。建议的自动化方法也可以合并到持续集成队列中，定义基础设施即代码(IaC) DevSecOps策略。开发的可扩展和可重用平台允许通过自动化模拟恶意网络活动。演习和使用的工具使安全专业人员对对手的行动有了很好的了解。同时将漏洞映射到MITRE ATT &CK知识库。结果表明，所开发的网络靶场平台适用于网络安全培训，效果良好。尽管如此，我们建议未来进行研究，以调查虚拟现实(VR)平台中的网络安全问题。
{"title":"Security Scenarios Automation and Deployment in Virtual Environment using Ansible","authors":"Rebecca Acheampong, T. Balan, D. Popovici, Alexandre Rekeraho","doi":"10.1109/comm54429.2022.9817150","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817150","url":null,"abstract":"Cyber security is a 21st-century challenge doubled by the risk introduced by the lack of cyber security workforce. To bridge this workforce gap, cyber security learning environment are necessary. The Cyber range is cyber virtual environment for security training for cyber security professionals. This study presents the next-generation tools for developing Cyber ranges in the cloud and integrates automation with Ansible hence, orchestrating the virtual environment for the efficacy of security testsbeds, equipping the user with cyber security readiness to defend against real-life cyber-attacks. The proposed automation methodology can be also incorporated in continuous integration queues, defining the Infrastructure as Code (IaC) DevSecOps strategy. The developed scalable and reusable platform allows malicious cyber activities to be simulated through automation. The exercises and the tools employed gives security professionals a good understanding of adversary's operations. while the vulnerabilities are mapped to MITRE ATT &CK knowledge base. Our results show that, the developed Cyber range platform is suitable and effective for cyber security training. Notwithstanding, we propose future studies be carried out to look into cyber security problematic in virtual reality (VR) platforms.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132118084","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817315
S. Zoican, Marius Constantin Vochin, R. Zoican, D. Galatchi
This paper analyzes and evaluates the performance of neural network testing using microcontrollers. The finite precision of microcontrollers registers and the computation time were considered for evaluation. The paper illustrates a general framework in which a neural network application can be easily integrated. This framework is based on Contiki IoT operating system that ensure support for wireless communication and efficient implementation of processes in a resource constrained system. As a practical example, a human activity recognition system, based on Long Short-Term Memory recurrent neural network, was implemented, and evaluated using a modified evaluation board from Analog Devices. The obtained system accuracy is very good, very close to the accuracy obtained in infinite precision simulation, the computation time is low enough and the system works in real time.
{"title":"Neural Network Testing Framework for Microcontrollers","authors":"S. Zoican, Marius Constantin Vochin, R. Zoican, D. Galatchi","doi":"10.1109/comm54429.2022.9817315","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817315","url":null,"abstract":"This paper analyzes and evaluates the performance of neural network testing using microcontrollers. The finite precision of microcontrollers registers and the computation time were considered for evaluation. The paper illustrates a general framework in which a neural network application can be easily integrated. This framework is based on Contiki IoT operating system that ensure support for wireless communication and efficient implementation of processes in a resource constrained system. As a practical example, a human activity recognition system, based on Long Short-Term Memory recurrent neural network, was implemented, and evaluated using a modified evaluation board from Analog Devices. The obtained system accuracy is very good, very close to the accuracy obtained in infinite precision simulation, the computation time is low enough and the system works in real time.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"97 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116104242","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817178
Alexandru Bobaru, C. Nafornita, Vladimir Cristian Vesa
This paper proposes a stationary target based online unsupervised calibration algorithm that can be applied on both 4D and 3D automotive radars for its horizontal alignment and misalignment detection. The calibration process requires no special EOL (End of Line) setup or stationary structures of reference. The method is based on the accurate determination of the own vehicle velocity and by using stationary targets. The approach provides both a long-term stable azimuth mounting compensation value as well as a separate, more dynamic angle value that converges faster than the long-term value in case of small accidents. The proposed method considers the systematic errors resulted from the vehicle integration and bumper tolerances and delivers an accurate horizontal alignment correction by using filtering outlier rejection techniques. The performance is evaluated used real world data from drive tests executed with a 77 GHz series automotive radar, showing promising results.
本文提出了一种基于静止目标的在线无监督标定算法，该算法可同时应用于四维和三维汽车雷达的水平对准和不对准检测。校准过程不需要特殊的EOL (End of Line)设置或固定的参考结构。该方法是在准确确定车辆自身速度的基础上，利用静止目标实现的。该方法既提供了一个长期稳定的方位角安装补偿值，也提供了一个单独的、更动态的角度值，在发生小事故时，它的收敛速度比长期值更快。该方法考虑了车辆集成和保险杠公差引起的系统误差，并利用滤波离群值抑制技术提供了精确的水平对中校正。使用77 GHz系列汽车雷达进行的驾驶测试的真实数据对性能进行了评估，显示出令人鼓舞的结果。
{"title":"Unsupervised Online Horizontal Misalignment Detection Algorithm for Automotive Radar","authors":"Alexandru Bobaru, C. Nafornita, Vladimir Cristian Vesa","doi":"10.1109/comm54429.2022.9817178","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817178","url":null,"abstract":"This paper proposes a stationary target based online unsupervised calibration algorithm that can be applied on both 4D and 3D automotive radars for its horizontal alignment and misalignment detection. The calibration process requires no special EOL (End of Line) setup or stationary structures of reference. The method is based on the accurate determination of the own vehicle velocity and by using stationary targets. The approach provides both a long-term stable azimuth mounting compensation value as well as a separate, more dynamic angle value that converges faster than the long-term value in case of small accidents. The proposed method considers the systematic errors resulted from the vehicle integration and bumper tolerances and delivers an accurate horizontal alignment correction by using filtering outlier rejection techniques. The performance is evaluated used real world data from drive tests executed with a 77 GHz series automotive radar, showing promising results.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124710560","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817351
Ana Ghiulai, Gordana Barb, F. Alexa, M. Otesteanu
Dynamic spectrum sharing is a new technology that enables LTE and 5G to be used in parallel in the same frequency band. In other words, it is an activation mechanism that allows 5G NR and 4G LTE to coexist while sharing the same frequency spectrum. In the long run, DSS allows mobile network operators to re-use their frequency bands in order to provide increased 5G coverage. In this paper we measure the effect of interference for a DL 5G system using two different mobile terminals. The main purpose of this study is to compare the DL interference of an adaptive system model for 5G mobile communications applications.
{"title":"Downlink Interference Measurement in 4G/5G Systems with Dynamic Spectrum Sharing","authors":"Ana Ghiulai, Gordana Barb, F. Alexa, M. Otesteanu","doi":"10.1109/comm54429.2022.9817351","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817351","url":null,"abstract":"Dynamic spectrum sharing is a new technology that enables LTE and 5G to be used in parallel in the same frequency band. In other words, it is an activation mechanism that allows 5G NR and 4G LTE to coexist while sharing the same frequency spectrum. In the long run, DSS allows mobile network operators to re-use their frequency bands in order to provide increased 5G coverage. In this paper we measure the effect of interference for a DL 5G system using two different mobile terminals. The main purpose of this study is to compare the DL interference of an adaptive system model for 5G mobile communications applications.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122075152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817234
Mircea Nicolaescu, P. Svasta, Mihai Enache
The increase of the operating frequency in order to fulfill the requirements for high speed digital applications makes the electromagnetic compatibility an important issue in the design stage. This paper studies the reflection, transmission and coupling parameters of two singled ended microstrip lines as a function of the separation between them, in the frequency band 0.5-50 GHz. The microstrip traces are printed on high frequency, low loss microwave substrate, Rogers 4003 C. Full wave electromagnetic method had been employed for the study.
{"title":"Full Wave Electromagnetic Analysis of Coupling in High Speed Microstrip Single Ended Lines with Different Separations","authors":"Mircea Nicolaescu, P. Svasta, Mihai Enache","doi":"10.1109/comm54429.2022.9817234","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817234","url":null,"abstract":"The increase of the operating frequency in order to fulfill the requirements for high speed digital applications makes the electromagnetic compatibility an important issue in the design stage. This paper studies the reflection, transmission and coupling parameters of two singled ended microstrip lines as a function of the separation between them, in the frequency band 0.5-50 GHz. The microstrip traces are printed on high frequency, low loss microwave substrate, Rogers 4003 C. Full wave electromagnetic method had been employed for the study.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129847059","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-16DOI: 10.1109/comm54429.2022.9817344
Ionut Girla, V. Neagoe
This paper proposes a novel approach to improve accuracy of weakly-supervised change detection for multispectral Earth Observation (EO) imagery. The method is based on the idea to use an initial small-size EO labeled dataset to generate a larger set of virtual data. We have considered two variants of virtual data-generators based on the general architecture called Generative Adversarial Network (GAN): MLPGAN and LSGAN. The resulting virtual dataset is used to train a simple Long Short-Term Memory (LSTM) classifier. The proposed method is evaluated using the Mexico dataset acquired by the Thematic Mapper (TM) sensor of the Landsat 5 satellite. For each acquisition date, two spectral bands are considered (B4, B5). The two images have been acquired in April 2000 and May 2002, respectively. We have evaluated the change detection performances (OA, Kappa, MAR, and FAR) using two virtual data generators corresponding to considered GAN architectures. As a benchmark method, we have considered the case when the LSTM classifier is trained with the original small-size dataset without synthetic data generation.
{"title":"A Weakly-Supervised Change Detection for Multispectral Earth Observation Imagery using a Long Short- Term Memory Classifier with a Virtual Training Data Neural Generator","authors":"Ionut Girla, V. Neagoe","doi":"10.1109/comm54429.2022.9817344","DOIUrl":"https://doi.org/10.1109/comm54429.2022.9817344","url":null,"abstract":"This paper proposes a novel approach to improve accuracy of weakly-supervised change detection for multispectral Earth Observation (EO) imagery. The method is based on the idea to use an initial small-size EO labeled dataset to generate a larger set of virtual data. We have considered two variants of virtual data-generators based on the general architecture called Generative Adversarial Network (GAN): MLPGAN and LSGAN. The resulting virtual dataset is used to train a simple Long Short-Term Memory (LSTM) classifier. The proposed method is evaluated using the Mexico dataset acquired by the Thematic Mapper (TM) sensor of the Landsat 5 satellite. For each acquisition date, two spectral bands are considered (B4, B5). The two images have been acquired in April 2000 and May 2002, respectively. We have evaluated the change detection performances (OA, Kappa, MAR, and FAR) using two virtual data generators corresponding to considered GAN architectures. As a benchmark method, we have considered the case when the LSTM classifier is trained with the original small-size dataset without synthetic data generation.","PeriodicalId":118077,"journal":{"name":"2022 14th International Conference on Communications (COMM)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128843986","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：