Pub Date : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00074
Ioannis Dimolitsas, Dimitrios Spatharakis, Dimitrios Dechouniotis, Anastasios Zafeiropoulos, S. Papavassiliou
The dynamic workload demands of smart city applications hosted on edge infrastructures require the development of advanced scaling mechanisms. Recent studies proposed single-application autoscaling solutions based on various technical approaches. However, for edge infrastructures with limited resource availability, it is essential to simultaneously manage heterogeneous application requirements, aiming at optimal resource allocation and minimal operational costs. This study introduces a multi-application hierarchical autoscaling framework for Kubernetes Edge Clusters. An application-based mechanism nominates the best applications’ deployments based on workload prediction and several criteria that guarantee the application’s performance while minimizing the infrastructure provider’s cost. For the joint application orchestration, an aggregation mechanism composes the candidate scaling solutions for the cluster. Then, a cluster autoscaling mechanism, based on the Analytic Hierarchy Process, undertakes the cluster’s scaling decision to optimize the resource allocation and energy consumption of the cluster. The evaluation illustrates the benefits of the proposed scaling strategy, achieving significant improvement in the average allocated resources and energy consumption compared to single-application approaches.
{"title":"Multi-Application Hierarchical Autoscaling for Kubernetes Edge Clusters","authors":"Ioannis Dimolitsas, Dimitrios Spatharakis, Dimitrios Dechouniotis, Anastasios Zafeiropoulos, S. Papavassiliou","doi":"10.1109/SMARTCOMP58114.2023.00074","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00074","url":null,"abstract":"The dynamic workload demands of smart city applications hosted on edge infrastructures require the development of advanced scaling mechanisms. Recent studies proposed single-application autoscaling solutions based on various technical approaches. However, for edge infrastructures with limited resource availability, it is essential to simultaneously manage heterogeneous application requirements, aiming at optimal resource allocation and minimal operational costs. This study introduces a multi-application hierarchical autoscaling framework for Kubernetes Edge Clusters. An application-based mechanism nominates the best applications’ deployments based on workload prediction and several criteria that guarantee the application’s performance while minimizing the infrastructure provider’s cost. For the joint application orchestration, an aggregation mechanism composes the candidate scaling solutions for the cluster. Then, a cluster autoscaling mechanism, based on the Analytic Hierarchy Process, undertakes the cluster’s scaling decision to optimize the resource allocation and energy consumption of the cluster. The evaluation illustrates the benefits of the proposed scaling strategy, achieving significant improvement in the average allocated resources and energy consumption compared to single-application approaches.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121460874","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00019
Md Abdullah Al Rumon, Veeturi Suparna, Mehmet Seckin, Dhaval Solanki, K. Mankodiya
Breathing exercises are gaining attention in managing anxiety and stress in daily life. Diaphragmatic breathing, in particular, fosters tranquility for both body and mind. Existing methods, such as meditation, yoga, and medical devices for guided breathing, often require expert guidance, complex instruments, cumbersome devices, and sticky electrodes. To address these challenges, we present Nisshash, an IoT-based smart T-shirt offering a personalized solution for regulated breathing exercises. Nisshash is embedded with three-channel e-textile respiration sensors and a tailored analog front-end (AFE) board to simultaneously monitor respiration rate (RR) and heart rate (HR). In this work, we seamlessly integrate soft textile sensors into a T-shirt and develop a detachable and Wi-Fi-enabled (2.4GHz) bio-instrumentation board, creating a pervasive wireless system (WPS) for guided breathing exercises (GBE). The system features an intuitive graphical user interface (GUI) and a seamless IoT-based control and computing system (CCS). It offers real-time instructions for inhaling and exhaling at various breathing speeds, including slow, normal, and fast breathing. Functions such as filtering, peak detections for respiration, and heart rate analysis are computed conjointly at the sender and receiver ends. We utilized the Pan-Tompkins and custom algorithms to calculate HR and RR from the filtered time-series signals. We conducted a study with 10 healthy adult participants who wore the T-shirt and performed guided breathing exercises. The average respiration event (inhale-exhale) detection accuracy was ≈98%. We validated the recorded HR against the 3-lead standard ECG monitoring device, achieving an accuracy of ≈99%. The RR-HR correlation analysis showed an R square value of 0.987. Collectively, these results demonstrate Nisshash’s potential as a personal guided breathing exercise solution.
呼吸练习在管理日常生活中的焦虑和压力方面越来越受到关注。尤其是横膈膜呼吸法，能促进身心的平静。现有的方法，如冥想、瑜伽和引导呼吸的医疗设备，通常需要专家指导、复杂的仪器、笨重的设备和粘性电极。为了应对这些挑战，我们推出了Nisshash，一款基于物联网的智能t恤，为调节呼吸练习提供个性化解决方案。Nisshash内置了三通道电子纺织呼吸传感器和定制的模拟前端(AFE)板，可同时监测呼吸速率(RR)和心率(HR)。在这项工作中，我们将柔软的纺织品传感器无缝集成到t恤中，并开发了一个可拆卸的、支持wi - fi (2.4GHz)的生物仪器板，为引导呼吸练习(GBE)创建了一个普适无线系统(WPS)。该系统具有直观的图形用户界面(GUI)和无缝的基于物联网的控制和计算系统(CCS)。它为在各种呼吸速度下吸气和呼气提供实时指示，包括慢的、正常的和快速的呼吸。诸如滤波、呼吸峰值检测和心率分析等功能在发送端和接收端联合计算。我们利用Pan-Tompkins和自定义算法从滤波后的时间序列信号中计算HR和RR。我们对10名健康的成年参与者进行了一项研究，他们穿着t恤，进行有指导的呼吸练习。平均呼吸事件(吸入-呼出)检测准确率≈98%。我们将记录的HR与3导联标准心电监护装置进行验证，准确率达到约99%。RR-HR相关分析R平方值为0.987。总的来说，这些结果证明了Nisshash作为个人指导呼吸练习解决方案的潜力。
{"title":"Nisshash: Design of An IoT-based Smart T-Shirt for Guided Breathing Exercises","authors":"Md Abdullah Al Rumon, Veeturi Suparna, Mehmet Seckin, Dhaval Solanki, K. Mankodiya","doi":"10.1109/SMARTCOMP58114.2023.00019","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00019","url":null,"abstract":"Breathing exercises are gaining attention in managing anxiety and stress in daily life. Diaphragmatic breathing, in particular, fosters tranquility for both body and mind. Existing methods, such as meditation, yoga, and medical devices for guided breathing, often require expert guidance, complex instruments, cumbersome devices, and sticky electrodes. To address these challenges, we present Nisshash, an IoT-based smart T-shirt offering a personalized solution for regulated breathing exercises. Nisshash is embedded with three-channel e-textile respiration sensors and a tailored analog front-end (AFE) board to simultaneously monitor respiration rate (RR) and heart rate (HR). In this work, we seamlessly integrate soft textile sensors into a T-shirt and develop a detachable and Wi-Fi-enabled (2.4GHz) bio-instrumentation board, creating a pervasive wireless system (WPS) for guided breathing exercises (GBE). The system features an intuitive graphical user interface (GUI) and a seamless IoT-based control and computing system (CCS). It offers real-time instructions for inhaling and exhaling at various breathing speeds, including slow, normal, and fast breathing. Functions such as filtering, peak detections for respiration, and heart rate analysis are computed conjointly at the sender and receiver ends. We utilized the Pan-Tompkins and custom algorithms to calculate HR and RR from the filtered time-series signals. We conducted a study with 10 healthy adult participants who wore the T-shirt and performed guided breathing exercises. The average respiration event (inhale-exhale) detection accuracy was ≈98%. We validated the recorded HR against the 3-lead standard ECG monitoring device, achieving an accuracy of ≈99%. The RR-HR correlation analysis showed an R square value of 0.987. Collectively, these results demonstrate Nisshash’s potential as a personal guided breathing exercise solution.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115929416","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00021
Zahid Hasan, A. Faridee, Masud Ahmed, Shibi Ayyanar, Nirmalya Roy
Remote Photoplethysmography (rPPG) systems offer contactless, low-cost, and ubiquitous heart rate (HR) monitoring by leveraging the skin-tissue blood volumetric variation-induced reflection. However, collecting large-scale time-synchronized rPPG data is costly and impedes the development of generalized end-to-end deep learning (DL) rPPG models to perform under diverse scenarios. We formulate the rPPG estimation as a generative task of recovering time-series PPG from facial videos and propose SrPPG, a novel semi-supervised adversarial learning framework using heterogeneous, asynchronous, and noisy rPPG data. More specifically, we develop a novel encoder-decoder architecture, where rPPG features are learned from video in a self-supervised manner (encoder) to reconstruct the time-series PPG (decoder/generator) with physics-inspired novel temporal consistency regularization. The generated PPG is scrutinized against the real rPPG signals by a frequency-class conditioned discriminator, forming a generative adversarial network. Thus, SrPPG generates samples without point-wise supervision, alleviating the need for time-synchronized data collection. We experiment and validate SrPPG by amassing three public datasets in heterogeneous settings. SrPPG outperforms both supervised and self-supervised state-of-the-art methods in HR estimation across all datasets without any time-synchronous rPPG data. We also perform extensive experiments to study the optimal generative setting (architecture, joint optimization) and provide insight into the SrPPG behavior.
{"title":"SrPPG: Semi-Supervised Adversarial Learning for Remote Photoplethysmography with Noisy Data","authors":"Zahid Hasan, A. Faridee, Masud Ahmed, Shibi Ayyanar, Nirmalya Roy","doi":"10.1109/SMARTCOMP58114.2023.00021","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00021","url":null,"abstract":"Remote Photoplethysmography (rPPG) systems offer contactless, low-cost, and ubiquitous heart rate (HR) monitoring by leveraging the skin-tissue blood volumetric variation-induced reflection. However, collecting large-scale time-synchronized rPPG data is costly and impedes the development of generalized end-to-end deep learning (DL) rPPG models to perform under diverse scenarios. We formulate the rPPG estimation as a generative task of recovering time-series PPG from facial videos and propose SrPPG, a novel semi-supervised adversarial learning framework using heterogeneous, asynchronous, and noisy rPPG data. More specifically, we develop a novel encoder-decoder architecture, where rPPG features are learned from video in a self-supervised manner (encoder) to reconstruct the time-series PPG (decoder/generator) with physics-inspired novel temporal consistency regularization. The generated PPG is scrutinized against the real rPPG signals by a frequency-class conditioned discriminator, forming a generative adversarial network. Thus, SrPPG generates samples without point-wise supervision, alleviating the need for time-synchronized data collection. We experiment and validate SrPPG by amassing three public datasets in heterogeneous settings. SrPPG outperforms both supervised and self-supervised state-of-the-art methods in HR estimation across all datasets without any time-synchronous rPPG data. We also perform extensive experiments to study the optimal generative setting (architecture, joint optimization) and provide insight into the SrPPG behavior.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116823340","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00087
Gautam Mundewadi, R. Wolski, C. Krintz
To cultivate healthy plants and high crop yields, growers must be able to measure soil moisture and irrigate accordingly. Errors in soil moisture measurements can lead to irrigation mismanagement with costly consequences. In this paper, we present a new approach to smart computing for irrigation management to address these challenges at a lower cost. We calibrate low cost, low precision soil moisture sensors to more accurately distinguish wet from dry soils using high cost, high precision Davis Instrument sensors. We investigate different modeling techniques including the natural log of the odds ratio (Log-odds), Monte Carlo simulation, and linear regression to distinguish between wet and moist soils and to establish a trustworthy threshold between these two moisture states. We have also developed a new smartphone application that simplifies the process of data collection and implements our analysis approach. The application is extensible by others and provides growers with low cost, data-driven decision support for irrigation. We implement our approach for UCSB’s Edible Campus student farm and empirically evaluate it using multiple test beds. Our results show an accuracy rate of 91% and lowers costs by 4x per deployment, making it useful for gardeners and farmers alike.
{"title":"Data Acquisition and Analysis for Improving the Utility of Low Cost Soil Moisture Sensors","authors":"Gautam Mundewadi, R. Wolski, C. Krintz","doi":"10.1109/SMARTCOMP58114.2023.00087","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00087","url":null,"abstract":"To cultivate healthy plants and high crop yields, growers must be able to measure soil moisture and irrigate accordingly. Errors in soil moisture measurements can lead to irrigation mismanagement with costly consequences. In this paper, we present a new approach to smart computing for irrigation management to address these challenges at a lower cost. We calibrate low cost, low precision soil moisture sensors to more accurately distinguish wet from dry soils using high cost, high precision Davis Instrument sensors. We investigate different modeling techniques including the natural log of the odds ratio (Log-odds), Monte Carlo simulation, and linear regression to distinguish between wet and moist soils and to establish a trustworthy threshold between these two moisture states. We have also developed a new smartphone application that simplifies the process of data collection and implements our analysis approach. The application is extensible by others and provides growers with low cost, data-driven decision support for irrigation. We implement our approach for UCSB’s Edible Campus student farm and empirically evaluate it using multiple test beds. Our results show an accuracy rate of 91% and lowers costs by 4x per deployment, making it useful for gardeners and farmers alike.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116140441","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00062
Ayanfeoluwa Oluyomi
Utility companies rely on accurate data (e.g. energy or water usage) to monitor and determine the pricing and distribution of resources. In most cities, a utility company tends to service a large number of houses in that city. These houses may not be concentrated in a neighborhood and this can make it difficult for them to manage because of the different patterns of water usage that exist in various neighborhoods. An adversary can take advantage of this by injecting false data into a subset of the houses such that the difference will not be noticed by the utility. False data injection (FDI) attacks compromise the integrity of the data, leading to inaccurate decision-making and potential water resource wastage. To address this problem, this research aims to study a clustering algorithm that leverages graph theory to cluster houses with similar water usage patterns in a city. After this, an FDI detection model is run on each cluster to identify any attack.
{"title":"Detecting False Data Injection in a Large-Scale Water Distribution Network","authors":"Ayanfeoluwa Oluyomi","doi":"10.1109/SMARTCOMP58114.2023.00062","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00062","url":null,"abstract":"Utility companies rely on accurate data (e.g. energy or water usage) to monitor and determine the pricing and distribution of resources. In most cities, a utility company tends to service a large number of houses in that city. These houses may not be concentrated in a neighborhood and this can make it difficult for them to manage because of the different patterns of water usage that exist in various neighborhoods. An adversary can take advantage of this by injecting false data into a subset of the houses such that the difference will not be noticed by the utility. False data injection (FDI) attacks compromise the integrity of the data, leading to inaccurate decision-making and potential water resource wastage. To address this problem, this research aims to study a clustering algorithm that leverages graph theory to cluster houses with similar water usage patterns in a city. After this, an FDI detection model is run on each cluster to identify any attack.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134522306","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00048
Xiaoming Guo, Xiao Hong
The paper introduces a deep reinforcement learning model for a special scenario in future smart transportation. The scenario describes a mobile edge computing platform hosted by a group of self-organized connected vehicles for sharing computation resources. The presented DQN model is to solve the trade-offs between the computing capability and the traffic state. Results show the existence of the trade-off and the need for future research in a few areas.
{"title":"DQN for Smart Transportation Supporting V2V Mobile Edge Computing","authors":"Xiaoming Guo, Xiao Hong","doi":"10.1109/SMARTCOMP58114.2023.00048","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00048","url":null,"abstract":"The paper introduces a deep reinforcement learning model for a special scenario in future smart transportation. The scenario describes a mobile edge computing platform hosted by a group of self-organized connected vehicles for sharing computation resources. The presented DQN model is to solve the trade-offs between the computing capability and the traffic state. Results show the existence of the trade-off and the need for future research in a few areas.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132373921","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00053
Elijah Spicer, S. Baidya
Modern cyber-physical systems use deep-learning based algorithms for many applications for intelligent decision-making. Many of these systems are resource-constrained due to small form factor or finite energy budget. However, these systems often use multiple deep-learning algorithms simultaneously for a given mission or task. Due to the diverse nature of the algorithms and their performance needs, we need to allocate optimal software and hardware resources for their coexistence. To this aim, in this paper, we study and evaluate the performance tradeoff which will enable the users to choose the size and complexity of the deep learning models, the capacity of the device and also the software framework. With real-world experiments with a wide range of hardware and software, we demonstrate and evaluate the performance of the coexisting deep neural networks (DNN) based applications.
{"title":"Performance Tradeoff in DNN-based Coexisting Applications in Resource-Constrained Cyber-Physical Systems","authors":"Elijah Spicer, S. Baidya","doi":"10.1109/SMARTCOMP58114.2023.00053","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00053","url":null,"abstract":"Modern cyber-physical systems use deep-learning based algorithms for many applications for intelligent decision-making. Many of these systems are resource-constrained due to small form factor or finite energy budget. However, these systems often use multiple deep-learning algorithms simultaneously for a given mission or task. Due to the diverse nature of the algorithms and their performance needs, we need to allocate optimal software and hardware resources for their coexistence. To this aim, in this paper, we study and evaluate the performance tradeoff which will enable the users to choose the size and complexity of the deep learning models, the capacity of the device and also the software framework. With real-world experiments with a wide range of hardware and software, we demonstrate and evaluate the performance of the coexisting deep neural networks (DNN) based applications.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133393313","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00032
Masud Ahmed, Zahid Hasan, Tim Yingling, Eric O'Leary, S. Purushotham, Suya You, Nirmalya Roy
The annotation load for a new dataset has been greatly decreased using domain adaptation based semantic segmentation, which iteratively constructs pseudo labels on unlabeled target data and retrains the network. However, realistic segmentation datasets are often imbalanced, with pseudo-labels tending to favor certain "head" classes while neglecting other "tail" classes. This can lead to an inaccurate and noisy mask. To address this issue, we propose a novel hard sample mining strategy for an active domain adaptation based semantic segmentation network, with the aim of automatically selecting a small subset of labeled target data to fine-tune the network. By calculating class-wise entropy, we are able to rank the difficulty level of different samples. We use a fusion of focal loss and regional mutual information loss instead of cross-entropy loss for the domain adaptation based semantic segmentation network. Our entire framework has been implemented in real-time using the Robotics Operating System (ROS) with a server PC and a small Unmanned Ground Vehicle (UGV) known as the ROSbot2.0 Pro. This implementation allows ROSbot2.0 Pro to access any type of data at any time, enabling it to perform a variety of tasks with ease. Our approach has been thoroughly evaluated through a series of extensive experiments, which demonstrate its superior performance compared to existing state-of-the-art methods. Remarkably, by using just 20% of hard samples for fine-tuning, our network has achieved a level of performance that is comparable (≈88%) to that of a fully supervised approach, with mIOU scores of 60.51% in the In-house dataset.
{"title":"An Online Continuous Semantic Segmentation Framework With Minimal Labeling Efforts","authors":"Masud Ahmed, Zahid Hasan, Tim Yingling, Eric O'Leary, S. Purushotham, Suya You, Nirmalya Roy","doi":"10.1109/SMARTCOMP58114.2023.00032","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00032","url":null,"abstract":"The annotation load for a new dataset has been greatly decreased using domain adaptation based semantic segmentation, which iteratively constructs pseudo labels on unlabeled target data and retrains the network. However, realistic segmentation datasets are often imbalanced, with pseudo-labels tending to favor certain \"head\" classes while neglecting other \"tail\" classes. This can lead to an inaccurate and noisy mask. To address this issue, we propose a novel hard sample mining strategy for an active domain adaptation based semantic segmentation network, with the aim of automatically selecting a small subset of labeled target data to fine-tune the network. By calculating class-wise entropy, we are able to rank the difficulty level of different samples. We use a fusion of focal loss and regional mutual information loss instead of cross-entropy loss for the domain adaptation based semantic segmentation network. Our entire framework has been implemented in real-time using the Robotics Operating System (ROS) with a server PC and a small Unmanned Ground Vehicle (UGV) known as the ROSbot2.0 Pro. This implementation allows ROSbot2.0 Pro to access any type of data at any time, enabling it to perform a variety of tasks with ease. Our approach has been thoroughly evaluated through a series of extensive experiments, which demonstrate its superior performance compared to existing state-of-the-art methods. Remarkably, by using just 20% of hard samples for fine-tuning, our network has achieved a level of performance that is comparable (≈88%) to that of a fully supervised approach, with mIOU scores of 60.51% in the In-house dataset.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128671127","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 : 2023-06-01DOI: 10.1109/SMARTCOMP58114.2023.00088
Thai Thao Nguyen, Jesse Parron, Omar Obidat, A. Tuininga, Weitian Wang
As robotics and artificial intelligence (AI) technologies have become increasingly relevant over the past couple of years, they will inevitably be key components for industries of all aspects which continue to expand to technological solutions. Particularly, the agricultural industry has progressed to using such means to minimize human involvement and reduce tasks that are time-consuming and costly. Motivated by this, we developed a robot-assisted crop maturity recognition and harvest system to accurately classify and detect the stages of ripeness the crops are in—ripe, medium ripe, and not ripe. Our proposed approach integrates computer vision, image processing, collaborative robotics, and a subcategory of artificial intelligence—transfer learning. The transfer learning-based model is trained to classify and recognize the crop in its maturity stages and locate the crop during real-time detection. Experimental results and analysis in real-world robot-assisted smart agriculture environments successfully demonstrated crop ripeness recognition accuracy, proving transfer learning could be utilized to effectively improve the efficiency and productivity of harvesting processes in the agricultural industry. The future work of this study is also discussed.
{"title":"Ready or Not? A Robot-Assisted Crop Harvest Solution in Smart Agriculture Contexts","authors":"Thai Thao Nguyen, Jesse Parron, Omar Obidat, A. Tuininga, Weitian Wang","doi":"10.1109/SMARTCOMP58114.2023.00088","DOIUrl":"https://doi.org/10.1109/SMARTCOMP58114.2023.00088","url":null,"abstract":"As robotics and artificial intelligence (AI) technologies have become increasingly relevant over the past couple of years, they will inevitably be key components for industries of all aspects which continue to expand to technological solutions. Particularly, the agricultural industry has progressed to using such means to minimize human involvement and reduce tasks that are time-consuming and costly. Motivated by this, we developed a robot-assisted crop maturity recognition and harvest system to accurately classify and detect the stages of ripeness the crops are in—ripe, medium ripe, and not ripe. Our proposed approach integrates computer vision, image processing, collaborative robotics, and a subcategory of artificial intelligence—transfer learning. The transfer learning-based model is trained to classify and recognize the crop in its maturity stages and locate the crop during real-time detection. Experimental results and analysis in real-world robot-assisted smart agriculture environments successfully demonstrated crop ripeness recognition accuracy, proving transfer learning could be utilized to effectively improve the efficiency and productivity of harvesting processes in the agricultural industry. The future work of this study is also discussed.","PeriodicalId":163556,"journal":{"name":"2023 IEEE International Conference on Smart Computing (SMARTCOMP)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133936286","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：