Pub Date : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931257
Xi Lin, Lianfang Tian, Qiliang Du, Chuanbo Qin
This paper proposes an Improved Decision Based Adaptive Threshold Median Filter(IDBATMF) for fingerprint image salt and pepper noise denoising, which can more effectively remove the salt and pepper noise in fingerprint images and preserve the details of the image. The algorithm first detects noise, introduces the minimum absolute brightness difference to reflect the difference between candidate noise pixels and surrounding non-noise pixels, and innovatively sets a linear threshold, which is adapted to the local noise density and compared with the minimum absolute brightness difference to distinguish noise pixels. When removing noise, the window size used is determined by the extreme pixel density in the window, and the median of the non-extreme pixel in the window is used for replacement. The methods mentioned in this paper are compared with Standard Median Filter (SMF), Adaptive Median Filter (AMF), Modified Adaptive Median Filter (MAMF), Switching Median Filter (SWMF), Adaptive Switching Median Filter (ASMF), the Decision Based Algorithm (DBA), and the Modified Decision Based Algorithm (MDBA). The experimental results show that, compared with the existing methods, the method proposed in this paper better considers the local characteristics of the image, and has better processing effect under various noise densities.
{"title":"Improved Decision Based Adaptive Threshold Median Filter for Fingerprint Image Salt and Pepper Noise Denoising","authors":"Xi Lin, Lianfang Tian, Qiliang Du, Chuanbo Qin","doi":"10.1109/ISCIT55906.2022.9931257","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931257","url":null,"abstract":"This paper proposes an Improved Decision Based Adaptive Threshold Median Filter(IDBATMF) for fingerprint image salt and pepper noise denoising, which can more effectively remove the salt and pepper noise in fingerprint images and preserve the details of the image. The algorithm first detects noise, introduces the minimum absolute brightness difference to reflect the difference between candidate noise pixels and surrounding non-noise pixels, and innovatively sets a linear threshold, which is adapted to the local noise density and compared with the minimum absolute brightness difference to distinguish noise pixels. When removing noise, the window size used is determined by the extreme pixel density in the window, and the median of the non-extreme pixel in the window is used for replacement. The methods mentioned in this paper are compared with Standard Median Filter (SMF), Adaptive Median Filter (AMF), Modified Adaptive Median Filter (MAMF), Switching Median Filter (SWMF), Adaptive Switching Median Filter (ASMF), the Decision Based Algorithm (DBA), and the Modified Decision Based Algorithm (MDBA). The experimental results show that, compared with the existing methods, the method proposed in this paper better considers the local characteristics of the image, and has better processing effect under various noise densities.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124504742","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 : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931289
Qiwen Wu, Zijing Feng, Yingkai Huang, Jiehui Zhong, Xiaolong Liu
The research on the application of adversarial examples mainly focuses on considering adversarial examples as a threat in the past. In order to make better use of the adversarial examples, rather than just taking it as the defects of neural network, a covert communication method based on adversarial examples is proposed in this paper. Using the characteristics that adversarial examples can carry information, we combine it with specific coding rules to develop a covert communication algorithm. Unlike the traditional steganography, the secret information is not contained in the communication content of the sender and the receiver itself in the proposed scheme. The mapping relationship between adversarial examples and secret information is hidden in the neural network model, so as to realize the hidden transmission of information and improve the concealment and security of communication. At the same time, the tamper identification is embedded in the tensor of adversarial output. When the image is modified during transmission, the tamper identification will also change, so that the image can be authenticated. Experiments show the feasibility of the algorithm and verify that it can completely extract secret information from the encrypted image adversarial example and authenticate the integrity of the image.
{"title":"Covert Communication and Image Authentication Algorithm Based on Adversarial Examples","authors":"Qiwen Wu, Zijing Feng, Yingkai Huang, Jiehui Zhong, Xiaolong Liu","doi":"10.1109/ISCIT55906.2022.9931289","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931289","url":null,"abstract":"The research on the application of adversarial examples mainly focuses on considering adversarial examples as a threat in the past. In order to make better use of the adversarial examples, rather than just taking it as the defects of neural network, a covert communication method based on adversarial examples is proposed in this paper. Using the characteristics that adversarial examples can carry information, we combine it with specific coding rules to develop a covert communication algorithm. Unlike the traditional steganography, the secret information is not contained in the communication content of the sender and the receiver itself in the proposed scheme. The mapping relationship between adversarial examples and secret information is hidden in the neural network model, so as to realize the hidden transmission of information and improve the concealment and security of communication. At the same time, the tamper identification is embedded in the tensor of adversarial output. When the image is modified during transmission, the tamper identification will also change, so that the image can be authenticated. Experiments show the feasibility of the algorithm and verify that it can completely extract secret information from the encrypted image adversarial example and authenticate the integrity of the image.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125276041","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 : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931199
Chenggang Xu, Yi Wu, Lei Chen, Zihao Liu, Ceyao Cao
The third generation of Audio Video coding Standard (AVS3) is an emerging video coding standard that surpasses High Efficiency Video Coding (HEVC). AVS3 allows flexible block subdivision to achieve great coding performance by applying Quad- Tree (QT), Binary-Tree (BT) and Extend Quad-Tree (EQT) partition structure, while the increased flexibility comes at the cost of enormous computational complexity. In this paper, we present a novel fast partition algorithm for all coding units (CUs) based on Discrete cosine transform (DCT) coefficients in frequency domain, which can also obtain the image texture features like Spatial domain and guide to skip horizontal/vertical BT/EQT partition of CUs or early terminate the partition of flat CUs. Experimental results show that the proposed fast CU partition algorithm achieves about a 26 % encoding time reduction on average with only a 0.45 % Bjantegaard Delta Bit Rate (BDBR) increase for the AVS3 reference software HPM 14.0 under All-Intra configuration.
第三代音频视频编码标准(AVS3)是超越HEVC (High Efficiency Video coding)的新兴视频编码标准。AVS3允许灵活的块细分通过应用四叉树(QT)、二叉树(BT)和扩展四叉树(EQT)分区结构来实现出色的编码性能，而增加的灵活性是以巨大的计算复杂度为代价的。本文提出了一种基于频域离散余弦变换(DCT)系数对所有编码单元(cu)进行快速划分的新算法，该算法还可以获得图像的空间域等纹理特征，并指导跳过cu的水平/垂直BT/EQT划分或提前终止平面cu的划分。实验结果表明，对于AVS3参考软件HPM 14.0，在All-Intra配置下，所提出的快速CU划分算法平均可减少26%的编码时间，仅增加0.45%的BDBR (Bjantegaard Delta比特率)。
{"title":"Fast CU Partition Decision Algorithm for AVS3 Based on Frequency Domain","authors":"Chenggang Xu, Yi Wu, Lei Chen, Zihao Liu, Ceyao Cao","doi":"10.1109/ISCIT55906.2022.9931199","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931199","url":null,"abstract":"The third generation of Audio Video coding Standard (AVS3) is an emerging video coding standard that surpasses High Efficiency Video Coding (HEVC). AVS3 allows flexible block subdivision to achieve great coding performance by applying Quad- Tree (QT), Binary-Tree (BT) and Extend Quad-Tree (EQT) partition structure, while the increased flexibility comes at the cost of enormous computational complexity. In this paper, we present a novel fast partition algorithm for all coding units (CUs) based on Discrete cosine transform (DCT) coefficients in frequency domain, which can also obtain the image texture features like Spatial domain and guide to skip horizontal/vertical BT/EQT partition of CUs or early terminate the partition of flat CUs. Experimental results show that the proposed fast CU partition algorithm achieves about a 26 % encoding time reduction on average with only a 0.45 % Bjantegaard Delta Bit Rate (BDBR) increase for the AVS3 reference software HPM 14.0 under All-Intra configuration.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131235575","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 : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931189
Li Li, Hongfeng Tian, Yapeng Wang, Tiankui Zhang
With the development of Internet and mobile communication technology, the mobile network traffic is increasing at exponential rates. Edge caching is a promising technology to reduce network load and content distribution delay. Through content popularity prediction, cache revenue and network per-formance can be improved. This paper proposes a temporal graph convolutional network (TGCN) based content popularity prediction algorithm, which explore the spatial-temporal two-dimensional features in the cellular networks. The proposed TGCN algorithm captures the temporal-dimension dependence from the content request sequence in the base stations (BSs) and the spatial-dimension dependence from different BSs. Then the content request at each BS in the next time cycle is predicted by TGCN. Simulation results show that, compared with the existing algorithms, the proposed algorithm can effectively improve the prediction accuracy of content requests, at least 3%, and improve the cache hit rate of the networks.
{"title":"Spatial-Temporal Content Popularity Prediction in Cache Enabled Cellular Networks","authors":"Li Li, Hongfeng Tian, Yapeng Wang, Tiankui Zhang","doi":"10.1109/ISCIT55906.2022.9931189","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931189","url":null,"abstract":"With the development of Internet and mobile communication technology, the mobile network traffic is increasing at exponential rates. Edge caching is a promising technology to reduce network load and content distribution delay. Through content popularity prediction, cache revenue and network per-formance can be improved. This paper proposes a temporal graph convolutional network (TGCN) based content popularity prediction algorithm, which explore the spatial-temporal two-dimensional features in the cellular networks. The proposed TGCN algorithm captures the temporal-dimension dependence from the content request sequence in the base stations (BSs) and the spatial-dimension dependence from different BSs. Then the content request at each BS in the next time cycle is predicted by TGCN. Simulation results show that, compared with the existing algorithms, the proposed algorithm can effectively improve the prediction accuracy of content requests, at least 3%, and improve the cache hit rate of the networks.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134189713","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}
When making a new microprocessor (MPU), the instruction set architecture is considered. It is also necessary to design how to work the instructions in hardware circuits of the MPU. As a solution to these, we previously proposed a method of expressing the instruction in two ways of semantic expression and functional expression. We also developed a tool that can visualize the instruction. In this paper, we have improved this method by extending a usual MPU to a meta-instruction for any MPU. We have also developed a conversion method for the meta-instruction using a relationship between the semantic and the functional expressions. Then we verified how well our meta-instruction can represent the usual instructions. It is confirmed that most of them can be represented by the meta-instruction. Since our developed tool user can design the meta-instructions from either semantic or functional expression and both expressions can be automatically converted to each other, they can design the instruction of a new MPU and can understand hardware circuits for the MPU instruction by the functional expression.
{"title":"Visualization Tool of Microprocessor Instruction for Designing New Architecture with Meta-Instructions","authors":"Tomoyuki Morimoto, Shunsuke Onogi, Tadahiro Ogita, Ryota Kawamata, Ziming Wang, Toshiyuki Tsutsumi","doi":"10.1109/ISCIT55906.2022.9931311","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931311","url":null,"abstract":"When making a new microprocessor (MPU), the instruction set architecture is considered. It is also necessary to design how to work the instructions in hardware circuits of the MPU. As a solution to these, we previously proposed a method of expressing the instruction in two ways of semantic expression and functional expression. We also developed a tool that can visualize the instruction. In this paper, we have improved this method by extending a usual MPU to a meta-instruction for any MPU. We have also developed a conversion method for the meta-instruction using a relationship between the semantic and the functional expressions. Then we verified how well our meta-instruction can represent the usual instructions. It is confirmed that most of them can be represented by the meta-instruction. Since our developed tool user can design the meta-instructions from either semantic or functional expression and both expressions can be automatically converted to each other, they can design the instruction of a new MPU and can understand hardware circuits for the MPU instruction by the functional expression.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117097831","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}
With the vigorous development of the Internet of Things (IoT), technologies based on IoT have been widely used in various fields. Among them, the electronic auction technology based on the IoT has been applied. The traditional electronic auction scheme cannot be flexibly applied in IoT scenario and most auction schemes often have only one authority, which has a great burden on operation and low efficiency, resulting in a single-point performance bottleneck. In order to solve above problems, we use a hierarchical multiple authorities system to propose a Robust Auction scheme with Multiple Authorities in IoT (RAMA). Besides, a lightweight range query protocol which can get the highest bidder in electronic auction is adopted to avoid deploying expensive sensors and aggregators in the IoT configuration. Moreover, security of the hierarchical multiple authorities system have been proved through rigorous security analysis, which indicate that the adopted range query protocol not only ensures the security of encrypted bidding but also prevents it from being stolen by malicious adversaries. Finally, we conduct simulation experiments and analyze experimental results to verify the feasibility of the scheme. Specifically, when the number of bidders is 1000, the time-consuming to generate the secret key and the time-consuming to obtain the maximum bidding price are not exceed 0.5s in the simulation experiment.
{"title":"RAMA: Robust Auction Scheme with Multiple Authorities in IoT","authors":"Qian Meng, Zhuoyang Liang, Zhonghua Shen, Yong Liu, Yuying Liu, Jiamin Hu","doi":"10.1109/ISCIT55906.2022.9931197","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931197","url":null,"abstract":"With the vigorous development of the Internet of Things (IoT), technologies based on IoT have been widely used in various fields. Among them, the electronic auction technology based on the IoT has been applied. The traditional electronic auction scheme cannot be flexibly applied in IoT scenario and most auction schemes often have only one authority, which has a great burden on operation and low efficiency, resulting in a single-point performance bottleneck. In order to solve above problems, we use a hierarchical multiple authorities system to propose a Robust Auction scheme with Multiple Authorities in IoT (RAMA). Besides, a lightweight range query protocol which can get the highest bidder in electronic auction is adopted to avoid deploying expensive sensors and aggregators in the IoT configuration. Moreover, security of the hierarchical multiple authorities system have been proved through rigorous security analysis, which indicate that the adopted range query protocol not only ensures the security of encrypted bidding but also prevents it from being stolen by malicious adversaries. Finally, we conduct simulation experiments and analyze experimental results to verify the feasibility of the scheme. Specifically, when the number of bidders is 1000, the time-consuming to generate the secret key and the time-consuming to obtain the maximum bidding price are not exceed 0.5s in the simulation experiment.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117345203","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 : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931255
Yun Fen Yong, Chee Keong Tan, Ian K. T. Tan, Su-Wei Tan
Bluetooth low energy (BLE)-based fingerprinting technique has received great attention in indoor localization systems. Despite its significant advantages, the offline site surveys to collect fingerprints to construct a radio map for precise localization in the online phase remain the key challenge because it requires tremendous human effort, time, and cost. To alleviate this issue, this paper presents a novel fingerprint interpolation technique for constructing the radio map based on reference point (RP) clustering and the Voronoi diagram. Firstly, the collected RPs are clustered based on a threshold value of received signal strength difference using the proposed clustering algorithm. A Voronoi diagram is drawn using the centroid of each cluster to partition the clusters in which virtual fingerprints are then generated using the Kriging interpolation algorithm to build a complete radio map. By grouping RPs with similar characteristics in the same region, more accurate virtual fingerprints can be inferred since the RPs in the same region have the tendency to experience similar multipath fading and signal shadowing effects. Experimental results show that the proposed scheme reduces the localization error up to 14% compared to the interpolation without clustering. As a result, we can overcome the site survey issues for IPS by constructing a radio map with more accurate localization results.
{"title":"Radio Map Construction Using Fingerprints Clustering and Voronoi Diagram for Indoor Positioning","authors":"Yun Fen Yong, Chee Keong Tan, Ian K. T. Tan, Su-Wei Tan","doi":"10.1109/ISCIT55906.2022.9931255","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931255","url":null,"abstract":"Bluetooth low energy (BLE)-based fingerprinting technique has received great attention in indoor localization systems. Despite its significant advantages, the offline site surveys to collect fingerprints to construct a radio map for precise localization in the online phase remain the key challenge because it requires tremendous human effort, time, and cost. To alleviate this issue, this paper presents a novel fingerprint interpolation technique for constructing the radio map based on reference point (RP) clustering and the Voronoi diagram. Firstly, the collected RPs are clustered based on a threshold value of received signal strength difference using the proposed clustering algorithm. A Voronoi diagram is drawn using the centroid of each cluster to partition the clusters in which virtual fingerprints are then generated using the Kriging interpolation algorithm to build a complete radio map. By grouping RPs with similar characteristics in the same region, more accurate virtual fingerprints can be inferred since the RPs in the same region have the tendency to experience similar multipath fading and signal shadowing effects. Experimental results show that the proposed scheme reduces the localization error up to 14% compared to the interpolation without clustering. As a result, we can overcome the site survey issues for IPS by constructing a radio map with more accurate localization results.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125638967","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 : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931221
Zhenjie Pan, Wei Xia
Bacterial foraging optimization (BFO) has attracted much attention and been widely applied in a variety of scientific and engineering applications since its inception. This paper proposes a DV -Hop localization algorithm based on bacterial foraging optimization by analysing the DV -Hop localization algorithm which lack of positioning in the wireless sensor network location. To improve the positioning accuracy of DV -Hop algorithm the algorithm use bacterial foraging algorithm to locate the position of the third step in the DV-Hop unknown node. Experimental results show that the algorithm proposed in this paper has better localization effect than the basic DV _Hop algorithm, and it effectively improves the location accuracy, which proves the validity of our proposed method.
{"title":"Research on Bacterial Foraging Optimization Algorithm for DV-Hop Localization","authors":"Zhenjie Pan, Wei Xia","doi":"10.1109/ISCIT55906.2022.9931221","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931221","url":null,"abstract":"Bacterial foraging optimization (BFO) has attracted much attention and been widely applied in a variety of scientific and engineering applications since its inception. This paper proposes a DV -Hop localization algorithm based on bacterial foraging optimization by analysing the DV -Hop localization algorithm which lack of positioning in the wireless sensor network location. To improve the positioning accuracy of DV -Hop algorithm the algorithm use bacterial foraging algorithm to locate the position of the third step in the DV-Hop unknown node. Experimental results show that the algorithm proposed in this paper has better localization effect than the basic DV _Hop algorithm, and it effectively improves the location accuracy, which proves the validity of our proposed method.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125929464","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 : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931323
Tianxiang Cao, Xin Song, Jun Wang
The Invoice Disputes Team wants to identify further opportunities to reduce invoice disputes and the team would like to explore whether data analytics can drive further improvements. The purpose of this paper is to compare the effectiveness of the eight approaches to predict binary dependent variables according to the specified data. The techniques examined are Logistic Regression, Probit Regression, CHAID, CART, Neural Networks, Bagging, Random Forests and Boosting. This paper describes the data set, the effectiveness measures used and the approaches, and also shows the results for each of the eight approaches that are examined. The simulation results show that both Bagging and Random forests seem to do better than other approaches.
{"title":"A Comparison of the Effectiveness of Techniques for Predicting Binary Dependent Variables","authors":"Tianxiang Cao, Xin Song, Jun Wang","doi":"10.1109/ISCIT55906.2022.9931323","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931323","url":null,"abstract":"The Invoice Disputes Team wants to identify further opportunities to reduce invoice disputes and the team would like to explore whether data analytics can drive further improvements. The purpose of this paper is to compare the effectiveness of the eight approaches to predict binary dependent variables according to the specified data. The techniques examined are Logistic Regression, Probit Regression, CHAID, CART, Neural Networks, Bagging, Random Forests and Boosting. This paper describes the data set, the effectiveness measures used and the approaches, and also shows the results for each of the eight approaches that are examined. The simulation results show that both Bagging and Random forests seem to do better than other approaches.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126119376","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 : 2022-09-27DOI: 10.1109/ISCIT55906.2022.9931195
Jing Wang, Siwen Wei, Kexin Wang, Jianhong Li
Aiming at the problems of low detection accuracy and high false detection rate in traditional defect detection methods, an improved YOLOv4 model for surface defect detection of aircraft glass canopy is proposed. In this paper, an improved FReLU activation function is used instead of Mish to better adaptively capture the spatial correlation to improve the defect detection efficiency. It is worth mentioning that due to the high aspect ratio of defects, this paper defines the rotated rectangle method without adding rotation anchors, which greatly improves the prediction speed of the model. Finally, this paper solves the problem of model fitting by optimizing the loss function of the model. Experiments show that the mAP value of our model on the test set reaches 93.34 %, which is 4.58 % higher than original YOLOv4 model, which proves the effectiveness of our model on the aircraft glass canopy surface defect dataset.
{"title":"Surface Defect Detection of Aircraft Glass Canopy Based on Improved YOLOv4","authors":"Jing Wang, Siwen Wei, Kexin Wang, Jianhong Li","doi":"10.1109/ISCIT55906.2022.9931195","DOIUrl":"https://doi.org/10.1109/ISCIT55906.2022.9931195","url":null,"abstract":"Aiming at the problems of low detection accuracy and high false detection rate in traditional defect detection methods, an improved YOLOv4 model for surface defect detection of aircraft glass canopy is proposed. In this paper, an improved FReLU activation function is used instead of Mish to better adaptively capture the spatial correlation to improve the defect detection efficiency. It is worth mentioning that due to the high aspect ratio of defects, this paper defines the rotated rectangle method without adding rotation anchors, which greatly improves the prediction speed of the model. Finally, this paper solves the problem of model fitting by optimizing the loss function of the model. Experiments show that the mAP value of our model on the test set reaches 93.34 %, which is 4.58 % higher than original YOLOv4 model, which proves the effectiveness of our model on the aircraft glass canopy surface defect dataset.","PeriodicalId":325919,"journal":{"name":"2022 21st International Symposium on Communications and Information Technologies (ISCIT)","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122449458","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：