Pub Date : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263638
R. A. Zitar, Mirna Nachouki, Hanan Hussain, Farid Alzboun
A new technique for producing hash values for text documents is introduced in this report. The method uses Recurrent Neural Networks (RNN). RNNs are functionally and temporally dependent on the input vectors of the neural networks (RNN). RNN 's capacity to integrate current values of inputs with previous values that manipulate the associations and the semanticists of the document constitutes a competitive framework for discovering internal interpretations of document details in a special way. In contrast to conventional approaches, two forms of RNNs are evaluated. Current approaches have been adequately examined and the effects of this study reveal the applicability of this artificial intelligence model to construct hash values for plain text. RNNs are very lightweight , portable and parallel in nature and their abilities are used as a potential professional document hashing technology is presented in this article.
{"title":"Recurrent Neural Networks for Signature Generation","authors":"R. A. Zitar, Mirna Nachouki, Hanan Hussain, Farid Alzboun","doi":"10.1109/CISP-BMEI51763.2020.9263638","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263638","url":null,"abstract":"A new technique for producing hash values for text documents is introduced in this report. The method uses Recurrent Neural Networks (RNN). RNNs are functionally and temporally dependent on the input vectors of the neural networks (RNN). RNN 's capacity to integrate current values of inputs with previous values that manipulate the associations and the semanticists of the document constitutes a competitive framework for discovering internal interpretations of document details in a special way. In contrast to conventional approaches, two forms of RNNs are evaluated. Current approaches have been adequately examined and the effects of this study reveal the applicability of this artificial intelligence model to construct hash values for plain text. RNNs are very lightweight , portable and parallel in nature and their abilities are used as a potential professional document hashing technology is presented in this article.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132139146","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263511
Runxin Liu, Jingfeng Bai, Kejun Zhao, Kang Zhang, Cheng Ni
Deep neural networks have been proved to be able to predict accurate dose prediction to improve radiotherapy planning efficiency. However, existing deep-learning-based methods could not predict dose distribution accurately for complicated cases, e.g. tumors at various locations and multi- prescriptions. Based on a new network Channel Attention Densely-connected U-Net (CAD-UNet) proposed by the authors, volume-normalized weight was firstly multiplied to the Mean Squared Error, defined as VN-MSE, as the loss function in the dose prediction area. A cohort of VMAT plans for lung cancer patients was selected for this study. The results show that the new model CAD-UNet with VN-MSE can successfully predict dose distribution of lung cancer cases with single and multiple prescriptions, outperforming CAD-UNet with MSE loss and HD-UNet. The new model demonstrates its potential to be applied for dose prediction in more complicated scenarios.
{"title":"A New Deep-Learning-based Model for Predicting 3D Radiotherapy Dose Distribution In Various Scenarios","authors":"Runxin Liu, Jingfeng Bai, Kejun Zhao, Kang Zhang, Cheng Ni","doi":"10.1109/CISP-BMEI51763.2020.9263511","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263511","url":null,"abstract":"Deep neural networks have been proved to be able to predict accurate dose prediction to improve radiotherapy planning efficiency. However, existing deep-learning-based methods could not predict dose distribution accurately for complicated cases, e.g. tumors at various locations and multi- prescriptions. Based on a new network Channel Attention Densely-connected U-Net (CAD-UNet) proposed by the authors, volume-normalized weight was firstly multiplied to the Mean Squared Error, defined as VN-MSE, as the loss function in the dose prediction area. A cohort of VMAT plans for lung cancer patients was selected for this study. The results show that the new model CAD-UNet with VN-MSE can successfully predict dose distribution of lung cancer cases with single and multiple prescriptions, outperforming CAD-UNet with MSE loss and HD-UNet. The new model demonstrates its potential to be applied for dose prediction in more complicated scenarios.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"82 10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133767666","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263530
Dongjian Cai, Shun Yue, J. Yue
Traffic congestion caused by traffic accidents has seriously affected daily life. The cellular automata model can predict traffic congestion after the traffic accident by simulating the characteristics of vehicle movement. However, the prediction accuracy is poor. Aiming at the shortcomings of the cellular automata model, we studied the characteristics of urban traffic flow, integrated the passenger car unit and random traffic flow. We also improved the probability optimization design in the traditional cellular automata model. Thus, an improved cellular automata model was put forward. The prediction accuracy of the improved model was higher and more stable than that of the traditional model. The model provided technical references for traffic congestion.
{"title":"Dynamic evolution of urban traffic based on improved Cellular Automata","authors":"Dongjian Cai, Shun Yue, J. Yue","doi":"10.1109/CISP-BMEI51763.2020.9263530","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263530","url":null,"abstract":"Traffic congestion caused by traffic accidents has seriously affected daily life. The cellular automata model can predict traffic congestion after the traffic accident by simulating the characteristics of vehicle movement. However, the prediction accuracy is poor. Aiming at the shortcomings of the cellular automata model, we studied the characteristics of urban traffic flow, integrated the passenger car unit and random traffic flow. We also improved the probability optimization design in the traditional cellular automata model. Thus, an improved cellular automata model was put forward. The prediction accuracy of the improved model was higher and more stable than that of the traditional model. The model provided technical references for traffic congestion.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130535454","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263596
TongJia Hou, Liang Zhou
In the named entity recognition task of Chinese electronic ship failure, traditional named entity recognition methods highly rely on manual feature extraction. Therefore, this paper designs a bidirectional long short-term memory (Bi-LSTM) network combined with conditional random field (CRF) network model to optimize the accuracy of ship fault named entity recognition. Firstly, the Chinese ship fault data set is desensitized, and the desensitized text sequence is preprocessed; secondly, the text sequence of ship fault is mapped to the low dimensional vector space by combining the word embedding technology, using the bidirectional long short-term (Bi-LSTM) network model to construct forward and backward semantic features; finally, the input and output of the data are analyzed after entering the conditional random field (CRF) layer, the optimal label of the whole text sequence is obtained through the conditional random field (CRF) layer, and the entity is extracted on this basis. The experimental results show that the model method of combining bilayer bidirectional long short-term memory (Bi-LSTM) network and conditional random field (CRF) can effectively improve the accuracy of named entity recognition of Chinese ship fault.
{"title":"Ship Fault Named Entity Recognition Based on Bilayer Bi-LSTM-CRF","authors":"TongJia Hou, Liang Zhou","doi":"10.1109/CISP-BMEI51763.2020.9263596","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263596","url":null,"abstract":"In the named entity recognition task of Chinese electronic ship failure, traditional named entity recognition methods highly rely on manual feature extraction. Therefore, this paper designs a bidirectional long short-term memory (Bi-LSTM) network combined with conditional random field (CRF) network model to optimize the accuracy of ship fault named entity recognition. Firstly, the Chinese ship fault data set is desensitized, and the desensitized text sequence is preprocessed; secondly, the text sequence of ship fault is mapped to the low dimensional vector space by combining the word embedding technology, using the bidirectional long short-term (Bi-LSTM) network model to construct forward and backward semantic features; finally, the input and output of the data are analyzed after entering the conditional random field (CRF) layer, the optimal label of the whole text sequence is obtained through the conditional random field (CRF) layer, and the entity is extracted on this basis. The experimental results show that the model method of combining bilayer bidirectional long short-term memory (Bi-LSTM) network and conditional random field (CRF) can effectively improve the accuracy of named entity recognition of Chinese ship fault.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127025570","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263695
Heng Dong, Lifang Wei
The retinal vessels segmentation by computer usually assist doctor to detect diabetic retinopathy. Due to the characteristic of retinal vessels structure is complex and changeable, the automatic vessels segmentation still is a challenging task. In this paper, the Mixed filter method is proposed for the retinal vessels segmentation, which utilizes matched filter (MF) combining B-COSFIRE filter to extract the vessels network. Firstly, the CLAFLE algorithm is used to processe with the green channel of retinal image for enhancing the contrast between blood vessel and background. Then, in the matched filter channel, morphological top-hat and bottom-hat are used to further enhance the contrast and Gaussian kernel is used to to extract the thin vessels tree. At the same time, B-COSFIRE filter is make use of filtering the thick vessels tree for green channel of retinal image. The corresponding results of dual channel filtering are segmented and fused to achieve retinal vessels segmentation map. Experimental results show that the proposed algorithm can effectively improve the performance of retinal vessels segmentation compared with single filtering method.
{"title":"Vessels Segmentation Base on Mixed Filter for Retinal Image","authors":"Heng Dong, Lifang Wei","doi":"10.1109/CISP-BMEI51763.2020.9263695","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263695","url":null,"abstract":"The retinal vessels segmentation by computer usually assist doctor to detect diabetic retinopathy. Due to the characteristic of retinal vessels structure is complex and changeable, the automatic vessels segmentation still is a challenging task. In this paper, the Mixed filter method is proposed for the retinal vessels segmentation, which utilizes matched filter (MF) combining B-COSFIRE filter to extract the vessels network. Firstly, the CLAFLE algorithm is used to processe with the green channel of retinal image for enhancing the contrast between blood vessel and background. Then, in the matched filter channel, morphological top-hat and bottom-hat are used to further enhance the contrast and Gaussian kernel is used to to extract the thin vessels tree. At the same time, B-COSFIRE filter is make use of filtering the thick vessels tree for green channel of retinal image. The corresponding results of dual channel filtering are segmented and fused to achieve retinal vessels segmentation map. Experimental results show that the proposed algorithm can effectively improve the performance of retinal vessels segmentation compared with single filtering method.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125785063","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263488
Xi Rubing, Yin Dawei
The multichannel image restoration is a kind of ill-posed inverse problem, which is usually solved by the vectorial variational regularization methods based on the independent and related prior information in the channels of the image.Focus on this, this paper propose a variational regularization model for multi-polarimetric SAR image speckle reduction based on multiplicative-additive noise model. Two level alternating minimization algorithm is designed for solving this new variational regularization model. According to the scattering matrix representation model of the multi-polarimetric SAR image, the amplitude coupling term of the arbitrarily two channels of the multi-polarimetric SAR image satisfies a multiplicative-additive noise model. The distribution of the two kinds of noise is determined by the correlation coefficient of the two channels. This paper establishes a variational regularization model for denoising this kind of multiplicative-additive noise. A auxiliary variable is introduced into the observe model to decompose it into an additive noise model and a multiplicative noise model. Then the variational regularization model for restoring the two channel coupling term from the multiplicative-additive noise is obtained by using the MAP method. To solve this model, it is considered as a minimization model of the primal and the auxiliary variables. Then an alternating minimization algorithm is used to solve the problem. The sub-model with respect to the primal variable is non-convex, which is convexed by a variable substitution technique. Then according to the separation of variables, the penalty method, and the iterative reweighted least squares method, the convex model is transformed into a minimization model with respect to three variables, which is solved again by an alternating minimization algorithm. On the other hand, the sub-model with respect to the auxiliary variable is quadratic convex, which can be easily solved by the Newton iteration method. In this paper, this new model is applied to the multi-polarimetric SAR image, and a fine despeckling result is obtained.
{"title":"A variational regularization model for multi-channel SAR image speckle reduction based on multiplicative-additive noise model","authors":"Xi Rubing, Yin Dawei","doi":"10.1109/CISP-BMEI51763.2020.9263488","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263488","url":null,"abstract":"The multichannel image restoration is a kind of ill-posed inverse problem, which is usually solved by the vectorial variational regularization methods based on the independent and related prior information in the channels of the image.Focus on this, this paper propose a variational regularization model for multi-polarimetric SAR image speckle reduction based on multiplicative-additive noise model. Two level alternating minimization algorithm is designed for solving this new variational regularization model. According to the scattering matrix representation model of the multi-polarimetric SAR image, the amplitude coupling term of the arbitrarily two channels of the multi-polarimetric SAR image satisfies a multiplicative-additive noise model. The distribution of the two kinds of noise is determined by the correlation coefficient of the two channels. This paper establishes a variational regularization model for denoising this kind of multiplicative-additive noise. A auxiliary variable is introduced into the observe model to decompose it into an additive noise model and a multiplicative noise model. Then the variational regularization model for restoring the two channel coupling term from the multiplicative-additive noise is obtained by using the MAP method. To solve this model, it is considered as a minimization model of the primal and the auxiliary variables. Then an alternating minimization algorithm is used to solve the problem. The sub-model with respect to the primal variable is non-convex, which is convexed by a variable substitution technique. Then according to the separation of variables, the penalty method, and the iterative reweighted least squares method, the convex model is transformed into a minimization model with respect to three variables, which is solved again by an alternating minimization algorithm. On the other hand, the sub-model with respect to the auxiliary variable is quadratic convex, which can be easily solved by the Newton iteration method. In this paper, this new model is applied to the multi-polarimetric SAR image, and a fine despeckling result is obtained.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123583355","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}
The requirements of maneuverability, velocity and control ability of modern aircraft are getting higher and higher. Effectively controlling the aircraft when flying at a high attack angle can greatly improve the maneuverability of the aircraft and reduce the stall speed of the aircraft. Flush Air Data Sensing System can indirectly obtain the angle of attack, sideslip angle, and the change of the incoming flow speed which provides necessary parameters for the flight control of the aircraft and meets the parameter measurement requirements of the aircraft. The algorithm research on pressure model and model trimming was conducted based on the three-point method with good real-time performance and high accuracy after study of traditional pressure models. And a series of experiments were performed to modify the accuracy of the model. After experimental verification, the model can be used to calculate the attack angle, sideslip angle, and flow velocity in flight projects, which can provide certain parameters or basis for aircraft flight control.
{"title":"Research on Embedded Atmospheric Measurement Method Based on Three-point Method","authors":"Heng Wang, Jiamei Zhao, Weihe Shen, Hai Jiang, Zhilong Zhang, Jintian Tang","doi":"10.1109/CISP-BMEI51763.2020.9263570","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263570","url":null,"abstract":"The requirements of maneuverability, velocity and control ability of modern aircraft are getting higher and higher. Effectively controlling the aircraft when flying at a high attack angle can greatly improve the maneuverability of the aircraft and reduce the stall speed of the aircraft. Flush Air Data Sensing System can indirectly obtain the angle of attack, sideslip angle, and the change of the incoming flow speed which provides necessary parameters for the flight control of the aircraft and meets the parameter measurement requirements of the aircraft. The algorithm research on pressure model and model trimming was conducted based on the three-point method with good real-time performance and high accuracy after study of traditional pressure models. And a series of experiments were performed to modify the accuracy of the model. After experimental verification, the model can be used to calculate the attack angle, sideslip angle, and flow velocity in flight projects, which can provide certain parameters or basis for aircraft flight control.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124769495","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263681
Hui Sun, Wenju Zhou, M. Fei
Graph matching (GM) which is the problem of finding vertex correspondence among two or multiple graphs is a fundamental problem in computer vision and pattern recognition. GM problem is a discrete combinatorial optimization problem. the property of this problem is NP-hard. Starting with a detailed introduction for modeling methods of graph matching. We walk through the recent development of two-graph matching and multi-graph matching. In two-graph matching, we focus on the continuous domain algorithms and briefly introduce the discrete domain algorithms. In the continuous domain method, we explain the method of transforming the problem from the discrete domain to the continuous domain and those state-of-the-arts algorithms in each type of algorithms in detail, including spectral methods, continuous methods, and deep learning methods. After two-graph matching, we introduce some typical multi-graph matching algorithms. In addition, the research activities of graph matching applications in computer vision and multimedia are displayed. In the end, several directions for future work are discussed.
{"title":"A Survey On Graph Matching In Computer Vision","authors":"Hui Sun, Wenju Zhou, M. Fei","doi":"10.1109/CISP-BMEI51763.2020.9263681","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263681","url":null,"abstract":"Graph matching (GM) which is the problem of finding vertex correspondence among two or multiple graphs is a fundamental problem in computer vision and pattern recognition. GM problem is a discrete combinatorial optimization problem. the property of this problem is NP-hard. Starting with a detailed introduction for modeling methods of graph matching. We walk through the recent development of two-graph matching and multi-graph matching. In two-graph matching, we focus on the continuous domain algorithms and briefly introduce the discrete domain algorithms. In the continuous domain method, we explain the method of transforming the problem from the discrete domain to the continuous domain and those state-of-the-arts algorithms in each type of algorithms in detail, including spectral methods, continuous methods, and deep learning methods. After two-graph matching, we introduce some typical multi-graph matching algorithms. In addition, the research activities of graph matching applications in computer vision and multimedia are displayed. In the end, several directions for future work are discussed.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127093750","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263537
Bo Yang, Yumin He, Honghao Yin
Statistical process control (SPC) emphasizes real-time monitoring of a process and uses statistical methods to provide early warning for a process. This paper applies SPC and proposes an P control chart-based method for process control. An example is provided, which applies the proposed method. The application example illustrates that P control chart can make process control for quality improvement.
{"title":"Research on Data Analysis and Quality Control based on P Control Chart","authors":"Bo Yang, Yumin He, Honghao Yin","doi":"10.1109/CISP-BMEI51763.2020.9263537","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263537","url":null,"abstract":"Statistical process control (SPC) emphasizes real-time monitoring of a process and uses statistical methods to provide early warning for a process. This paper applies SPC and proposes an P control chart-based method for process control. An example is provided, which applies the proposed method. The application example illustrates that P control chart can make process control for quality improvement.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"304 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127265746","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 : 2020-10-17DOI: 10.1109/CISP-BMEI51763.2020.9263636
Hongjiang Liu, Mao Wang, Lihua Liu, Jibing Wu, Hongbin Huang
At present, natural scene text detection methods based on deep learning have achieved outstanding results in many applications. Hull number belongs to the text object, detecting the hull number successfully plays an important role in maritime military and shipping. However, the hull number occupies a relatively small area in the ship image as well as it could be blurred or deformed due to the reason of the photo environment, which made the accuracy of detecting the hull number directly on ship images has great room for improvement. Therefore, this article proposes a hull number detection method based on image super-resolution(SR), which first performs SR on a single ship image, then implements hull number detection on the SR ship image. To reduce the time consumption of the SR process, the original image is divided into multiple grids that perform SR in parallel. Finally, these multiple SR grids are synthesized into a new SR image. Experiments proved that the detection accuracy of the hull number is significantly improved on SR ship images.
{"title":"Hull Number Detection for Ship Images Based on Image Super-Resolution","authors":"Hongjiang Liu, Mao Wang, Lihua Liu, Jibing Wu, Hongbin Huang","doi":"10.1109/CISP-BMEI51763.2020.9263636","DOIUrl":"https://doi.org/10.1109/CISP-BMEI51763.2020.9263636","url":null,"abstract":"At present, natural scene text detection methods based on deep learning have achieved outstanding results in many applications. Hull number belongs to the text object, detecting the hull number successfully plays an important role in maritime military and shipping. However, the hull number occupies a relatively small area in the ship image as well as it could be blurred or deformed due to the reason of the photo environment, which made the accuracy of detecting the hull number directly on ship images has great room for improvement. Therefore, this article proposes a hull number detection method based on image super-resolution(SR), which first performs SR on a single ship image, then implements hull number detection on the SR ship image. To reduce the time consumption of the SR process, the original image is divided into multiple grids that perform SR in parallel. Finally, these multiple SR grids are synthesized into a new SR image. Experiments proved that the detection accuracy of the hull number is significantly improved on SR ship images.","PeriodicalId":346757,"journal":{"name":"2020 13th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127430763","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：