Pub Date : 2021-11-01DOI: 10.1109/ITME53901.2021.00126
Guo-Hong Lei, Liu Xingjuan, Ji Shaoli
Mathematics is a basic subject to realize the combination of number and form. Intuitive imagination literacy is a necessary literacy to abstract visual images into mathematical language. Junior middle school is an important period to lay the foundation for senior high school. Therefore, it is very meaningful to study the strategy of cultivating junior middle school students' intuitive imagination literacy. This paper mainly studies the application of the Geometer's Sketchpad in exploring the conditions of congruence of triangles in junior middle school. Based on this application, this paper gives the strategy of using the Geometer's Sketchpad to Cultivate Junior Middle School Students' intuitive imagination literacy. Establishing an intuitive situation, cultivating the idea of combining numbers and shapes, and students' independent exploration through geometric sketchpad are important strategies to Cultivate Junior Middle School Students' intuitive imagination literacy.
{"title":"Cultivation strategy of junior middle school students' intuitive imagination literacy based on computer software - the Geometer's Sketchpad","authors":"Guo-Hong Lei, Liu Xingjuan, Ji Shaoli","doi":"10.1109/ITME53901.2021.00126","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00126","url":null,"abstract":"Mathematics is a basic subject to realize the combination of number and form. Intuitive imagination literacy is a necessary literacy to abstract visual images into mathematical language. Junior middle school is an important period to lay the foundation for senior high school. Therefore, it is very meaningful to study the strategy of cultivating junior middle school students' intuitive imagination literacy. This paper mainly studies the application of the Geometer's Sketchpad in exploring the conditions of congruence of triangles in junior middle school. Based on this application, this paper gives the strategy of using the Geometer's Sketchpad to Cultivate Junior Middle School Students' intuitive imagination literacy. Establishing an intuitive situation, cultivating the idea of combining numbers and shapes, and students' independent exploration through geometric sketchpad are important strategies to Cultivate Junior Middle School Students' intuitive imagination literacy.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"39 1","pages":"595-599"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87207964","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00072
Li-hao Shen, Wei Zhou
In the context of the increasing pace of information technology in hospitals, the construction of “intelligent medicine”, with new technologies and methods as the main means of innovation, is also becoming an important part of the development of high-quality medical care. In the care system, vital signs such as blood pressure and body temperature are important indicators of the health of the human body. This paper aims to build a vital signs measurement system using RFID technology as the main technical means, by combining sensor technology and RFID transmission technology, using active RFID tags to build body temperature and blood pressure sensor tags [1], and latching on to RFID technology to complete the measurement and transmission of data. The RFID technology can be used for real-time monitoring of blood pressure and body temperature, and can play an active role in the diagnosis and health monitoring of patients with acute and critical illnesses, which will ultimately promote the development of intelligent care and high quality of care.
{"title":"Development of RFID-based “Smart+” Vital Signs Measurement Instruments and Systems","authors":"Li-hao Shen, Wei Zhou","doi":"10.1109/ITME53901.2021.00072","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00072","url":null,"abstract":"In the context of the increasing pace of information technology in hospitals, the construction of “intelligent medicine”, with new technologies and methods as the main means of innovation, is also becoming an important part of the development of high-quality medical care. In the care system, vital signs such as blood pressure and body temperature are important indicators of the health of the human body. This paper aims to build a vital signs measurement system using RFID technology as the main technical means, by combining sensor technology and RFID transmission technology, using active RFID tags to build body temperature and blood pressure sensor tags [1], and latching on to RFID technology to complete the measurement and transmission of data. The RFID technology can be used for real-time monitoring of blood pressure and body temperature, and can play an active role in the diagnosis and health monitoring of patients with acute and critical illnesses, which will ultimately promote the development of intelligent care and high quality of care.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"124 1","pages":"319-323"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82989291","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00068
Zhu Yuming, Xu Wenlong
Respiratory diseases have a significant impact on the health and social economy of the population, and there are currently limited ways to detect respiratory diseases in hospitals. To this end, we proposed a cascade neural network model based on multi-features fusion to classify respiratory diseases. Meanwhile, we also used two different pre-processings to input respiratory sounds into three different deep neural networks for comparative experiments. In order to solve the problem of class- imbalance of the dataset, we extend the dataset. Our system classifies six respiratory diseases, and achieves 88.3% ICBHI average accuracy, respectively. The average accuracy is repeated on ten random splittings of 80% training and 20% testing data using the ICBHI 2017 dataset of respiratory cycles.
{"title":"Research on Classification of Respiratory Diseases Based on Multi-features Fusion Cascade Neural Network","authors":"Zhu Yuming, Xu Wenlong","doi":"10.1109/ITME53901.2021.00068","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00068","url":null,"abstract":"Respiratory diseases have a significant impact on the health and social economy of the population, and there are currently limited ways to detect respiratory diseases in hospitals. To this end, we proposed a cascade neural network model based on multi-features fusion to classify respiratory diseases. Meanwhile, we also used two different pre-processings to input respiratory sounds into three different deep neural networks for comparative experiments. In order to solve the problem of class- imbalance of the dataset, we extend the dataset. Our system classifies six respiratory diseases, and achieves 88.3% ICBHI average accuracy, respectively. The average accuracy is repeated on ten random splittings of 80% training and 20% testing data using the ICBHI 2017 dataset of respiratory cycles.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"24 1","pages":"298-301"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91185909","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00107
Jiyan Wang, Congcong Li
Engineering mechanics is a fundamental major course for engineering majors in application-oriented universities. Based on reviewing current situation of classroom teaching of engineering mechanics, the paper puts forward the countermeasures for reforming classroom teaching in the student-centered principle, including making the innovation of teaching pattern, integrating theoretical knowledge with engineering practice and introducing curriculum ideological and political education, in order to achieve the cultivation of comprehensive and all-round applied-type talents. The research can provide reference for the education of fundamental mechanics courses.
{"title":"Research on Classroom Teaching Reform of Engineering Mechanics in Application-Oriented Universities","authors":"Jiyan Wang, Congcong Li","doi":"10.1109/ITME53901.2021.00107","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00107","url":null,"abstract":"Engineering mechanics is a fundamental major course for engineering majors in application-oriented universities. Based on reviewing current situation of classroom teaching of engineering mechanics, the paper puts forward the countermeasures for reforming classroom teaching in the student-centered principle, including making the innovation of teaching pattern, integrating theoretical knowledge with engineering practice and introducing curriculum ideological and political education, in order to achieve the cultivation of comprehensive and all-round applied-type talents. The research can provide reference for the education of fundamental mechanics courses.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"22 1","pages":"505-508"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82675410","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00092
Li Ruiyu, Pan Rufang
Objective:By summarizing the methods and drugs used in the treatment of cervical cancer by traditional Chinese medicine(TCM) in the past five years, reveal the great potential of TCM in the treatment of cervical cancer, and then provide relevant theoretical support and medication reference for clinical treatment. Methods: Through computer search of Wanfang.com, CNKI, Weipu.com, Pubmed and other databases, collect relevant literature from 2016 to 2021, summarize high-frequency Chinese herbal medicine and traditional Chinese medicine treatment methods for the treatment of cervical cancer, and select typical cases to demonstrate the application and effect of these methods. Result: (1)Different external treatment methods such as acupuncture, enema, external application, and external washing can treat and relieve cervical cancer complications or adverse reactions related to surgery, radiotherapy and chemotherapy, and have certain effects on reducing HPV infection. (2)Oral use of TCM decoction can significantly relieve the clinical symptoms of patients with cervical cancer and improve the quality of life.(3) Comprehensive treatment has the advantages and disadvantages of internal treatment and external treatment. Clinically, choose the best treatment method to avoid excessive treatment.(4)Different TCM doctors have different medication characteristics. On the whole, the current high-frequency Chinese medicines for the treatment of cervical cancer are mainly detoxification, tonic, heat-clearing, and blood- activating medicines. Conclusion: TCM can effectively alleviate the clinical symptoms of cervical cancer patients, reduce the side effects caused by surgery, radiotherapy, radiotherapy and chemotherapy, and has important therapeutic value. However, there is no TCM therapy that can directly cure and remove solid cervical cancer tumors.
{"title":"2016–2021 domestic and foreign research progress in the treatment of cervical cancer with traditional Chinese medicine","authors":"Li Ruiyu, Pan Rufang","doi":"10.1109/ITME53901.2021.00092","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00092","url":null,"abstract":"Objective:By summarizing the methods and drugs used in the treatment of cervical cancer by traditional Chinese medicine(TCM) in the past five years, reveal the great potential of TCM in the treatment of cervical cancer, and then provide relevant theoretical support and medication reference for clinical treatment. Methods: Through computer search of Wanfang.com, CNKI, Weipu.com, Pubmed and other databases, collect relevant literature from 2016 to 2021, summarize high-frequency Chinese herbal medicine and traditional Chinese medicine treatment methods for the treatment of cervical cancer, and select typical cases to demonstrate the application and effect of these methods. Result: (1)Different external treatment methods such as acupuncture, enema, external application, and external washing can treat and relieve cervical cancer complications or adverse reactions related to surgery, radiotherapy and chemotherapy, and have certain effects on reducing HPV infection. (2)Oral use of TCM decoction can significantly relieve the clinical symptoms of patients with cervical cancer and improve the quality of life.(3) Comprehensive treatment has the advantages and disadvantages of internal treatment and external treatment. Clinically, choose the best treatment method to avoid excessive treatment.(4)Different TCM doctors have different medication characteristics. On the whole, the current high-frequency Chinese medicines for the treatment of cervical cancer are mainly detoxification, tonic, heat-clearing, and blood- activating medicines. Conclusion: TCM can effectively alleviate the clinical symptoms of cervical cancer patients, reduce the side effects caused by surgery, radiotherapy, radiotherapy and chemotherapy, and has important therapeutic value. However, there is no TCM therapy that can directly cure and remove solid cervical cancer tumors.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"15 1","pages":"425-431"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81402474","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00059
Yezhen Wang, Haobin Zheng, Changjiang Mao, Jing Zhang, Xiao Ke
With the economic and social development and the substantial improvement of material conditions, the generation of domestic waste has grown rapidly and has become a constraint factor for the development of new urbanization. In the past few years, research on the domestic waste industry has been limited to intelligent waste sorting, neglecting the role of intelligent management of waste storage sites. To relieve it, We propose a deep learning-based multi-task network for intelligent management of garbage deposit points, which combines algorithms such as YoloV5,Deepsort, Insightface, and Openpose to achieve waste bin detection, waste bin status recognition and analysis, face recognition, action recognition, and multiple object tracking based on real-time surveillance video. Besides, we propose a new dataset named Waste Bin Status, which provides a meaningful addition to the existing field of waste bin identification. Experiments on WBS dataset validate that our method is superior to other methods for garbage point status identification. Moreover, our network is trained to work with different scenarios of garbage deposits, demonstrating state-of-the-art performance in real-world tests.
{"title":"Deep Learning-based Multi-task Network for Intelligent Management of Garbage Deposit Points","authors":"Yezhen Wang, Haobin Zheng, Changjiang Mao, Jing Zhang, Xiao Ke","doi":"10.1109/ITME53901.2021.00059","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00059","url":null,"abstract":"With the economic and social development and the substantial improvement of material conditions, the generation of domestic waste has grown rapidly and has become a constraint factor for the development of new urbanization. In the past few years, research on the domestic waste industry has been limited to intelligent waste sorting, neglecting the role of intelligent management of waste storage sites. To relieve it, We propose a deep learning-based multi-task network for intelligent management of garbage deposit points, which combines algorithms such as YoloV5,Deepsort, Insightface, and Openpose to achieve waste bin detection, waste bin status recognition and analysis, face recognition, action recognition, and multiple object tracking based on real-time surveillance video. Besides, we propose a new dataset named Waste Bin Status, which provides a meaningful addition to the existing field of waste bin identification. Experiments on WBS dataset validate that our method is superior to other methods for garbage point status identification. Moreover, our network is trained to work with different scenarios of garbage deposits, demonstrating state-of-the-art performance in real-world tests.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"5 1","pages":"251-256"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86489742","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00035
Xiang Chen, Zongwen Fan, Jin Gou
The task of finding similar patterns in a long time series, commonly called motifs, has received continuous and increasing attention from diverse scientific fields. Although numerous approaches have been proposed for motif discovery, they cannot discover the motifs in an exact and efficient manner. Furthermore, domain knowledge is required from the experts for those methods to predefine the pattern length, which is also quite objective. In addiction, it is very time-consuming to extract the exact motifs and sometimes the extracted motif has no specific meanings. Especially in the field of financial and hydrology, many studies are focused on whether there is a fixed pattern including trend information hidden in the data. To address the above problems, we proposed a framework to automatically discovery the trend motifs without predefining the length of patterns. It has four main steps, (1) singular spectrum analysis is first applied to removed noise; (2) segmentation by extracting extreme points is then employed to automatically obtain the unequal length of time series pattern; (3) symbolic aggregate approximation is introduced to discretize the data and transform them into string sequences; (4) finally, the trend motifs are selected by measuring their similarity. Experimental results on the real-world time-series datasets reveal that our framework fit well in different circumstances, indicating our proposed framework is effective for trend motif discovery.
{"title":"Detecting trend motifs: an efficient framework for time series motif discovery","authors":"Xiang Chen, Zongwen Fan, Jin Gou","doi":"10.1109/ITME53901.2021.00035","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00035","url":null,"abstract":"The task of finding similar patterns in a long time series, commonly called motifs, has received continuous and increasing attention from diverse scientific fields. Although numerous approaches have been proposed for motif discovery, they cannot discover the motifs in an exact and efficient manner. Furthermore, domain knowledge is required from the experts for those methods to predefine the pattern length, which is also quite objective. In addiction, it is very time-consuming to extract the exact motifs and sometimes the extracted motif has no specific meanings. Especially in the field of financial and hydrology, many studies are focused on whether there is a fixed pattern including trend information hidden in the data. To address the above problems, we proposed a framework to automatically discovery the trend motifs without predefining the length of patterns. It has four main steps, (1) singular spectrum analysis is first applied to removed noise; (2) segmentation by extracting extreme points is then employed to automatically obtain the unequal length of time series pattern; (3) symbolic aggregate approximation is introduced to discretize the data and transform them into string sequences; (4) finally, the trend motifs are selected by measuring their similarity. Experimental results on the real-world time-series datasets reveal that our framework fit well in different circumstances, indicating our proposed framework is effective for trend motif discovery.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"26 1","pages":"122-126"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82407420","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00028
Zhibo Gu, Zhiming Luo, Min Huang, Yuanzheng Cai, Shaozi Li
Over the past few years, deep convolutional neural networks (CNNs) based semantic segmentation methods reached the state-of-the-art performance. To train a model with the ability to know a concept, a lot of pixel level annotated images are required, which is time consuming and hard to cover unseen object categories. Thus, few-shot semantic segmentation has been developed to implement segmentation with a few annotation images. In this paper, we proposed a novel prototype mixing model for few shot segmentation. Different with other works which only produce prototypes form support set, our proposed model learn a group of concept-specific prototypes from support set and then generate prototypes from query set. With prototypes from both query set and support set, we proposed a GCN(Graphic Convolutional Network) module to generate mixing prototypes for better utilizing of informations from different categories. We also proposed a clustering module to produce multi-prototypes for representing different parts of a single semantic class, which reach better performance than single prototype. Our model achieve 48.8% and 55.9%mIoU score on PASCAL-5i for 1-shot and 5-shot settings respectively.
{"title":"A Graph-Convolutional-Network based Prototype Mixing Model for Few-shot Segmentation","authors":"Zhibo Gu, Zhiming Luo, Min Huang, Yuanzheng Cai, Shaozi Li","doi":"10.1109/ITME53901.2021.00028","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00028","url":null,"abstract":"Over the past few years, deep convolutional neural networks (CNNs) based semantic segmentation methods reached the state-of-the-art performance. To train a model with the ability to know a concept, a lot of pixel level annotated images are required, which is time consuming and hard to cover unseen object categories. Thus, few-shot semantic segmentation has been developed to implement segmentation with a few annotation images. In this paper, we proposed a novel prototype mixing model for few shot segmentation. Different with other works which only produce prototypes form support set, our proposed model learn a group of concept-specific prototypes from support set and then generate prototypes from query set. With prototypes from both query set and support set, we proposed a GCN(Graphic Convolutional Network) module to generate mixing prototypes for better utilizing of informations from different categories. We also proposed a clustering module to produce multi-prototypes for representing different parts of a single semantic class, which reach better performance than single prototype. Our model achieve 48.8% and 55.9%mIoU score on PASCAL-5i for 1-shot and 5-shot settings respectively.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"37 1","pages":"86-90"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81652303","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00075
Meng Zhao, Lingmin Jin, Shenghua Teng, Zuoyong Li
In medicine, white blood cell (WBC) classification plays an important role in clinical diagnosis and treatment. Due to the similarity between classes and lack of training data, the precise classification of WBC is still challenging. To alleviate this problem, we propose an attention residual network for WBC image classification on the basis of data augmentation. Specifically, the attention residual network is composed of multiple attention residual blocks, an adaptive average pooling layer, and a full connection layer. The channel attention mechanism is introduced in each residual block to use the feature maps of WBC learned by a high layer to generate the attention map for a low layer. Each attention residual block also introduces depth separable convolution to extract the feature of WBC and decrease the training costs. The Wasserstein Generative adversarial network (WGAN) is used to create synthetic instances to enhance the size of training data. Experiments on two image datasets show the superiority of the proposed method over several state-of-the-art methods.
{"title":"Attention Residual Network for White Blood Cell Classification with WGAN Data Augmentation","authors":"Meng Zhao, Lingmin Jin, Shenghua Teng, Zuoyong Li","doi":"10.1109/ITME53901.2021.00075","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00075","url":null,"abstract":"In medicine, white blood cell (WBC) classification plays an important role in clinical diagnosis and treatment. Due to the similarity between classes and lack of training data, the precise classification of WBC is still challenging. To alleviate this problem, we propose an attention residual network for WBC image classification on the basis of data augmentation. Specifically, the attention residual network is composed of multiple attention residual blocks, an adaptive average pooling layer, and a full connection layer. The channel attention mechanism is introduced in each residual block to use the feature maps of WBC learned by a high layer to generate the attention map for a low layer. Each attention residual block also introduces depth separable convolution to extract the feature of WBC and decrease the training costs. The Wasserstein Generative adversarial network (WGAN) is used to create synthetic instances to enhance the size of training data. Experiments on two image datasets show the superiority of the proposed method over several state-of-the-art methods.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"41 1","pages":"336-340"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78931146","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 : 2021-11-01DOI: 10.1109/ITME53901.2021.00063
Lang Xi, Xinyu Jin
At present, most of the work is based on deep neural network to construct simultaneous peritoneal tumor detection algorithm. The prerequisite for the successful application of these algorithms is that the training set and the test set are independent and identically distributed, that is, the algorithm needs a large number of training samples with the same distribution as the target application. In order to effectively use the public data set with sufficient data to assist the training, and to get the model with superior performance index even when the data amount is limited, we propose a simultaneous peritoneal tumor detection algorithm based on domain adaptation. Specifically, we realize edge distribution alignment based on covariance matrix, and propose two constraints based on feature space optimization and conditional distribution alignment, so that the algorithm can effectively transfer knowledge by using data sets with the same tasks but different distributions. The model can learn the interface fitting to the specific data set even if there is only a small amount of labeled data. Extensive experiments show that the proposed algorithm based on domain adaptation can significantly improve the recognition performance of the model.
{"title":"Simultaneous Peritoneal Tumor Detection Algorithm based on Domain Adaptation","authors":"Lang Xi, Xinyu Jin","doi":"10.1109/ITME53901.2021.00063","DOIUrl":"https://doi.org/10.1109/ITME53901.2021.00063","url":null,"abstract":"At present, most of the work is based on deep neural network to construct simultaneous peritoneal tumor detection algorithm. The prerequisite for the successful application of these algorithms is that the training set and the test set are independent and identically distributed, that is, the algorithm needs a large number of training samples with the same distribution as the target application. In order to effectively use the public data set with sufficient data to assist the training, and to get the model with superior performance index even when the data amount is limited, we propose a simultaneous peritoneal tumor detection algorithm based on domain adaptation. Specifically, we realize edge distribution alignment based on covariance matrix, and propose two constraints based on feature space optimization and conditional distribution alignment, so that the algorithm can effectively transfer knowledge by using data sets with the same tasks but different distributions. The model can learn the interface fitting to the specific data set even if there is only a small amount of labeled data. Extensive experiments show that the proposed algorithm based on domain adaptation can significantly improve the recognition performance of the model.","PeriodicalId":6774,"journal":{"name":"2021 11th International Conference on Information Technology in Medicine and Education (ITME)","volume":"67 1","pages":"271-276"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86893959","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：