Pub Date : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00018
Chiabwoot Ratanavilisagul
Particle Swarm Optimization (PSO) and K-Means (KM) are widely used for solving data clustering. KM encounters the problem of initializing the cluster centers and the problem of trapping in local optimum. When PSO is applied with KM, it can decrease two problems from KM. Hence, the hybrid clustering technique based on PSO and KM that can enhance performance of clustering is more than using KM alone. However, the hybrid clustering technique encounters the trapping in local optimum problem. To solve this problem, this paper proposed improving hybrid technique by the mutation operation is applied with particles of PSO when swarm traps in local optimum. The proposed technique is tested on eight datasets from the UCI Machine Learning Repository and gives more satisfied search results in comparison with PSOs for the data clustering problems.
{"title":"A novel modified particle swarm optimization algorithm with mutation for data clustering problem","authors":"Chiabwoot Ratanavilisagul","doi":"10.1109/ICCIA49625.2020.00018","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00018","url":null,"abstract":"Particle Swarm Optimization (PSO) and K-Means (KM) are widely used for solving data clustering. KM encounters the problem of initializing the cluster centers and the problem of trapping in local optimum. When PSO is applied with KM, it can decrease two problems from KM. Hence, the hybrid clustering technique based on PSO and KM that can enhance performance of clustering is more than using KM alone. However, the hybrid clustering technique encounters the trapping in local optimum problem. To solve this problem, this paper proposed improving hybrid technique by the mutation operation is applied with particles of PSO when swarm traps in local optimum. The proposed technique is tested on eight datasets from the UCI Machine Learning Repository and gives more satisfied search results in comparison with PSOs for the data clustering problems.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123822214","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00047
Dan Shan, Guotao Cong, W. Lu
Most of the existing convolutional neural networks (CNNs) are based on PC software, which cannot meet the real-time, low power and miniaturization requirements of the systems. In this paper, a CNN accelerator with flexible structure based on Field-Programmable Gate Array (FPGA) is proposed to achieve recognition of MNIST handwritten numeric characters. The system adopts deep pipeline processing and optimizes inter-layer and intra-layer parallelism from two levels of coarse and fine granularity. In view of the similarity of convolution structure, this design adopts structured circuit, which can easily expand the number of layers and neurons. The classification throughput and inter-layer data throughput capability can be improved by rationally organizing the internal memory resources of the FPGA. Compared with the general CPU, it achieves 3 times acceleration at 50MHz frequency, while the power consumption is only 2% of the CPU. Finally performance and power consumption are compared with other accelerators by VGG16.
{"title":"A CNN Accelerator on FPGA with a Flexible Structure","authors":"Dan Shan, Guotao Cong, W. Lu","doi":"10.1109/ICCIA49625.2020.00047","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00047","url":null,"abstract":"Most of the existing convolutional neural networks (CNNs) are based on PC software, which cannot meet the real-time, low power and miniaturization requirements of the systems. In this paper, a CNN accelerator with flexible structure based on Field-Programmable Gate Array (FPGA) is proposed to achieve recognition of MNIST handwritten numeric characters. The system adopts deep pipeline processing and optimizes inter-layer and intra-layer parallelism from two levels of coarse and fine granularity. In view of the similarity of convolution structure, this design adopts structured circuit, which can easily expand the number of layers and neurons. The classification throughput and inter-layer data throughput capability can be improved by rationally organizing the internal memory resources of the FPGA. Compared with the general CPU, it achieves 3 times acceleration at 50MHz frequency, while the power consumption is only 2% of the CPU. Finally performance and power consumption are compared with other accelerators by VGG16.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121102340","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00039
Jinhua Zheng, Tian Chen, H. Xie, Shengxiang Yang
In evolutionary dynamic multiobjective optimization (EDMO), the memory strategy and prediction method are considered as effective and efficient methods. To handling dynamic multiobjective problems (DMOPs), this paper studies the behavior of environment change and tries to make use of the historical information appropriately. And then, this paper proposes an improved memory prediction model that uses the memory strategy to provide valuable information to the prediction model to predict the POS of the new environment more accurately. This memory prediction model is incorporated into a multiobjective evolutionary algorithm based on decomposition (MOEA/D). In particular, the resultant algorithm (MOEA/D-MP) adopts a sensor-based method to detect the environment change and find a similar one in history to reuse the information of it in the prediction process. The proposed algorithm is compared with several state-of-the-art dynamic multiobjective evolutionary algorithms (DMOEA) on six typical benchmark problems with different dynamic characteristics. Experimental results demonstrate that the proposed algorithm can effectively tackle DMOPs.
{"title":"An improved memory prediction strategy for dynamic multiobjective optimization","authors":"Jinhua Zheng, Tian Chen, H. Xie, Shengxiang Yang","doi":"10.1109/ICCIA49625.2020.00039","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00039","url":null,"abstract":"In evolutionary dynamic multiobjective optimization (EDMO), the memory strategy and prediction method are considered as effective and efficient methods. To handling dynamic multiobjective problems (DMOPs), this paper studies the behavior of environment change and tries to make use of the historical information appropriately. And then, this paper proposes an improved memory prediction model that uses the memory strategy to provide valuable information to the prediction model to predict the POS of the new environment more accurately. This memory prediction model is incorporated into a multiobjective evolutionary algorithm based on decomposition (MOEA/D). In particular, the resultant algorithm (MOEA/D-MP) adopts a sensor-based method to detect the environment change and find a similar one in history to reuse the information of it in the prediction process. The proposed algorithm is compared with several state-of-the-art dynamic multiobjective evolutionary algorithms (DMOEA) on six typical benchmark problems with different dynamic characteristics. Experimental results demonstrate that the proposed algorithm can effectively tackle DMOPs.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122998216","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00021
Gaoxiang Cong, Jianxiong Wan, T. Hua, Jie Zhou, Hongxun Niu
Data Centers (DC) requires massive monitoring for thermal and energy efficiency. Currently, popular wireless DC monitoring solutions include Zigbee and Bluetooth, etc. However, these solutions are typically short-range wireless communication technologies, leading to serious scalability issues. In this paper, we design and implement a wireless DC thermal monitoring system based on LoRa (Long Range). The system consists of Data Acquisition Subsystem (DAS), Data Transmission Subsystem (DTS), and Backend Monitoring Subsystem (BMS), where the thermal data are collected via LoRa network with star topology and routed to the BMS for thermal monitoring and fault diagnosis. An advantage of our solution is that the number of nodes that is necessary to cover the data center is significantly reduced due to the long-range communication of LoRa technology. In addition, we further cut the energy consumption of the system by a customized design of the end device such that all irrelevant peripheral components are removed. Finally, we show how dependable and real-time DC thermal monitoring can be achieved by using our solution in a field deployment.
{"title":"A Data Center Thermal Monitoring System Based on LoRa","authors":"Gaoxiang Cong, Jianxiong Wan, T. Hua, Jie Zhou, Hongxun Niu","doi":"10.1109/ICCIA49625.2020.00021","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00021","url":null,"abstract":"Data Centers (DC) requires massive monitoring for thermal and energy efficiency. Currently, popular wireless DC monitoring solutions include Zigbee and Bluetooth, etc. However, these solutions are typically short-range wireless communication technologies, leading to serious scalability issues. In this paper, we design and implement a wireless DC thermal monitoring system based on LoRa (Long Range). The system consists of Data Acquisition Subsystem (DAS), Data Transmission Subsystem (DTS), and Backend Monitoring Subsystem (BMS), where the thermal data are collected via LoRa network with star topology and routed to the BMS for thermal monitoring and fault diagnosis. An advantage of our solution is that the number of nodes that is necessary to cover the data center is significantly reduced due to the long-range communication of LoRa technology. In addition, we further cut the energy consumption of the system by a customized design of the end device such that all irrelevant peripheral components are removed. Finally, we show how dependable and real-time DC thermal monitoring can be achieved by using our solution in a field deployment.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117034568","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00049
Rubén Álvarez-González, Andres Mendez-Vazquez
Vision is the dominant sensory channel by which humans acquire external information. Understanding how the human brain responds to a visual stimulus will help us develop better brain-machine interfaces and describe the human-brain activity response. One technique for tracking brain activity is functional magnetic resonance imaging (fMRI) using blood-oxygen-level-dependent imaging or BOLD-contrast imaging to show the blood oxygenation in the brain before, during and after a stimulus. Identifying the brain activity provoked by a given stimulus is a topic in different research centers.When popular classifiers do not provide perfect accuracy in a practical application, possible causes of their failure can be deficiencies in the algorithms and intrinsic difficulties in the data. In machine and deep learning, models mostly remain black boxes; convolutional neural networks (CNN) are no exception. This understanding of the design of the machine-learning pipeline and the feature-extraction process will provide insight into what a classification model could be.
{"title":"ERP Detector using Texture Filters and Tucker Decomposition","authors":"Rubén Álvarez-González, Andres Mendez-Vazquez","doi":"10.1109/ICCIA49625.2020.00049","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00049","url":null,"abstract":"Vision is the dominant sensory channel by which humans acquire external information. Understanding how the human brain responds to a visual stimulus will help us develop better brain-machine interfaces and describe the human-brain activity response. One technique for tracking brain activity is functional magnetic resonance imaging (fMRI) using blood-oxygen-level-dependent imaging or BOLD-contrast imaging to show the blood oxygenation in the brain before, during and after a stimulus. Identifying the brain activity provoked by a given stimulus is a topic in different research centers.When popular classifiers do not provide perfect accuracy in a practical application, possible causes of their failure can be deficiencies in the algorithms and intrinsic difficulties in the data. In machine and deep learning, models mostly remain black boxes; convolutional neural networks (CNN) are no exception. This understanding of the design of the machine-learning pipeline and the feature-extraction process will provide insight into what a classification model could be.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130798661","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00026
Tiancheng Xia, Yong Qing Fu, Nanlin Jin, P. Chazot, P. Angelov, Richard Jiang
Microscopic blood cell analysis is an important methodology for medical diagnosis, and complete blood cell counts (CBCs) are one of the routine tests operated in hospitals. Results of the CBCs include amounts of red blood cells, white blood cells and platelets in a unit blood sample. It is possible to diagnose diseases such as anemia when the numbers or shapes of red blood cells become abnormal. The percentage of white blood cells is one of the important indicators of many severe illnesses such as infection and cancer. The amounts of platelets are decreased when the patient suffers hemophilia. Doctors often use these as criteria to monitor the general health conditions and recovery stages of the patients in the hospital. However, many hospitals are relying on expensive hematology analyzers to perform these tests, and these procedures are often time consuming. There is a huge demand for an automated, fast and easily used CBCs method in order to avoid redundant procedures and minimize patients’ burden on costs of healthcare. In this research, we investigate a new CBC detection method by using deep neural networks, and discuss state of the art machine learning methods in order to meet the medical usage requirements. The approach we applied in this work is based on YOLOv3 algorithm, and our experimental results show the applied deep learning algorithms have a great potential for CBCs tests, promising for deployment of deep learning methods into microfluidic point-of-care medical devices. As a case of study, we applied our blood cell detector to the blood samples of COVID-19 patients, where blood cell clots are a typical symptom of COVID-19.
{"title":"AI-enabled Microscopic Blood Analysis for Microfluidic COVID-19 Hematology","authors":"Tiancheng Xia, Yong Qing Fu, Nanlin Jin, P. Chazot, P. Angelov, Richard Jiang","doi":"10.1109/ICCIA49625.2020.00026","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00026","url":null,"abstract":"Microscopic blood cell analysis is an important methodology for medical diagnosis, and complete blood cell counts (CBCs) are one of the routine tests operated in hospitals. Results of the CBCs include amounts of red blood cells, white blood cells and platelets in a unit blood sample. It is possible to diagnose diseases such as anemia when the numbers or shapes of red blood cells become abnormal. The percentage of white blood cells is one of the important indicators of many severe illnesses such as infection and cancer. The amounts of platelets are decreased when the patient suffers hemophilia. Doctors often use these as criteria to monitor the general health conditions and recovery stages of the patients in the hospital. However, many hospitals are relying on expensive hematology analyzers to perform these tests, and these procedures are often time consuming. There is a huge demand for an automated, fast and easily used CBCs method in order to avoid redundant procedures and minimize patients’ burden on costs of healthcare. In this research, we investigate a new CBC detection method by using deep neural networks, and discuss state of the art machine learning methods in order to meet the medical usage requirements. The approach we applied in this work is based on YOLOv3 algorithm, and our experimental results show the applied deep learning algorithms have a great potential for CBCs tests, promising for deployment of deep learning methods into microfluidic point-of-care medical devices. As a case of study, we applied our blood cell detector to the blood samples of COVID-19 patients, where blood cell clots are a typical symptom of COVID-19.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121098091","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 : 2020-06-01DOI: 10.1109/iccia49625.2020.00003
{"title":"ICCIA 2020 Breaker Page","authors":"","doi":"10.1109/iccia49625.2020.00003","DOIUrl":"https://doi.org/10.1109/iccia49625.2020.00003","url":null,"abstract":"","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115450604","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00032
Meiyan Zhang, Jinwei Sun, Dan Liu, Qisong Wang
Alertness (also called continuous attention) is a description of a person's ability to maintain attention over a period of time and make appropriate timely feedback to external stimuli. It includes three aspects: the degree of awakening, the concentration of attention and the ability to respond to emergencies. Many human-computer interaction positions, all require alertness maintaining a high level. The accurate assessment and estimation of alertness has become a hot topic in international research. Many researchers use electroencephalogram to evaluate drowsiness and wakefulness, finding that different levels of alertness correspond to different brain activities. This paper uses power spectral density and short-time Fourier transform to extract feature of the denoised brain signals, then proposes the method of Support Vector Machine-DS to evaluate brain alertness based on EEG.
{"title":"Brain alertness evaluation based on SVM-DS","authors":"Meiyan Zhang, Jinwei Sun, Dan Liu, Qisong Wang","doi":"10.1109/ICCIA49625.2020.00032","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00032","url":null,"abstract":"Alertness (also called continuous attention) is a description of a person's ability to maintain attention over a period of time and make appropriate timely feedback to external stimuli. It includes three aspects: the degree of awakening, the concentration of attention and the ability to respond to emergencies. Many human-computer interaction positions, all require alertness maintaining a high level. The accurate assessment and estimation of alertness has become a hot topic in international research. Many researchers use electroencephalogram to evaluate drowsiness and wakefulness, finding that different levels of alertness correspond to different brain activities. This paper uses power spectral density and short-time Fourier transform to extract feature of the denoised brain signals, then proposes the method of Support Vector Machine-DS to evaluate brain alertness based on EEG.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127818520","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00048
Min-Ho Kim, Kyeong-Bin Park, Ki-Seok Chung
The extended min-sum algorithm (EMS) for decoding non-binary low density parity check (NB-LDPC) codes reduces the decoding complexity by truncating the message vector by retaining only the most reliable symbols. However, the EMS algorithm does not consider that the noise of the received codeword is gradually reduced as the iteration count goes up. In this paper, we propose a low-complexity adaptive EMS algorithm, called threshold-based EMS (TB-EMS). The TB-EMS algorithm has a simple adaptive rule to calculate the new message vector length compared to the A-EMS. The proposed algorithm selects one of two message vector lengths. Experimental results show that the proposed algorithm reduces the decoding complexity with minimal performance degradation compared with the EMS algorithm. Further, the decoding performance of the TB-EMS is better than A-EMS.
{"title":"A Low-Complexity Adaptive Extended Min-Sum Algorithm for Non-Binary LDPC Codes","authors":"Min-Ho Kim, Kyeong-Bin Park, Ki-Seok Chung","doi":"10.1109/ICCIA49625.2020.00048","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00048","url":null,"abstract":"The extended min-sum algorithm (EMS) for decoding non-binary low density parity check (NB-LDPC) codes reduces the decoding complexity by truncating the message vector by retaining only the most reliable symbols. However, the EMS algorithm does not consider that the noise of the received codeword is gradually reduced as the iteration count goes up. In this paper, we propose a low-complexity adaptive EMS algorithm, called threshold-based EMS (TB-EMS). The TB-EMS algorithm has a simple adaptive rule to calculate the new message vector length compared to the A-EMS. The proposed algorithm selects one of two message vector lengths. Experimental results show that the proposed algorithm reduces the decoding complexity with minimal performance degradation compared with the EMS algorithm. Further, the decoding performance of the TB-EMS is better than A-EMS.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127651761","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 : 2020-06-01DOI: 10.1109/ICCIA49625.2020.00030
Ming Meng, Shaojun Liu
In this paper, we propose a panorama stitching algorithm based on asymmetric bidirectional optical flow. This algorithm expects multiple photos captured by fisheye lens cameras as input, and then, through the proposed algorithm, these photos can be merged into a high-quality 360-degree spherical panoramic image. For photos taken from a distant perspective, the parallax among them is relatively small, and the obtained panoramic image can be nearly seamless and undistorted. For photos taken from a close perspective or with a relatively large parallax, a seamless though partially distorted panoramic image can also be obtained. Besides, with the help of Graphics Processing Unit (GPU), this algorithm can complete the whole stitching process at a very fast speed: typically, it only takes less than 30s to obtain a panoramic image of 9000-by-4000 pixels, which means our panorama stitching algorithm is of high value in many real-time applications. Our code is available at https://github.com/MungoMeng/Panorama-OpticalFlow.
{"title":"High-quality Panorama Stitching based on Asymmetric Bidirectional Optical Flow","authors":"Ming Meng, Shaojun Liu","doi":"10.1109/ICCIA49625.2020.00030","DOIUrl":"https://doi.org/10.1109/ICCIA49625.2020.00030","url":null,"abstract":"In this paper, we propose a panorama stitching algorithm based on asymmetric bidirectional optical flow. This algorithm expects multiple photos captured by fisheye lens cameras as input, and then, through the proposed algorithm, these photos can be merged into a high-quality 360-degree spherical panoramic image. For photos taken from a distant perspective, the parallax among them is relatively small, and the obtained panoramic image can be nearly seamless and undistorted. For photos taken from a close perspective or with a relatively large parallax, a seamless though partially distorted panoramic image can also be obtained. Besides, with the help of Graphics Processing Unit (GPU), this algorithm can complete the whole stitching process at a very fast speed: typically, it only takes less than 30s to obtain a panoramic image of 9000-by-4000 pixels, which means our panorama stitching algorithm is of high value in many real-time applications. Our code is available at https://github.com/MungoMeng/Panorama-OpticalFlow.","PeriodicalId":237536,"journal":{"name":"2020 5th International Conference on Computational Intelligence and Applications (ICCIA)","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124747975","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：