Pub Date : 2019-07-01DOI: 10.1109/NAECON46414.2019.9057919
B. Jin, Grant W. Bidney, A. Brettin, N. Limberopoulos, I. Anisimov, A. Urbas, Hanyang Li, V. Astratov
Two types of high-index truncated micro-cones are proposed as concentrators for focusing light on pixels of mid-wave infrared (MWIR) focal plane arrays (FPAs). They can be used to miniaturise the pixels for reducing their thermal noise and, potentially, increasing the operation temperature of FPAs without sacrificing their sensitivity.
{"title":"High-index micro-cones for focusing and concentrating light in MWIR focal plane arrays","authors":"B. Jin, Grant W. Bidney, A. Brettin, N. Limberopoulos, I. Anisimov, A. Urbas, Hanyang Li, V. Astratov","doi":"10.1109/NAECON46414.2019.9057919","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9057919","url":null,"abstract":"Two types of high-index truncated micro-cones are proposed as concentrators for focusing light on pixels of mid-wave infrared (MWIR) focal plane arrays (FPAs). They can be used to miniaturise the pixels for reducing their thermal noise and, potentially, increasing the operation temperature of FPAs without sacrificing their sensitivity.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123073464","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9058290
A. Razaque, Fathi H. Amsaad, Cai Mengjie Cherry, Lin Jiahui Linda, A. Oun
Many researchers focus on using smart phone as toolbox, such as measuring height via barometer on smart phone. However, none of them compared the error rate that exists in different brands of mobile phones (e.g., Samsung, HUAWEI and iPhone). This paper carries out sets of control experiments to acquire plenty of data to analyze and contrast the error rate in those mobile systems. This paper uses elementary effects (EE) method to recognize variables that are non-influential. EE method is also used to sort affecting factors such as temperature and humidity, according to factors’ significance. Furthermore, this paper minimizes the burst affecting variable random error by using a method called Filtering Burst Error and Random Error Process (FBEREP) to ensure the error rate. The validation is conducted using real smart phones. Based on the result, it can be observed that error rate is controlled within 2%.
{"title":"Smart Phone as Toolbox for Height Measurement","authors":"A. Razaque, Fathi H. Amsaad, Cai Mengjie Cherry, Lin Jiahui Linda, A. Oun","doi":"10.1109/NAECON46414.2019.9058290","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9058290","url":null,"abstract":"Many researchers focus on using smart phone as toolbox, such as measuring height via barometer on smart phone. However, none of them compared the error rate that exists in different brands of mobile phones (e.g., Samsung, HUAWEI and iPhone). This paper carries out sets of control experiments to acquire plenty of data to analyze and contrast the error rate in those mobile systems. This paper uses elementary effects (EE) method to recognize variables that are non-influential. EE method is also used to sort affecting factors such as temperature and humidity, according to factors’ significance. Furthermore, this paper minimizes the burst affecting variable random error by using a method called Filtering Burst Error and Random Error Process (FBEREP) to ensure the error rate. The validation is conducted using real smart phones. Based on the result, it can be observed that error rate is controlled within 2%.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126850448","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9058104
Grant W. Bidney, A. Brettin, B. Jin, N. Limberopoulos, I. Anisimov, Hanyang Li, A. Maslov, V. Astratov
It is shown that the resolution of cellphone-based microscopy can be improved with the use of millimeter-scale ball lenses with index n=1.85 placed in contact with fluorescent (FL) objects represented by dye-doped microspheres. Using the virtual imaging through such ball lenses, the resolution improvement from ~35 µm to ~2 µm is achieved over a field-of-view of about 0.1 × 0.1 mm2. Further improvement of resolution and field-of-view can be expected with optimization of parameters of the contact ball lenses that can lead to wide applications of the proposed technology in mobile biomedical imaging.
{"title":"Improving Cellphone Microscopy Imaging with Contact Ball Lenses","authors":"Grant W. Bidney, A. Brettin, B. Jin, N. Limberopoulos, I. Anisimov, Hanyang Li, A. Maslov, V. Astratov","doi":"10.1109/NAECON46414.2019.9058104","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9058104","url":null,"abstract":"It is shown that the resolution of cellphone-based microscopy can be improved with the use of millimeter-scale ball lenses with index n=1.85 placed in contact with fluorescent (FL) objects represented by dye-doped microspheres. Using the virtual imaging through such ball lenses, the resolution improvement from ~35 µm to ~2 µm is achieved over a field-of-view of about 0.1 × 0.1 mm2. Further improvement of resolution and field-of-view can be expected with optimization of parameters of the contact ball lenses that can lead to wide applications of the proposed technology in mobile biomedical imaging.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"106 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116012174","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9058217
Jayan Kant Duggal, M. El-Sharkawy
CNNs is the foundation for deep learning and computer vision domain enabling applications such as autonomous driving, face recognition, automatic radiology image reading, etc. But, CNN is a algorithm which is memory and computationally intensive. DSE of neural networks and compression techniques have made convolution neural networks memory and computationally efficient. It improved the CNN architectures and made it more suitable to implement on real-time embedded systems. This paper proposes an efficient and a compact CNN to ameliorate the performance of existing CNN architectures. The intuition behind this proposed architecture is to supplant convolution layers with a more sophisticated block module and to develop a compact architecture with a competitive accuracy. Further, explores the bottleneck module and squeezenext basic block structure. The state-of-the-art squeezenext baseline architecture is used as a foundation to recreate and propose a high performance squeezenext architecture. The proposed architecture is further trained on the CIFAR-10 dataset from scratch. All the training and testing results are visualized with live loss and accuracy graphs. Focus of this paper is to make an adaptable and a flexible model for efficient CNN performance which can perform better with the minimum tradeoff between model accuracy, size, and speed. Finally, the conclusion is made that the performance of CNN can be improved by developing an architecture for a specific dataset. The purpose of this paper is to introduce and propose high performance squeezenext for CIFAR-10.
{"title":"High Performance SqueezeNext for CIFAR-10","authors":"Jayan Kant Duggal, M. El-Sharkawy","doi":"10.1109/NAECON46414.2019.9058217","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9058217","url":null,"abstract":"CNNs is the foundation for deep learning and computer vision domain enabling applications such as autonomous driving, face recognition, automatic radiology image reading, etc. But, CNN is a algorithm which is memory and computationally intensive. DSE of neural networks and compression techniques have made convolution neural networks memory and computationally efficient. It improved the CNN architectures and made it more suitable to implement on real-time embedded systems. This paper proposes an efficient and a compact CNN to ameliorate the performance of existing CNN architectures. The intuition behind this proposed architecture is to supplant convolution layers with a more sophisticated block module and to develop a compact architecture with a competitive accuracy. Further, explores the bottleneck module and squeezenext basic block structure. The state-of-the-art squeezenext baseline architecture is used as a foundation to recreate and propose a high performance squeezenext architecture. The proposed architecture is further trained on the CIFAR-10 dataset from scratch. All the training and testing results are visualized with live loss and accuracy graphs. Focus of this paper is to make an adaptable and a flexible model for efficient CNN performance which can perform better with the minimum tradeoff between model accuracy, size, and speed. Finally, the conclusion is made that the performance of CNN can be improved by developing an architecture for a specific dataset. The purpose of this paper is to introduce and propose high performance squeezenext for CIFAR-10.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116093068","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9057803
Yaping Cai, Massimo Cancian, M. D’Arpino, G. Rizzoni
A critical issue in large-scale battery pack is the capability of assessing the impact of cell-to-cell variation on the pack/module performance. The inhomogeneity of cell parameters is mainly due to manufacturing tolerances, cell degradation, and temperature gradients, and leads to unbalanced current and voltage distribution in the pack. In this paper, a generalized equivalent circuit model is proposed that effectively allows estimating the current and voltage differences between cells/strings in a pack as function of the system architecture and parameters. The model is based on a simplified equivalent circuit model that separates the healthy cells from the ones affected by parameters variation. The proposed methodology has been applied to several pack configurations, typically used for automotive and aerospace applications, and demonstrates high benefits for large scale systems due to the reduced computational effort. The proposed model can be used during the design of the battery pack for estimating the impact of degradation and thermal unbalance in a pack. Look-up-tables can be carried out to describe the behavior of battery pack at steady-state conditions and can be used by the Battery Management System (BMS) for on-line estimation of the operating conditions of the cells.
{"title":"A generalized equivalent circuit model for large-scale battery packs with cell-to-cell variation","authors":"Yaping Cai, Massimo Cancian, M. D’Arpino, G. Rizzoni","doi":"10.1109/NAECON46414.2019.9057803","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9057803","url":null,"abstract":"A critical issue in large-scale battery pack is the capability of assessing the impact of cell-to-cell variation on the pack/module performance. The inhomogeneity of cell parameters is mainly due to manufacturing tolerances, cell degradation, and temperature gradients, and leads to unbalanced current and voltage distribution in the pack. In this paper, a generalized equivalent circuit model is proposed that effectively allows estimating the current and voltage differences between cells/strings in a pack as function of the system architecture and parameters. The model is based on a simplified equivalent circuit model that separates the healthy cells from the ones affected by parameters variation. The proposed methodology has been applied to several pack configurations, typically used for automotive and aerospace applications, and demonstrates high benefits for large scale systems due to the reduced computational effort. The proposed model can be used during the design of the battery pack for estimating the impact of degradation and thermal unbalance in a pack. Look-up-tables can be carried out to describe the behavior of battery pack at steady-state conditions and can be used by the Battery Management System (BMS) for on-line estimation of the operating conditions of the cells.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"126 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114543075","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9057872
Ying Li, Sihao Ding
Lane filtering is a necessary process applied after lane detection in autonomous vehicle applications. The unprocessed result from lane detection can usually be noisy. The length and position of the detected lanes are often changing abruptly across frames due to imperfect detection, which would introduce noise to downstream processes. In order to obtain steady lane detection result, we develop a new method to filter the raw output of detection. We first perform a prepossessing to filter out large obvious inconsistency. A compact lane representation is designed, to convert the various length into fixed-dimension vector representation. The general shape of the lanes is kept while a low computational complexity is maintained. We then apply the Kalman filter to perform filtering in temporal domain, and estimate the location of the lanes. Qualitative and quantitative experiments are conducted on real data collected from vehicle driving in urban area, showing improved results compared to unprocessed lane detection results.
{"title":"Fast Lane Filtering for Autonomous Vehicle","authors":"Ying Li, Sihao Ding","doi":"10.1109/NAECON46414.2019.9057872","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9057872","url":null,"abstract":"Lane filtering is a necessary process applied after lane detection in autonomous vehicle applications. The unprocessed result from lane detection can usually be noisy. The length and position of the detected lanes are often changing abruptly across frames due to imperfect detection, which would introduce noise to downstream processes. In order to obtain steady lane detection result, we develop a new method to filter the raw output of detection. We first perform a prepossessing to filter out large obvious inconsistency. A compact lane representation is designed, to convert the various length into fixed-dimension vector representation. The general shape of the lanes is kept while a low computational complexity is maintained. We then apply the Kalman filter to perform filtering in temporal domain, and estimate the location of the lanes. Qualitative and quantitative experiments are conducted on real data collected from vehicle driving in urban area, showing improved results compared to unprocessed lane detection results.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117006970","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9058166
Joshua K. Woodward, M. Rizkalla
A software approach was utilized to study the gain stability over a long frequency range in the order of 10 decades. The approach addresses the input and output power representations on the Linvill plane over a given frequency range. The study considers the mismatching approach by adding input and output admittances to keep the transducer gain of the system stable with a high stern stability factor. The study was extended to cover optimum noise figure over a frequency range of 10Hz-100GHz. The new selection technique uses a MATLAB script to calculate the Y parameters of a common emitter BJT amplifier for a wide frequency band. From the Y parameters, calculations for the optimum shunt admittances for the source and load are made based on a customizable threshold for the Stern stability factor. Additional calculations are made to optimize the source conductance to achieve the lowest noise figure. The program then calculates the noise figure at optimum stability and the stability factor at optimum noise. Initially, the device was instable for 2 decades within the MHz range, and after applying the described approach, the device was stable for the entire 10Hz100GHz frequency range. The transducer gain of the stabilized device was lower than that of the initial instable device.
{"title":"Low Noise High Stability Amplifiers over Very High Frequency Range Using Mismatching Approach within Linvill Plane Simulation","authors":"Joshua K. Woodward, M. Rizkalla","doi":"10.1109/NAECON46414.2019.9058166","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9058166","url":null,"abstract":"A software approach was utilized to study the gain stability over a long frequency range in the order of 10 decades. The approach addresses the input and output power representations on the Linvill plane over a given frequency range. The study considers the mismatching approach by adding input and output admittances to keep the transducer gain of the system stable with a high stern stability factor. The study was extended to cover optimum noise figure over a frequency range of 10Hz-100GHz. The new selection technique uses a MATLAB script to calculate the Y parameters of a common emitter BJT amplifier for a wide frequency band. From the Y parameters, calculations for the optimum shunt admittances for the source and load are made based on a customizable threshold for the Stern stability factor. Additional calculations are made to optimize the source conductance to achieve the lowest noise figure. The program then calculates the noise figure at optimum stability and the stability factor at optimum noise. Initially, the device was instable for 2 decades within the MHz range, and after applying the described approach, the device was stable for the entire 10Hz100GHz frequency range. The transducer gain of the stabilized device was lower than that of the initial instable device.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129808317","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9058150
W. Wilson, J. P. Moore, Katelyn R. Brinker
As the drive for increased aviation efficiency continues to push airframe designs toward lighter, more flexible configurations, the technology limits for modelling and monitoring airframe and control surface flutter events continue to be a challenge. Presented is a new flutter sensing technique that configures a commercial microwave reflectometer as a monostatic radar to measure low frequency vibrations typical of subsonic aircraft structural flutter. The stepped frequency continuous wave (SFCW) system with bandwidths of 90 MHz and 7.5 GHz was used to measure vibration frequencies from 1Hz to 125 Hz in a carbon fiber–reinforced polymer (CFRP) composite panel. In addition to the SFCW mode, a single frequency mode has been used to determine the velocity of vibration from 1Hz to 75 Hz. The data presented demonstrates that commercial reflectometers can be used to measure flutter-induced aircraft vibrations.
{"title":"The use of a Reflectometer as a Monostatic Radar for Measuring Aircraft Structural Flutter","authors":"W. Wilson, J. P. Moore, Katelyn R. Brinker","doi":"10.1109/NAECON46414.2019.9058150","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9058150","url":null,"abstract":"As the drive for increased aviation efficiency continues to push airframe designs toward lighter, more flexible configurations, the technology limits for modelling and monitoring airframe and control surface flutter events continue to be a challenge. Presented is a new flutter sensing technique that configures a commercial microwave reflectometer as a monostatic radar to measure low frequency vibrations typical of subsonic aircraft structural flutter. The stepped frequency continuous wave (SFCW) system with bandwidths of 90 MHz and 7.5 GHz was used to measure vibration frequencies from 1Hz to 125 Hz in a carbon fiber–reinforced polymer (CFRP) composite panel. In addition to the SFCW mode, a single frequency mode has been used to determine the velocity of vibration from 1Hz to 75 Hz. The data presented demonstrates that commercial reflectometers can be used to measure flutter-induced aircraft vibrations.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128421772","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9057847
A. Samiei, Sarah Ismail, Liang Sun
In the context of autonomy for Unmanned Aerial Vehicles (UAVs), task allocation plays a significant role for collaborative UAVs to make decisions in dynamic environments. This paper presents a novel Hungarian-based approach to challenging multi-task allocation (MTA) problems, where the number of UAVs is smaller than the number of tasks. We developed the Cluster-Based Hungarian Algorithm (CBHA), in which (1) tasks are grouped such that the number of UAVs is the same as the number of task groups, (2) the original Hungarian algorithm is applied, and (3) an algorithm for travel-salesman-problem (TSP) is applied to perform path planning for individual UAVs. The performance of the proposed CBHA was compared with the Consensus-Based Bundle Algorithm (CBBA) in Monti Carlo simulations, where different numbers of UAVs and tasks were adopted in the scenario of a team of unmanned aerial vehicles traveling through a number of targeting locations. The simulation result shows that the CBHA outperforms CBBA in all cases.
{"title":"Cluster-Based Hungarian Approach to Task Allocation for Unmanned Aerial Vehicles","authors":"A. Samiei, Sarah Ismail, Liang Sun","doi":"10.1109/NAECON46414.2019.9057847","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9057847","url":null,"abstract":"In the context of autonomy for Unmanned Aerial Vehicles (UAVs), task allocation plays a significant role for collaborative UAVs to make decisions in dynamic environments. This paper presents a novel Hungarian-based approach to challenging multi-task allocation (MTA) problems, where the number of UAVs is smaller than the number of tasks. We developed the Cluster-Based Hungarian Algorithm (CBHA), in which (1) tasks are grouped such that the number of UAVs is the same as the number of task groups, (2) the original Hungarian algorithm is applied, and (3) an algorithm for travel-salesman-problem (TSP) is applied to perform path planning for individual UAVs. The performance of the proposed CBHA was compared with the Consensus-Based Bundle Algorithm (CBBA) in Monti Carlo simulations, where different numbers of UAVs and tasks were adopted in the scenario of a team of unmanned aerial vehicles traveling through a number of targeting locations. The simulation result shows that the CBHA outperforms CBBA in all cases.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114341591","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 : 2019-07-01DOI: 10.1109/NAECON46414.2019.9058245
Redha Ali, R. Hardie, Barath Narayanan Narayanan, Supun de Silva
Skin cancer has a significant impact across the world. Melanoma is a malignant form of skin cancer. Skin lesion segmentation is an important step in computer-aided diagnosis (CAD) for automated diagnosis of melanoma. In this paper, we describe our research work and the submission to the International Skin Imaging Collaborations (ISIC) 2018 Challenge in Skin Lesion Analysis Towards Melanoma Detection. We propose Convolutional Neural Network (CNN) based ensemble methods for improving the existing performance of lesion segmentation. The proposed ensemble technique includes VGG19-UNet, DeeplabV3+ and other preprocessing methodologies. Extensive experiments are conducted on the ISIC 2018 challenge dataset to demonstrate the efficacy of the proposed model. For evaluation, we utilize the ISIC 2018 datasets that contains 2,594 dermoscopy images with their ground truth segmentation masks. We randomly divided the dataset into 80% for training and 20% for validation. Our proposed model provided an overall accuracy of 93.6%, average Jaccard Index of 0.815, and dice coefficient of 0.887 on the testing dataset.
皮肤癌在全世界都有重大影响。黑色素瘤是一种恶性皮肤癌。皮肤病灶分割是计算机辅助诊断(CAD)实现黑色素瘤自动诊断的重要步骤。在本文中，我们描述了我们的研究工作，并向国际皮肤成像合作组织(ISIC)提交了2018年皮肤病变分析与黑色素瘤检测的挑战。我们提出了基于卷积神经网络(CNN)的集成方法来改进现有的病灶分割性能。提出的集成技术包括VGG19-UNet、DeeplabV3+等预处理方法。在ISIC 2018挑战数据集上进行了大量实验，以证明所提出模型的有效性。为了进行评估，我们使用了ISIC 2018数据集，该数据集包含2594张皮肤镜图像及其ground truth segmentation mask。我们将数据集随机分为80%用于训练和20%用于验证。我们提出的模型在测试数据集上的总体准确率为93.6%，平均Jaccard指数为0.815，骰子系数为0.887。
{"title":"Deep Learning Ensemble Methods for Skin Lesion Analysis towards Melanoma Detection","authors":"Redha Ali, R. Hardie, Barath Narayanan Narayanan, Supun de Silva","doi":"10.1109/NAECON46414.2019.9058245","DOIUrl":"https://doi.org/10.1109/NAECON46414.2019.9058245","url":null,"abstract":"Skin cancer has a significant impact across the world. Melanoma is a malignant form of skin cancer. Skin lesion segmentation is an important step in computer-aided diagnosis (CAD) for automated diagnosis of melanoma. In this paper, we describe our research work and the submission to the International Skin Imaging Collaborations (ISIC) 2018 Challenge in Skin Lesion Analysis Towards Melanoma Detection. We propose Convolutional Neural Network (CNN) based ensemble methods for improving the existing performance of lesion segmentation. The proposed ensemble technique includes VGG19-UNet, DeeplabV3+ and other preprocessing methodologies. Extensive experiments are conducted on the ISIC 2018 challenge dataset to demonstrate the efficacy of the proposed model. For evaluation, we utilize the ISIC 2018 datasets that contains 2,594 dermoscopy images with their ground truth segmentation masks. We randomly divided the dataset into 80% for training and 20% for validation. Our proposed model provided an overall accuracy of 93.6%, average Jaccard Index of 0.815, and dice coefficient of 0.887 on the testing dataset.","PeriodicalId":193529,"journal":{"name":"2019 IEEE National Aerospace and Electronics Conference (NAECON)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116508784","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：