Pub Date : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584769
Alexander Castro, Nevo Magnezi, Biruk Sintayehu, Alexander Quinto, P. Abshire
We describe a nano-UAV system for odor source localization in a windless indoor environment. The central part of the system is a small drone (Crazyflie) that has been augmented with a commercial solid state gas sensor. The drone acquires data from onboard gas and optic flow sensors and is controlled by a laptop. We used the sensor to characterize an odor plume both manually and deployed on the Crazyflie. An odor source localization method is described and implemented on the drone. The proposed system uses low cost sensors and is small enough to comfortably and safely fly indoors.
{"title":"Odor Source Localization on a Nano Quadcopter","authors":"Alexander Castro, Nevo Magnezi, Biruk Sintayehu, Alexander Quinto, P. Abshire","doi":"10.1109/BIOCAS.2018.8584769","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584769","url":null,"abstract":"We describe a nano-UAV system for odor source localization in a windless indoor environment. The central part of the system is a small drone (Crazyflie) that has been augmented with a commercial solid state gas sensor. The drone acquires data from onboard gas and optic flow sensors and is controlled by a laptop. We used the sensor to characterize an odor plume both manually and deployed on the Crazyflie. An odor source localization method is described and implemented on the drone. The proposed system uses low cost sensors and is small enough to comfortably and safely fly indoors.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126880036","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584701
Mustafa A. Kanchwala, Grant A. McCallum, D. Durand
Bioelectronic Medicine Therapies offer a promising alternative to traditional procedures for diseases such as epilepsy, and implantable devices are crucial for its development. We present here a miniature, low power, 2 channel wireless neural recording system with sampling rates of 20ksps to allow researchers to understand the neurological functioning to develop therapies in freely moving small animals. The wireless implant uses Carbon Nano Tube Yarn (CNTY) electrodes to interface with the nervous system and record signals. High data transmission rates are achieved by using an Ultra-wideband Impulse Radio (UWB-IR) transmitter and wireless switching control is provided by Bluetooth Low Energy (BLE). The UWB transmitter is primarily designed to make it chronically implantable in freely moving rats to record neural activity but is also applicable to the telemetry of any signals such as surface EEG. Preliminary experiments and bench test results have confirmed its functioning for a distance range of more than 5m with high data transmission rate and low power consumption.
{"title":"A Miniature Wireless Neural Recording System for Chronic Implantation in Freely Moving Animals","authors":"Mustafa A. Kanchwala, Grant A. McCallum, D. Durand","doi":"10.1109/BIOCAS.2018.8584701","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584701","url":null,"abstract":"Bioelectronic Medicine Therapies offer a promising alternative to traditional procedures for diseases such as epilepsy, and implantable devices are crucial for its development. We present here a miniature, low power, 2 channel wireless neural recording system with sampling rates of 20ksps to allow researchers to understand the neurological functioning to develop therapies in freely moving small animals. The wireless implant uses Carbon Nano Tube Yarn (CNTY) electrodes to interface with the nervous system and record signals. High data transmission rates are achieved by using an Ultra-wideband Impulse Radio (UWB-IR) transmitter and wireless switching control is provided by Bluetooth Low Energy (BLE). The UWB transmitter is primarily designed to make it chronically implantable in freely moving rats to record neural activity but is also applicable to the telemetry of any signals such as surface EEG. Preliminary experiments and bench test results have confirmed its functioning for a distance range of more than 5m with high data transmission rate and low power consumption.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127055655","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584697
Khushal Sethi, V. Parmar, M. Suri
Deep learning research has generated widespread interest leading to emergence of a large variety of technological innovations and applications. As significant proportion of deep learning research focuses on vision based applications, there exists a potential for using some of these techniques to enable low-power portable health-care diagnostic support solutions. In this paper, we propose an embedded-hardware-based implementation of microscopy diagnostic support system for PoC case study on: (a) Malaria in thick blood smears, (b) Tuberculosis in sputum samples, and (c) Intestinal parasite infection in stool samples. We use a Squeeze-Net based model to reduce the network size and computation time. We also utilize the Trained Quantization technique to further reduce memory footprint of the learned models. This enables microscopy-based detection of pathogens that classifies with laboratory expert level accuracy as a standalone embedded hardware platform. The proposed implementation is 6x more power-efficient compared to conventional CPU-based implementation and has an inference time of ~ 3 ms/sample.
{"title":"Low-Power Hardware-Based Deep-Learning Diagnostics Support Case Study","authors":"Khushal Sethi, V. Parmar, M. Suri","doi":"10.1109/BIOCAS.2018.8584697","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584697","url":null,"abstract":"Deep learning research has generated widespread interest leading to emergence of a large variety of technological innovations and applications. As significant proportion of deep learning research focuses on vision based applications, there exists a potential for using some of these techniques to enable low-power portable health-care diagnostic support solutions. In this paper, we propose an embedded-hardware-based implementation of microscopy diagnostic support system for PoC case study on: (a) Malaria in thick blood smears, (b) Tuberculosis in sputum samples, and (c) Intestinal parasite infection in stool samples. We use a Squeeze-Net based model to reduce the network size and computation time. We also utilize the Trained Quantization technique to further reduce memory footprint of the learned models. This enables microscopy-based detection of pathogens that classifies with laboratory expert level accuracy as a standalone embedded hardware platform. The proposed implementation is 6x more power-efficient compared to conventional CPU-based implementation and has an inference time of ~ 3 ms/sample.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125906678","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584806
Y. Pu, D. Butterfield, Jorge A. García, Jing Xie, Mark Lin, Rohit Sauhta, Rick Farley, Steven Shellhammer, Moses Derkalousdian, Adam Newham, Chunlei Shi, R. Shenoy, Evgeni Gousev, Rashid Attar
This paper presents an ultra-low-power dual-die platform for (medical) smart hearables. It pairs two custom ASICs: i) Blackghost - a 28nm CMOS near-threshold-VDDpowered and highly integrated SoC with embedded PMU, MCU, 16-issue DSP engine and hardened audio sub-system island; ii) DIRAC - a 28nm CMOS always-on voiceband RF & mixed-signal audio codec frontend. With ~90dB of dynamic range, DIRAC codec consumes <200µW of total power. For fast wakeup, sleep and standby of Blackghost, DIRAC also features low latency microphone activity detection (MAD) and TX-RX cross fading scheme. This dual-die platform enables miniaturized hearable devices capable of running emerging audio algorithms like deep learning at an extremely low enerzy budget.
{"title":"An Ultra-low-power 28nm CMOS Dual-die ASIC Platform for Smart Hearables","authors":"Y. Pu, D. Butterfield, Jorge A. García, Jing Xie, Mark Lin, Rohit Sauhta, Rick Farley, Steven Shellhammer, Moses Derkalousdian, Adam Newham, Chunlei Shi, R. Shenoy, Evgeni Gousev, Rashid Attar","doi":"10.1109/BIOCAS.2018.8584806","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584806","url":null,"abstract":"This paper presents an ultra-low-power dual-die platform for (medical) smart hearables. It pairs two custom ASICs: i) Blackghost - a 28nm CMOS near-threshold-VDDpowered and highly integrated SoC with embedded PMU, MCU, 16-issue DSP engine and hardened audio sub-system island; ii) DIRAC - a 28nm CMOS always-on voiceband RF & mixed-signal audio codec frontend. With ~90dB of dynamic range, DIRAC codec consumes <200µW of total power. For fast wakeup, sleep and standby of Blackghost, DIRAC also features low latency microphone activity detection (MAD) and TX-RX cross fading scheme. This dual-die platform enables miniaturized hearable devices capable of running emerging audio algorithms like deep learning at an extremely low enerzy budget.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127252739","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584715
S. Hsu, Yihan Zi, Ying Choon Wu, P. Jackson, T. Jung
Advancing our understanding of neurocognitive systems impacted by hypnotherapy may improve therapeutic outcomes. This study addresses the challenge of decoding cortical state changes from continuous electroencephalographic (EEG) data recorded during hypnosis. We model changes in brain state dynamics over the course of hypnosis using Adaptive Mixture Independent Component Analysis (AMICA), an unsupervised approach that learns multiple ICA models for characterizing non-stationary, unlabeled data. Applied to EEG from six sessions of hypnosis, AMICA characterized changes in system-wide brain activity that corresponded to transitions between hypnosis stages. Moreover, the results showed consistent AMICA-based models across sessions and subjects that reflected distinct patterns of source activities in different hypnosis states. By analyzing independent component clusters associated with distinctive classes of model probability patterns, shifts in the theta, alpha, and other spectral features of source activities were characterized over the course of the therapy sessions. The AMICA approach offers a promising tool for linking brain-network changes during hypnotherapy with physiological and cognitive state changes brought about by this form of treatment. It can also ignite new research and developments toward brain-state monitoring for clinical applications.
{"title":"Exploring Mental State Changes during Hypnotherapy using Adaptive Mixture Independent Component Analysis of EEG","authors":"S. Hsu, Yihan Zi, Ying Choon Wu, P. Jackson, T. Jung","doi":"10.1109/BIOCAS.2018.8584715","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584715","url":null,"abstract":"Advancing our understanding of neurocognitive systems impacted by hypnotherapy may improve therapeutic outcomes. This study addresses the challenge of decoding cortical state changes from continuous electroencephalographic (EEG) data recorded during hypnosis. We model changes in brain state dynamics over the course of hypnosis using Adaptive Mixture Independent Component Analysis (AMICA), an unsupervised approach that learns multiple ICA models for characterizing non-stationary, unlabeled data. Applied to EEG from six sessions of hypnosis, AMICA characterized changes in system-wide brain activity that corresponded to transitions between hypnosis stages. Moreover, the results showed consistent AMICA-based models across sessions and subjects that reflected distinct patterns of source activities in different hypnosis states. By analyzing independent component clusters associated with distinctive classes of model probability patterns, shifts in the theta, alpha, and other spectral features of source activities were characterized over the course of the therapy sessions. The AMICA approach offers a promising tool for linking brain-network changes during hypnotherapy with physiological and cognitive state changes brought about by this form of treatment. It can also ignite new research and developments toward brain-state monitoring for clinical applications.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125079060","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584677
Marc Foster, Patrick D. Erb, B. Plank, H. West, J. Russenberger, M. Gruen, M. Daniele, D. Roberts, A. Bozkurt
This paper describes the design and fabrication of 3D-printed conductive electrodes for measuring heart rate and heart rate variability in animals. The customizable electrodes have a three-legged stool structure with rounded edges in order to provide optimized, balanced, and comfortable skin contact through the fur of the animal without needing to shave it. We explored two alternative designs for manufacturing: a flexible, insulated base structure coated with graphene for conductivity and a rigid, all-conductive, graphene-infused PLA base structure. To enable connection to standard female electrocardiogram snap connectors, we epoxied the former electrode with a metal male snap connector and the latter electrode had the standard-sized male snap connector 3D printed in one assembly. We characterized and compared the performance of these electrodes through electrochemical impedance spectroscopy and benchmarked with commercial electrodes traditionally used in veterinary clinics. Preliminary in vivo results demonstrate the feasibility of these electrodes to measure heart rate and heart rate variability in canine puppies.
{"title":"3D-Printed Electrocardiogram Electrodes for Heart Rate Detection in Canines","authors":"Marc Foster, Patrick D. Erb, B. Plank, H. West, J. Russenberger, M. Gruen, M. Daniele, D. Roberts, A. Bozkurt","doi":"10.1109/BIOCAS.2018.8584677","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584677","url":null,"abstract":"This paper describes the design and fabrication of 3D-printed conductive electrodes for measuring heart rate and heart rate variability in animals. The customizable electrodes have a three-legged stool structure with rounded edges in order to provide optimized, balanced, and comfortable skin contact through the fur of the animal without needing to shave it. We explored two alternative designs for manufacturing: a flexible, insulated base structure coated with graphene for conductivity and a rigid, all-conductive, graphene-infused PLA base structure. To enable connection to standard female electrocardiogram snap connectors, we epoxied the former electrode with a metal male snap connector and the latter electrode had the standard-sized male snap connector 3D printed in one assembly. We characterized and compared the performance of these electrodes through electrochemical impedance spectroscopy and benchmarked with commercial electrodes traditionally used in veterinary clinics. Preliminary in vivo results demonstrate the feasibility of these electrodes to measure heart rate and heart rate variability in canine puppies.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124160356","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584837
J. Birjandtalab, M. James, M. Nourani, J. Harvey
EEG data collected in EMU is highly imbalanced and accuracy of automatic epileptic seizure detection is naturally low. Our aim is to increase the accuracy by reducing the imbalance ratio of seizure and non-seizure classes. We hypothesis that the non-seizure class itself includes various daily brain activities and then the data points are distributed as clusters in this class. In training phase, we propose a technique to cluster the majority (non-seizure) class into k clusters. Then, we train k KNN classifiers using each of k non-seizure clusters plus seizure class. In testing phase, we classify an incoming sample using this model and the non-seizure cluster closest to the incoming sample. We employed a state-of-the-art visualization technique to illustrate clusters of majority non-seizure class in two dimensions. The results, applied to MIT EEG dataset, show that our technique provides a higher average F-Measure accuracy.
{"title":"Learning from Non-Seizure Clusters for EEG Analytics","authors":"J. Birjandtalab, M. James, M. Nourani, J. Harvey","doi":"10.1109/BIOCAS.2018.8584837","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584837","url":null,"abstract":"EEG data collected in EMU is highly imbalanced and accuracy of automatic epileptic seizure detection is naturally low. Our aim is to increase the accuracy by reducing the imbalance ratio of seizure and non-seizure classes. We hypothesis that the non-seizure class itself includes various daily brain activities and then the data points are distributed as clusters in this class. In training phase, we propose a technique to cluster the majority (non-seizure) class into k clusters. Then, we train k KNN classifiers using each of k non-seizure clusters plus seizure class. In testing phase, we classify an incoming sample using this model and the non-seizure cluster closest to the incoming sample. We employed a state-of-the-art visualization technique to illustrate clusters of majority non-seizure class in two dimensions. The results, applied to MIT EEG dataset, show that our technique provides a higher average F-Measure accuracy.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"600 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121980887","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584798
Z. Sun, Yingdan Li, Hanjun Jiang, Fei Chen, Zhihua Wang
A software defined binaural hearing aid system with a smartphone-centered architecture has been developed. This architecture takes advantage of the powerful computing hardware and memory capacity of smartphones. The sound signals are captured under complex acoustics scenes. And they need further enhancement to satisfy patients' personalized requirements. A minimum variance distortion response (MVDR) and binaural multichannel wiener filtering (MWF) combined algorithm (MMC) for speech enhancement is proposed. The proposed algorithm can achieve balance between noise reduction in complex acoustics environment and preservation of interaural cues therefore improve speech intelligence at the same time. The subjective and objective evaluation results as well as running on the binaural hearing aid with smartphone platform case demonstrate the efficiency of our work.
{"title":"A MVDR- MWF Combined Algorithm for Binaural Hearing Aid System","authors":"Z. Sun, Yingdan Li, Hanjun Jiang, Fei Chen, Zhihua Wang","doi":"10.1109/BIOCAS.2018.8584798","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584798","url":null,"abstract":"A software defined binaural hearing aid system with a smartphone-centered architecture has been developed. This architecture takes advantage of the powerful computing hardware and memory capacity of smartphones. The sound signals are captured under complex acoustics scenes. And they need further enhancement to satisfy patients' personalized requirements. A minimum variance distortion response (MVDR) and binaural multichannel wiener filtering (MWF) combined algorithm (MMC) for speech enhancement is proposed. The proposed algorithm can achieve balance between noise reduction in complex acoustics environment and preservation of interaural cues therefore improve speech intelligence at the same time. The subjective and objective evaluation results as well as running on the binaural hearing aid with smartphone platform case demonstrate the efficiency of our work.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122159197","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584665
Fatemeh Koochaki, L. Najafizadeh
Eye movement is a valuable (and in several cases, the only remaining) means of communication for impaired people with extremely limited motor or communication capabilities. In this paper, we present a new framework that utilizes eye gaze patterns as input, to predict user's intention for performing daily tasks. The proposed framework consists of two main modules. First, by clustering the eye gaze patterns, the regions of interest (ROIs) on the displayed image are extracted. A deep convolutional neural network is then trained and used to recognize the objects in each ROI. Finally, the intended task is predicted by using support vector machine (SVM) through learning the embedded relationship between recognized objects. The proposed framework is tested using data from 8 subjects, in an experiment considering 4 intended tasks as well as the scenario in which the user does not have a specific intention when looking at the displayed image. Results demonstrate an average accuracy of 95.68% across all tasks, confirming the efficacy of the proposed framework.
{"title":"Predicting Intention Through Eye Gaze Patterns","authors":"Fatemeh Koochaki, L. Najafizadeh","doi":"10.1109/BIOCAS.2018.8584665","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584665","url":null,"abstract":"Eye movement is a valuable (and in several cases, the only remaining) means of communication for impaired people with extremely limited motor or communication capabilities. In this paper, we present a new framework that utilizes eye gaze patterns as input, to predict user's intention for performing daily tasks. The proposed framework consists of two main modules. First, by clustering the eye gaze patterns, the regions of interest (ROIs) on the displayed image are extracted. A deep convolutional neural network is then trained and used to recognize the objects in each ROI. Finally, the intended task is predicted by using support vector machine (SVM) through learning the embedded relationship between recognized objects. The proposed framework is tested using data from 8 subjects, in an experiment considering 4 intended tasks as well as the scenario in which the user does not have a specific intention when looking at the displayed image. Results demonstrate an average accuracy of 95.68% across all tasks, confirming the efficacy of the proposed framework.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115041619","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 : 2018-10-01DOI: 10.1109/BIOCAS.2018.8584683
J. Birjandtalab, V. Jarmale, M. Nourani, J. Harvey
Around 1% of world's population suffer from epileptic seizures which can lead to injuries and even unexpected death. Making use of EEG signals, which are proven to be the best indicators of seizures, we aim to build an Artificial Neural Networks to classify seizure and non-seizure events. However, the limited availability of seizure events in the EEG data makes it difficult for the automatic classifiers in general to accurately classify seizure events. To improve this, we propose an imbalance learning approach to improve accuracy of highly imbalanced seizure dataset. Since each patient provides a different response to the seizure, we personalize the classification models in terms of training data and model parameters. The proposed imbalance learning method provides an average F-measure accuracy above 86% for Physionet MIT dataset.
{"title":"Imbalance Learning Using Neural Networks for Seizure Detection","authors":"J. Birjandtalab, V. Jarmale, M. Nourani, J. Harvey","doi":"10.1109/BIOCAS.2018.8584683","DOIUrl":"https://doi.org/10.1109/BIOCAS.2018.8584683","url":null,"abstract":"Around 1% of world's population suffer from epileptic seizures which can lead to injuries and even unexpected death. Making use of EEG signals, which are proven to be the best indicators of seizures, we aim to build an Artificial Neural Networks to classify seizure and non-seizure events. However, the limited availability of seizure events in the EEG data makes it difficult for the automatic classifiers in general to accurately classify seizure events. To improve this, we propose an imbalance learning approach to improve accuracy of highly imbalanced seizure dataset. Since each patient provides a different response to the seizure, we personalize the classification models in terms of training data and model parameters. The proposed imbalance learning method provides an average F-measure accuracy above 86% for Physionet MIT dataset.","PeriodicalId":259162,"journal":{"name":"2018 IEEE Biomedical Circuits and Systems Conference (BioCAS)","volume":"82 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134436313","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：