Pub Date : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9333017
Tai-Wei Chen, Wei-Liang Lin
This paper using a 3D pose estimator to predict human 3D poses. By combining the pose sequence information as a motion capture, we could reconstruct the human motion in Unity with any appearance. A potential application is collecting a compact human 3D activity dataset.
{"title":"3D Human Motion Reconstruction in Unity With Monocular Camera","authors":"Tai-Wei Chen, Wei-Liang Lin","doi":"10.1109/ISOCC50952.2020.9333017","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9333017","url":null,"abstract":"This paper using a 3D pose estimator to predict human 3D poses. By combining the pose sequence information as a motion capture, we could reconstruct the human motion in Unity with any appearance. A potential application is collecting a compact human 3D activity dataset.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115760398","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9333027
Dohyun Kim, Taeyang Jeong, Eui-Young Chung
In LTE(Long-Term Evolution) system, UE(User Equipment) performs synchronization processing with a specific cell to communicate. In that processing, the UE uses a matched filter to filter PSS(Primary Synchronization Signal) from downlink signals sent from the cell. There are various ways to design such a matched filter. In the most native design of the matched filter, the number of multipliers is required as much as a number of the taps which means filter length. If resources are limited, that is a very inefficient design approach. Therefore, we proposed filter design method to significantly reduce the number of multipliers in the matched filter by utilizing the difference of between sampling rate and operating clock frequency. When using FPGA resources for designing the filter, The filter design method proposed in this paper reduced the LUT(look-up table) utilization by 55.2% to 6.22%, the FF(flip-flop) utilization decreased by 24.95% to 4.44%, and the BRAM utilization decreased by 42.65% to 13.05% than the Natively design method.
{"title":"Resource Utilization Optimized Design Method for Matched Filter of PSS Searcher","authors":"Dohyun Kim, Taeyang Jeong, Eui-Young Chung","doi":"10.1109/ISOCC50952.2020.9333027","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9333027","url":null,"abstract":"In LTE(Long-Term Evolution) system, UE(User Equipment) performs synchronization processing with a specific cell to communicate. In that processing, the UE uses a matched filter to filter PSS(Primary Synchronization Signal) from downlink signals sent from the cell. There are various ways to design such a matched filter. In the most native design of the matched filter, the number of multipliers is required as much as a number of the taps which means filter length. If resources are limited, that is a very inefficient design approach. Therefore, we proposed filter design method to significantly reduce the number of multipliers in the matched filter by utilizing the difference of between sampling rate and operating clock frequency. When using FPGA resources for designing the filter, The filter design method proposed in this paper reduced the LUT(look-up table) utilization by 55.2% to 6.22%, the FF(flip-flop) utilization decreased by 24.95% to 4.44%, and the BRAM utilization decreased by 42.65% to 13.05% than the Natively design method.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127244690","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9332925
Jo Yoshimoto, Ittetsu Taniguchi, H. Tomiyama, T. Onoye
This paper evaluates the edge computing platform for the drone backup system, which enhances the reliability of automated drones. The drone backup system is assumed to be alternate to execute the critical applications, which used to be executed on edge or cloud, such as image recognition, path planning, etc. Since the drone is facing severe conditions in terms of computational capability, battery capacity, etc., the performance and energy consumption are key issues to support the operation of automated drones. In this paper, we measure the execution time and energy consumption on Raspberry Pi with Intel Neural Compute Stick 2 accelerator for three practical applications: Single Shot MultiBox Detector, State Lattice Planner, and Pix2Pix. The experimental results show the performance and energy consumption on the practical scenarios for the drone backup system. Based on these knowledge, the design optimization of the drone backup systems will be performed for safer drones.
{"title":"An Evaluation of Edge Computing Platform for Reliable Automated Drones","authors":"Jo Yoshimoto, Ittetsu Taniguchi, H. Tomiyama, T. Onoye","doi":"10.1109/ISOCC50952.2020.9332925","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9332925","url":null,"abstract":"This paper evaluates the edge computing platform for the drone backup system, which enhances the reliability of automated drones. The drone backup system is assumed to be alternate to execute the critical applications, which used to be executed on edge or cloud, such as image recognition, path planning, etc. Since the drone is facing severe conditions in terms of computational capability, battery capacity, etc., the performance and energy consumption are key issues to support the operation of automated drones. In this paper, we measure the execution time and energy consumption on Raspberry Pi with Intel Neural Compute Stick 2 accelerator for three practical applications: Single Shot MultiBox Detector, State Lattice Planner, and Pix2Pix. The experimental results show the performance and energy consumption on the practical scenarios for the drone backup system. Based on these knowledge, the design optimization of the drone backup systems will be performed for safer drones.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125954074","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9333124
Yupei Du, K. Man, E. Lim
Sensing, processing, and communication are the 3 key elements for Intelligent Transportation Systems (ITS), while processing is ever advancing on cloud and communication that seems to be solved already by the implementation of 5G communication protocol, sensing has become the most critical part. Traditional video dominated sensing system needs revolutions because of many physical limitations such as degraded performance under bad weather and low illumination conditions, incompetent of detection and tracking overlapped objects, deficient distance and speed detection ability as well as limited field of view. Thankfully, these limitations can be well compensated by radar technology. Radar is known as a kind of all-weather sensor with high accuracy and long-range sensing capability, a radar video fused sensing system could be the key to the next level of intelligent transportation system.
{"title":"Image Radar-based Traffic Surveillance System: An all-weather sensor as intelligent transportation infrastructure component","authors":"Yupei Du, K. Man, E. Lim","doi":"10.1109/ISOCC50952.2020.9333124","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9333124","url":null,"abstract":"Sensing, processing, and communication are the 3 key elements for Intelligent Transportation Systems (ITS), while processing is ever advancing on cloud and communication that seems to be solved already by the implementation of 5G communication protocol, sensing has become the most critical part. Traditional video dominated sensing system needs revolutions because of many physical limitations such as degraded performance under bad weather and low illumination conditions, incompetent of detection and tracking overlapped objects, deficient distance and speed detection ability as well as limited field of view. Thankfully, these limitations can be well compensated by radar technology. Radar is known as a kind of all-weather sensor with high accuracy and long-range sensing capability, a radar video fused sensing system could be the key to the next level of intelligent transportation system.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125009764","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9332929
Kensuke Iizuka, Kohe Ito, Kazuei Hironaka, H. Amano
We propose a partition method to improve the performance of convolutional neural networks (CNN) on a multi-FPGA system called Flow-in-Cloud (FiC) and implement the 2nd layer of AlexNet on FiC. As a result, our implementation is slightly more energy-efficient than the CPU and the GPU with an optimized machine learning framework.
{"title":"A Method of Partitioning Convolutional Layer to Multiple FPGAs","authors":"Kensuke Iizuka, Kohe Ito, Kazuei Hironaka, H. Amano","doi":"10.1109/ISOCC50952.2020.9332929","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9332929","url":null,"abstract":"We propose a partition method to improve the performance of convolutional neural networks (CNN) on a multi-FPGA system called Flow-in-Cloud (FiC) and implement the 2nd layer of AlexNet on FiC. As a result, our implementation is slightly more energy-efficient than the CPU and the GPU with an optimized machine learning framework.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123166648","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9332932
Min-Joon Kim, Sung-Hun Chae, Yeonsoo Moon
In this paper, the implementation result of real-time simulation system for Ii-ion battery protection circuit module (PCM) is presented. Battery protection is one of the most important factors to protect the electrical system. Especially, as the usage of the battery increases, an accurate monitoring of battery state becomes necessary for system safety. Therefore, we implement the PCM consisting of multiple ICs for battery protection and simulation board functioning as voltage load and power supply. The simulation board can show voltage and current autonomously, and also can be linked to the developed PC monitoring program with state-of-charge (SOC) estimation. Finally, the real-time simulation and output monitoring for battery protection is presented.
{"title":"Implementation of Real-time Simulation System for Li-ion Battery Protection Circuit Module","authors":"Min-Joon Kim, Sung-Hun Chae, Yeonsoo Moon","doi":"10.1109/ISOCC50952.2020.9332932","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9332932","url":null,"abstract":"In this paper, the implementation result of real-time simulation system for Ii-ion battery protection circuit module (PCM) is presented. Battery protection is one of the most important factors to protect the electrical system. Especially, as the usage of the battery increases, an accurate monitoring of battery state becomes necessary for system safety. Therefore, we implement the PCM consisting of multiple ICs for battery protection and simulation board functioning as voltage load and power supply. The simulation board can show voltage and current autonomously, and also can be linked to the developed PC monitoring program with state-of-charge (SOC) estimation. Finally, the real-time simulation and output monitoring for battery protection is presented.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126455130","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9332968
Amrita Rana, Kyung Ki Kim
Recent advances in the field of AI have proved that deep neural networks perform and recognize arrhythmia better than cardiologists when trained with a large chunk of data. However, despite the better performance, deep neural networks demand more resources. Therefore, in this paper, a new deep neural network using low resources has been proposed while maintaining high performance, and it is enhanced with a depthwise separable convolution layer for Electrocardiogram (ECG) classification. The algorithm is performed on the Physikalisch-Technische Bundesanstalt (PTB) diagnostic dataset taken from Physionet consisting of two classes: Myocardial Infarction (MI) and Normal (N). Our simulation results show that the proposed lightweight DNN provides high performance with almost the same accuracy as conventional SquezeNets.
{"title":"A Lightweight DNN for ECG Image Classification","authors":"Amrita Rana, Kyung Ki Kim","doi":"10.1109/ISOCC50952.2020.9332968","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9332968","url":null,"abstract":"Recent advances in the field of AI have proved that deep neural networks perform and recognize arrhythmia better than cardiologists when trained with a large chunk of data. However, despite the better performance, deep neural networks demand more resources. Therefore, in this paper, a new deep neural network using low resources has been proposed while maintaining high performance, and it is enhanced with a depthwise separable convolution layer for Electrocardiogram (ECG) classification. The algorithm is performed on the Physikalisch-Technische Bundesanstalt (PTB) diagnostic dataset taken from Physionet consisting of two classes: Myocardial Infarction (MI) and Normal (N). Our simulation results show that the proposed lightweight DNN provides high performance with almost the same accuracy as conventional SquezeNets.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116537594","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9333069
Wun-Siou Jhong, S. Chu, Yu-Jung Huang, Tsun-Yi Hsu, Wei-Chen Lin, Po-Chung Huang, Jia-Jung Wang
This paper presents a software/hardware co-design for classifying three most commonly heart sounds classes: normal, murmur and extrasystole heartbeat. The detection system extracts Mel Frequency Cepstral Coefficient (MFCC)-based heart sound features to train different deep learning network architectures for multiclass classification. The software/hardware co-design for Long Short-Term Memory (LSTM) implementation indicates the multiclass classification accuracy of 85% can be achieved. The proposed heart sound classification platform has great development potential and good application prospects.
{"title":"Deep Learning Hardware/Software Co-Design for Heart Sound Classification","authors":"Wun-Siou Jhong, S. Chu, Yu-Jung Huang, Tsun-Yi Hsu, Wei-Chen Lin, Po-Chung Huang, Jia-Jung Wang","doi":"10.1109/ISOCC50952.2020.9333069","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9333069","url":null,"abstract":"This paper presents a software/hardware co-design for classifying three most commonly heart sounds classes: normal, murmur and extrasystole heartbeat. The detection system extracts Mel Frequency Cepstral Coefficient (MFCC)-based heart sound features to train different deep learning network architectures for multiclass classification. The software/hardware co-design for Long Short-Term Memory (LSTM) implementation indicates the multiclass classification accuracy of 85% can be achieved. The proposed heart sound classification platform has great development potential and good application prospects.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131059972","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9333038
J. Bae, J. Park, J. Park, M. Sunwoo
Medical image diagnosis should consider the information contained in multiple images, not just a single image, such as natural image classification. Mammography is the most basic X-ray screening method for diagnosing breast cancer, and mammograms have four images per patient. Convolutional neural networks should be able to diagnose using these four images. This paper proposes a convolutional neural network to simultaneously concatenate four images to solve the multi-view problem. The proposed network was trained and validated with the digital database for screening mammography (DDSM) and achieved 0.952 area under the ROC curve (AUC) for the two-class problem (positive vs. negative). This paper also proposes a new approach to localize lesions without patch labels or mask labels.
{"title":"Multi-Channel Input Deep Convolutional Neural Network for Mammogram Diagnosis","authors":"J. Bae, J. Park, J. Park, M. Sunwoo","doi":"10.1109/ISOCC50952.2020.9333038","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9333038","url":null,"abstract":"Medical image diagnosis should consider the information contained in multiple images, not just a single image, such as natural image classification. Mammography is the most basic X-ray screening method for diagnosing breast cancer, and mammograms have four images per patient. Convolutional neural networks should be able to diagnose using these four images. This paper proposes a convolutional neural network to simultaneously concatenate four images to solve the multi-view problem. The proposed network was trained and validated with the digital database for screening mammography (DDSM) and achieved 0.952 area under the ROC curve (AUC) for the two-class problem (positive vs. negative). This paper also proposes a new approach to localize lesions without patch labels or mask labels.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132061129","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 : 2020-10-21DOI: 10.1109/ISOCC50952.2020.9333121
Zubair Mehmood, M. Seo
This paper presents a high speed On-Off Keying (OOK) modulator using local oscillator (LO) cancellation technique. Implemented in 28 nm bulk CMOS process, a 100 GHz modulator post layout full-wave EM simulation results are executed for data-rate up to 50 Gbps. The modulator has on-off isolation of 21.6 dB. The proposed modulator design consumes power up to 9.6 mW and occupies chip area of 0.025 mm2.
{"title":"A 100 GHz LO Cancellation Based High Speed OOK Modulator","authors":"Zubair Mehmood, M. Seo","doi":"10.1109/ISOCC50952.2020.9333121","DOIUrl":"https://doi.org/10.1109/ISOCC50952.2020.9333121","url":null,"abstract":"This paper presents a high speed On-Off Keying (OOK) modulator using local oscillator (LO) cancellation technique. Implemented in 28 nm bulk CMOS process, a 100 GHz modulator post layout full-wave EM simulation results are executed for data-rate up to 50 Gbps. The modulator has on-off isolation of 21.6 dB. The proposed modulator design consumes power up to 9.6 mW and occupies chip area of 0.025 mm2.","PeriodicalId":270577,"journal":{"name":"2020 International SoC Design Conference (ISOCC)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133429916","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：