Pub Date : 2020-11-25DOI: 10.1109/ICSPIS51252.2020.9340130
Aysha Alteneiji, U. Ahmad, Kin Poon, N. Ali, Nawaf I. Almoosa
Filter (PF) is a promising technique for indoor location estimation and tracking. In an indoor environment, localization has become significantly challenging due to multipath reflections. This work addresses the problem of indoor localization of a Moving Target (MT) in a rich multipath environment by fusing acceleration data obtained from Inertial Measurement Unit (IMU) sensors and Angle of Arrival (AoA) measurements. First, the moving target position is predicted using the IMU sensor data. Thereafter, MUltiple SIgnal Classification (MUSIC) algorithm is applied to estimate the AoA of the multipath components. IMU sensor data and the estimated AoA of the multipath components are then fused using the probabilistic framework of the PF to estimate the moving target location. Simulation results demonstrate that the proposed system can achieve a location accuracy of less than $2m$ in a rich multipath environment with only 2 WiFi Access Points (APs).
{"title":"Indoor Localization in Multi-Path Environment based on AoA with Particle Filter","authors":"Aysha Alteneiji, U. Ahmad, Kin Poon, N. Ali, Nawaf I. Almoosa","doi":"10.1109/ICSPIS51252.2020.9340130","DOIUrl":"https://doi.org/10.1109/ICSPIS51252.2020.9340130","url":null,"abstract":"Filter (PF) is a promising technique for indoor location estimation and tracking. In an indoor environment, localization has become significantly challenging due to multipath reflections. This work addresses the problem of indoor localization of a Moving Target (MT) in a rich multipath environment by fusing acceleration data obtained from Inertial Measurement Unit (IMU) sensors and Angle of Arrival (AoA) measurements. First, the moving target position is predicted using the IMU sensor data. Thereafter, MUltiple SIgnal Classification (MUSIC) algorithm is applied to estimate the AoA of the multipath components. IMU sensor data and the estimated AoA of the multipath components are then fused using the probabilistic framework of the PF to estimate the moving target location. Simulation results demonstrate that the proposed system can achieve a location accuracy of less than $2m$ in a rich multipath environment with only 2 WiFi Access Points (APs).","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114953221","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 : 2020-11-25DOI: 10.1109/ICSPIS51252.2020.9340144
Maha Al Hosani, Hamda Al Marzouqi, Shoug Al Junaibi, A. A. Hmoudi, J. Al-Karaki, A. Gawanmeh
Disruptive technologies have evolved dramatically rendering our world increasingly connected and introducing many business opportunities. The use of AI in various businesses has increased in unprecedented rate. In this paper, we design and implement RAASID, a video based facial and emotional recognition smart system that aims at automating the attendance taking procedure in institutes by autonomously marking the attendance of individuals in real-time with no direct physical interaction. In addition, RAASID simultaneously detect and analyze individuals' facial expressions in order to identify the current emotional state of the students at regular time points. The proposed solution guarantees the highest level of student's discipline in the classroom while simultaneously monitoring the student's emotional state regularly. Upon experimentation, the accuracy of the system scored 80% in differentiating students and classifying emotions during various tested classes. The proposed system can be applied to large scale classrooms or conference events with further enhancements.
{"title":"RAASID: A Multipurpose Crowd Sensing Smart System With Sentimental Analysis","authors":"Maha Al Hosani, Hamda Al Marzouqi, Shoug Al Junaibi, A. A. Hmoudi, J. Al-Karaki, A. Gawanmeh","doi":"10.1109/ICSPIS51252.2020.9340144","DOIUrl":"https://doi.org/10.1109/ICSPIS51252.2020.9340144","url":null,"abstract":"Disruptive technologies have evolved dramatically rendering our world increasingly connected and introducing many business opportunities. The use of AI in various businesses has increased in unprecedented rate. In this paper, we design and implement RAASID, a video based facial and emotional recognition smart system that aims at automating the attendance taking procedure in institutes by autonomously marking the attendance of individuals in real-time with no direct physical interaction. In addition, RAASID simultaneously detect and analyze individuals' facial expressions in order to identify the current emotional state of the students at regular time points. The proposed solution guarantees the highest level of student's discipline in the classroom while simultaneously monitoring the student's emotional state regularly. Upon experimentation, the accuracy of the system scored 80% in differentiating students and classifying emotions during various tested classes. The proposed system can be applied to large scale classrooms or conference events with further enhancements.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133119743","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 : 2020-11-25DOI: 10.1109/icspis51252.2020.9340129
Yashbir Singh, S. Deepa, W. Mansoor
Recent progress in computational biology has led to the identification of new genes in the genome of different organisms. In silico approaches enable characterization of structure and function of hypothetical proteins. Vaccinia virus has been used broadly for human immunization. Analysis of Vaccinia virus data determined that 45% of proteins are conserved hypothetical proteins whose function has not been determined. This analysis provides a platform to establish sequence-function relationships and to better understand the molecular machinery of organisms. In this study, we predicted the probable functions of Hypothetical proteins (HPs) and classified all HPs on the basis of sequence similarity, protein family, and domain assignment. The outcome of this work will be helpful for understanding mechanisms of pathogenesis, finding new therapeutic targets, and understanding adaptability to host.
{"title":"Functional Annotation and Identification of Putative Drug Target in VV","authors":"Yashbir Singh, S. Deepa, W. Mansoor","doi":"10.1109/icspis51252.2020.9340129","DOIUrl":"https://doi.org/10.1109/icspis51252.2020.9340129","url":null,"abstract":"Recent progress in computational biology has led to the identification of new genes in the genome of different organisms. In silico approaches enable characterization of structure and function of hypothetical proteins. Vaccinia virus has been used broadly for human immunization. Analysis of Vaccinia virus data determined that 45% of proteins are conserved hypothetical proteins whose function has not been determined. This analysis provides a platform to establish sequence-function relationships and to better understand the molecular machinery of organisms. In this study, we predicted the probable functions of Hypothetical proteins (HPs) and classified all HPs on the basis of sequence similarity, protein family, and domain assignment. The outcome of this work will be helpful for understanding mechanisms of pathogenesis, finding new therapeutic targets, and understanding adaptability to host.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124434658","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 : 2020-11-25DOI: 10.1109/ICSPIS51252.2020.9340131
A. Copiaco, C. Ritz, Stefano Fasciani, N. Abdulaziz
Dementia is an ailment heavily associated with cognitive decline and old age. Due to its progressive nature, several changes in sensory perceptions may be experienced by the individual. Thus, consistent monitoring of patients' assistance requirement, as well as the noise levels throughout their environment, can pose a challenge to caretakers. This is especially apparent for healthcare professionals working in nursing facilities. In this work, we propose an application with an intuitive interface that allows the acoustic monitoring of the patient without infringing their privacy. This is achieved through neural network-based sound scene classification and source location estimation models, which are trained with results of 98.80% and 99.68% F1-scores, respectively. Further, a sound level assessment tool is implemented, such that the time-average levels of the sound are compared to the recommended levels depending on the specific location and time of the day. Experimentation and implementation is carried out in MATLAB, while the interface was developed through the MATLAB App Designer, which can be exported into a mobile phone application as per required.
{"title":"An Application for Dementia Patient Monitoring with Sound Level Assessment Tool","authors":"A. Copiaco, C. Ritz, Stefano Fasciani, N. Abdulaziz","doi":"10.1109/ICSPIS51252.2020.9340131","DOIUrl":"https://doi.org/10.1109/ICSPIS51252.2020.9340131","url":null,"abstract":"Dementia is an ailment heavily associated with cognitive decline and old age. Due to its progressive nature, several changes in sensory perceptions may be experienced by the individual. Thus, consistent monitoring of patients' assistance requirement, as well as the noise levels throughout their environment, can pose a challenge to caretakers. This is especially apparent for healthcare professionals working in nursing facilities. In this work, we propose an application with an intuitive interface that allows the acoustic monitoring of the patient without infringing their privacy. This is achieved through neural network-based sound scene classification and source location estimation models, which are trained with results of 98.80% and 99.68% F1-scores, respectively. Further, a sound level assessment tool is implemented, such that the time-average levels of the sound are compared to the recommended levels depending on the specific location and time of the day. Experimentation and implementation is carried out in MATLAB, while the interface was developed through the MATLAB App Designer, which can be exported into a mobile phone application as per required.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"82 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122740530","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 : 2020-11-25DOI: 10.1109/ICSPIS51252.2020.9340134
M. Alkhodari, O. Hassanin, S. Dhou
Brain tumor segmentation from magnetic resonance (MR) images can have a great impact on improving diagnostics, growth rate prediction, and treatment planning. In this paper, we provide a comparative study of four well-known segmentation algorithms, namely k-means clustering, histogram thresholding (Otsu), fuzzy c-means thresholding, and region growing. For the region growing algorithm, the seed selection process is automated and enhanced by preprocessing the images and approximating the tumor regions using initial clustering and/or thresholding approaches. The evaluation and comparison of the algorithms is conducted using a data-set of T1-Weighted Contrast-Enhanced magnetic resonance imaging (MRI) brain images. Ground truth tumor images were provided by three experienced radiologists and are used in the evaluation process. Results showed that the enhanced region growing method had the highest mean dice similarity coefficient with a score of 0.87, and the lowest under-segmentation rate (17.46%). The fuzzy c-means thresholding method had the lowest over-segmentation rate (0.03%). This study serves as a baseline for other advanced tumor segmentation studies such as the ones using the emergent machine learning approaches.
{"title":"A Comparative Study of Meningioma Tumors Segmentation Methods from MR Images","authors":"M. Alkhodari, O. Hassanin, S. Dhou","doi":"10.1109/ICSPIS51252.2020.9340134","DOIUrl":"https://doi.org/10.1109/ICSPIS51252.2020.9340134","url":null,"abstract":"Brain tumor segmentation from magnetic resonance (MR) images can have a great impact on improving diagnostics, growth rate prediction, and treatment planning. In this paper, we provide a comparative study of four well-known segmentation algorithms, namely k-means clustering, histogram thresholding (Otsu), fuzzy c-means thresholding, and region growing. For the region growing algorithm, the seed selection process is automated and enhanced by preprocessing the images and approximating the tumor regions using initial clustering and/or thresholding approaches. The evaluation and comparison of the algorithms is conducted using a data-set of T1-Weighted Contrast-Enhanced magnetic resonance imaging (MRI) brain images. Ground truth tumor images were provided by three experienced radiologists and are used in the evaluation process. Results showed that the enhanced region growing method had the highest mean dice similarity coefficient with a score of 0.87, and the lowest under-segmentation rate (17.46%). The fuzzy c-means thresholding method had the lowest over-segmentation rate (0.03%). This study serves as a baseline for other advanced tumor segmentation studies such as the ones using the emergent machine learning approaches.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124071242","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 : 2020-11-25DOI: 10.1109/ICSPIS51252.2020.9340141
Mohammed Al Neaimi, H. A. Hamadi, C. Yeun, M. Zemerly
Digital forensic experts are responsible for assisting law enforcement in extracting evidence from electronic devices. Identifying a file type within digital evidence is an essential part of the forensic practice. This paper investigated the existing forensic approaches to identify the file type and developed a new approach based on deep learning and overcome previous approaches' limitations. This paper also highlighted the difference between modern and traditional methods to conduct such an analysis. Whereas, most traditional techniques have been identified to have challenges emanating from the approach structure, which influences how file types are identified, which has prompted researchers in the field to look for new systems that will address this gap. Thus, a new methodology is proposed, which will utilize deep learning techniques to provide a model able to predict corrupted files.
{"title":"Digital Forensic Analysis of Files Using Deep Learning","authors":"Mohammed Al Neaimi, H. A. Hamadi, C. Yeun, M. Zemerly","doi":"10.1109/ICSPIS51252.2020.9340141","DOIUrl":"https://doi.org/10.1109/ICSPIS51252.2020.9340141","url":null,"abstract":"Digital forensic experts are responsible for assisting law enforcement in extracting evidence from electronic devices. Identifying a file type within digital evidence is an essential part of the forensic practice. This paper investigated the existing forensic approaches to identify the file type and developed a new approach based on deep learning and overcome previous approaches' limitations. This paper also highlighted the difference between modern and traditional methods to conduct such an analysis. Whereas, most traditional techniques have been identified to have challenges emanating from the approach structure, which influences how file types are identified, which has prompted researchers in the field to look for new systems that will address this gap. Thus, a new methodology is proposed, which will utilize deep learning techniques to provide a model able to predict corrupted files.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122629752","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 : 2020-11-25DOI: 10.1109/icspis51252.2020.9340142
Debashis Das, Sourav Banerjee, W. Mansoor, U. Biswas, Pushpita Chatterjee, Uttam Ghosh
Blockchain is developing rapidly in various domains for its security. Nowadays, one of the most crucial fundamental concerns is internet security. Blockchain is a novel solution to enhance the security of network applications. However, there are no precise frameworks to secure the Internet of Vehicle (IoV) using Blockchain technology. In this paper, a blockchain-based smart internet of vehicle (BSIoV) framework has been proposed due to the cooperative, collaborative, transparent, and secure characteristics of Blockchain. The main contribution of the proposed work is to connect vehicle-related authorities together to fix a secure and transparent vehicle-to-everything (V2X) communication through the peer-to-peer network connection and provide secure services to the intelligent transport systems. A key management strategy has been included to identify a vehicle in this proposed system. The proposed framework can also provide a significant solution for the data security and safety of the connected vehicles in blockchain network.
{"title":"Design of a Secure Blockchain-Based Smart IoV Architecture","authors":"Debashis Das, Sourav Banerjee, W. Mansoor, U. Biswas, Pushpita Chatterjee, Uttam Ghosh","doi":"10.1109/icspis51252.2020.9340142","DOIUrl":"https://doi.org/10.1109/icspis51252.2020.9340142","url":null,"abstract":"Blockchain is developing rapidly in various domains for its security. Nowadays, one of the most crucial fundamental concerns is internet security. Blockchain is a novel solution to enhance the security of network applications. However, there are no precise frameworks to secure the Internet of Vehicle (IoV) using Blockchain technology. In this paper, a blockchain-based smart internet of vehicle (BSIoV) framework has been proposed due to the cooperative, collaborative, transparent, and secure characteristics of Blockchain. The main contribution of the proposed work is to connect vehicle-related authorities together to fix a secure and transparent vehicle-to-everything (V2X) communication through the peer-to-peer network connection and provide secure services to the intelligent transport systems. A key management strategy has been included to identify a vehicle in this proposed system. The proposed framework can also provide a significant solution for the data security and safety of the connected vehicles in blockchain network.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126296453","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 : 2020-11-25DOI: 10.1109/icspis51252.2020.9340153
{"title":"[Copyright notice]","authors":"","doi":"10.1109/icspis51252.2020.9340153","DOIUrl":"https://doi.org/10.1109/icspis51252.2020.9340153","url":null,"abstract":"","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114175148","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 : 2020-11-25DOI: 10.1109/ICSPIS51252.2020.9340156
K. Afsari, Maha K. Saadeh
The rise of Artificial Intelligence (AI) and robotics in the past decade has created various career opportunities in many industries such as robotics, manufacturing and healthcare. Thus, skills such as AI and robotics will play a critical role in the coming years. This project focuses on introduction of an AI platform for Low-cost robotics using MATLAB. The graphical based interface simplifies the task of design and implantation of machine learning functionalities such as object/pattern recognition and classification. The application offers three types of learning methods such as Machine learning, Deep Learning and Transfer Learning. Currently, the end devices are Arduino based and raspberry pi processors. The communication can be wired or wireless. The MATLAB app is tested using an Arduino based Robot (DFRobot Turtle 2WD) and a Raspberry pi-based robot to achieve object recognition and voice recognition.
{"title":"Artificial Intelligence Platform for Low-Cost Robotics","authors":"K. Afsari, Maha K. Saadeh","doi":"10.1109/ICSPIS51252.2020.9340156","DOIUrl":"https://doi.org/10.1109/ICSPIS51252.2020.9340156","url":null,"abstract":"The rise of Artificial Intelligence (AI) and robotics in the past decade has created various career opportunities in many industries such as robotics, manufacturing and healthcare. Thus, skills such as AI and robotics will play a critical role in the coming years. This project focuses on introduction of an AI platform for Low-cost robotics using MATLAB. The graphical based interface simplifies the task of design and implantation of machine learning functionalities such as object/pattern recognition and classification. The application offers three types of learning methods such as Machine learning, Deep Learning and Transfer Learning. Currently, the end devices are Arduino based and raspberry pi processors. The communication can be wired or wireless. The MATLAB app is tested using an Arduino based Robot (DFRobot Turtle 2WD) and a Raspberry pi-based robot to achieve object recognition and voice recognition.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132960624","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 : 2020-11-25DOI: 10.1109/ICSPIS51252.2020.9340135
M. Mathews, M. AnzarS., R. K. Krishnan, A. Panthakkan
Automated techniques for retinal vessel segmentation is an active research area for the past three decades. Features associated with retinal blood vessels like morphology, area, diameter, tortuosity are important to assess the onset and progression of many eye-related and cardiovascular diseases. For retinal vessel segmentation, we propose two deep neural networks: U-net with EfficientNet as the backbone and EfficientNet encoder with LinkNet decoder. Gamma adjustment and contrast limited histogram equalization is the pre-processing stages adopted. EfficientNetB3 with U-net provide significant improvement. Results are evaluated on benchmark fundus image datasets like DRIVE [1], STARE [2], HRF [3], and CHASE_DB1 [4]. The proposed architecture obtained 96.35% accuracy, 86.35% sensitivity, 97.67% specificity, and an F1 score of 0.8465 on the DRIVE dataset.
{"title":"EfficientNet for retinal blood vessel segmentation","authors":"M. Mathews, M. AnzarS., R. K. Krishnan, A. Panthakkan","doi":"10.1109/ICSPIS51252.2020.9340135","DOIUrl":"https://doi.org/10.1109/ICSPIS51252.2020.9340135","url":null,"abstract":"Automated techniques for retinal vessel segmentation is an active research area for the past three decades. Features associated with retinal blood vessels like morphology, area, diameter, tortuosity are important to assess the onset and progression of many eye-related and cardiovascular diseases. For retinal vessel segmentation, we propose two deep neural networks: U-net with EfficientNet as the backbone and EfficientNet encoder with LinkNet decoder. Gamma adjustment and contrast limited histogram equalization is the pre-processing stages adopted. EfficientNetB3 with U-net provide significant improvement. Results are evaluated on benchmark fundus image datasets like DRIVE [1], STARE [2], HRF [3], and CHASE_DB1 [4]. The proposed architecture obtained 96.35% accuracy, 86.35% sensitivity, 97.67% specificity, and an F1 score of 0.8465 on the DRIVE dataset.","PeriodicalId":373750,"journal":{"name":"2020 3rd International Conference on Signal Processing and Information Security (ICSPIS)","volume":"80 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125886835","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：