Pub Date : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065656
Lang Liu, Hengyi Yang, Yongqiang Huang
Based on the research of various information systems in air traffic management, this paper realizes the seamless communication between civil aviation systems through the mutual cooperation of information. When this achievement is applied to air transportation, it can achieve the following effects: (1) realize the sharing of civil aviation business information, support the collaborative decision-making of air traffic management, especially provide the Airport Collaborative Decision Making system with fast shared data, which can help the A-CDM system make fast and accurate decisions, and effectively solve the problems of flight delay and cancellation; (2) Improve the utilization rate of aviation data, so that airspace planning and route design are more scientific and reasonable, so that more flight routes can be used accurately, and improve the capacity of the existing civil aviation airspace system, which can improve the utilization rate of airspace, reduce flight delays and save energy.
{"title":"Sharing Big Data Storage for Air Traffic Management","authors":"Lang Liu, Hengyi Yang, Yongqiang Huang","doi":"10.1109/ICCC56324.2022.10065656","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065656","url":null,"abstract":"Based on the research of various information systems in air traffic management, this paper realizes the seamless communication between civil aviation systems through the mutual cooperation of information. When this achievement is applied to air transportation, it can achieve the following effects: (1) realize the sharing of civil aviation business information, support the collaborative decision-making of air traffic management, especially provide the Airport Collaborative Decision Making system with fast shared data, which can help the A-CDM system make fast and accurate decisions, and effectively solve the problems of flight delay and cancellation; (2) Improve the utilization rate of aviation data, so that airspace planning and route design are more scientific and reasonable, so that more flight routes can be used accurately, and improve the capacity of the existing civil aviation airspace system, which can improve the utilization rate of airspace, reduce flight delays and save energy.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":" 45","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113947110","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065677
Zhang Hongyi, W. Muqing, Zhao Min
As one of the most important tasks in the computer vision, vehicle re-identification aims to retrieve and identify the same vehicle under different surveillance cameras, which plays a key role in urban road traffic safety and intelligent traffic management system. However, the large intra-class difference and high inter-class similarity are still main challenges, as well as the diversity in lighting conditions, camera's shooting angle, and occlusion degrees. In order to further improve the average accuracy and algorithm performance, this paper proposes a vehicle re-identification algorithm based on multi-feature fusion and non-local operation. We embed non-local operation into the ResNet50 network, and employ feature slicing and reorganization to obtain multiple feature branches. Besides, learning rate warm-up and cosine annealing scheduler are also used. The experimental results show that our proposed method achieves higher accuracy on two commonly used datasets VeRi-776 and VehicleID.
{"title":"Multi-feature Fusion and Non-Local Operation for Vehicle Re-identification","authors":"Zhang Hongyi, W. Muqing, Zhao Min","doi":"10.1109/ICCC56324.2022.10065677","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065677","url":null,"abstract":"As one of the most important tasks in the computer vision, vehicle re-identification aims to retrieve and identify the same vehicle under different surveillance cameras, which plays a key role in urban road traffic safety and intelligent traffic management system. However, the large intra-class difference and high inter-class similarity are still main challenges, as well as the diversity in lighting conditions, camera's shooting angle, and occlusion degrees. In order to further improve the average accuracy and algorithm performance, this paper proposes a vehicle re-identification algorithm based on multi-feature fusion and non-local operation. We embed non-local operation into the ResNet50 network, and employ feature slicing and reorganization to obtain multiple feature branches. Besides, learning rate warm-up and cosine annealing scheduler are also used. The experimental results show that our proposed method achieves higher accuracy on two commonly used datasets VeRi-776 and VehicleID.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"190 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122426214","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065993
Dong Ma, B. Li, Bo Ran, Yonghao Wang, Xiao Huang, Kaibo Shi, Peng Kong, Wei Li
With the innovation of energy harvesting(EH) tech-nology and energy storage technology, renewable energy with energy storage batteries provides a new way to power future mobile communication base stations (BSs). However, a large number of BSs distributed energy storage resources are idle in most cases. In order to cope with this phenomenon, this study divides the battery energy storage zone into backup area and dispatchable capacity area according to the relationship between renewable energy collection and base station(BS) local load. On this basis, the battery control model and battery schedulable model are established to obtain the battery dispatchable capacity. In addition, deep Q learning (DQL) algorithms in machine learning are explored to optimize the model and maximize battery schedulable capacity. Finally, experimental cases show that battery energy dispatching is a win-win move for communication operators and distribution networks. Increasing the battery capacity can effectively smooth the local load curve of the distribution network.
{"title":"Green Base Station Battery Dispatchable Capacity Modeling and Optimization","authors":"Dong Ma, B. Li, Bo Ran, Yonghao Wang, Xiao Huang, Kaibo Shi, Peng Kong, Wei Li","doi":"10.1109/ICCC56324.2022.10065993","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065993","url":null,"abstract":"With the innovation of energy harvesting(EH) tech-nology and energy storage technology, renewable energy with energy storage batteries provides a new way to power future mobile communication base stations (BSs). However, a large number of BSs distributed energy storage resources are idle in most cases. In order to cope with this phenomenon, this study divides the battery energy storage zone into backup area and dispatchable capacity area according to the relationship between renewable energy collection and base station(BS) local load. On this basis, the battery control model and battery schedulable model are established to obtain the battery dispatchable capacity. In addition, deep Q learning (DQL) algorithms in machine learning are explored to optimize the model and maximize battery schedulable capacity. Finally, experimental cases show that battery energy dispatching is a win-win move for communication operators and distribution networks. Increasing the battery capacity can effectively smooth the local load curve of the distribution network.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127008790","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10066032
Danyang Wang, Peng Chen, Ruoyu Wang
Motivated by the improved direction of arrival (DOA) estimation performance by intelligent reflecting surface (IRS) in target detection systems, we propose a system using IRS for the target detection on the unmanned aerial vehicle (UAV) to improve the anti-interference capability and target detection performance. However, the UAV movement degrades the detection performance, so we formulate an UAV jitter model, in which the horizontal and vertical jitters with IRS model is considered. Then, we optimize the beamforming coefficients to maximize signal-to-noise ratio (SNR) of the received signals with UAV movement. Meanwhile, the performance improvement introduced by IRS is shown by the proposed optimization method with UAV. Simulation results illustrate that, When IRS is applied to UAV target detection, with the increased number of IRS units of IRS-aided UAV target detection system, the optimized method has better detection probability and anti-jitter interference capability compared with the existing non-IRS-assisted target detection systems.
{"title":"Cooperative Target Detection Based on UAV Jitter Model","authors":"Danyang Wang, Peng Chen, Ruoyu Wang","doi":"10.1109/ICCC56324.2022.10066032","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10066032","url":null,"abstract":"Motivated by the improved direction of arrival (DOA) estimation performance by intelligent reflecting surface (IRS) in target detection systems, we propose a system using IRS for the target detection on the unmanned aerial vehicle (UAV) to improve the anti-interference capability and target detection performance. However, the UAV movement degrades the detection performance, so we formulate an UAV jitter model, in which the horizontal and vertical jitters with IRS model is considered. Then, we optimize the beamforming coefficients to maximize signal-to-noise ratio (SNR) of the received signals with UAV movement. Meanwhile, the performance improvement introduced by IRS is shown by the proposed optimization method with UAV. Simulation results illustrate that, When IRS is applied to UAV target detection, with the increased number of IRS units of IRS-aided UAV target detection system, the optimized method has better detection probability and anti-jitter interference capability compared with the existing non-IRS-assisted target detection systems.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127019046","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065849
Wen-qiang Chen, Jiayi Zhang, Cheng Chi, Yan Luo, Xuemei Ding, Baoluo Ma
To align with the national climate pledges for long-term emissions goals, each industry needs consider using advanced digital technologies to mitigate carbon emissions and make progress in low-carbon transformation. The textile and apparel industry plays a very important role in economy and is one of the main contributors of greehouse gases emissions. This paper aims to show the feasibility of using blockchain technology for the carbon footprint of products accounting and trace. A framework of blockchain-based application of carbon footprint of products in textile industry has been proposed, in which a permissioned blockchchain infrastructure is integrated with enterprise information systems, automatic data collection, aggregation and reliable data sharing between multiple stakeholders is fulfilled. A case study of carbon footprint of a silk product has been conducted and the results show that the printing and dyeing process, and the silk reeling process are the most two contributors for carbon emissions, accounting for 44% of the total.
{"title":"Innovative Blockchain-Based Application of Carbon Footprint of Products: A Case Study in Textile and Apparel Industry","authors":"Wen-qiang Chen, Jiayi Zhang, Cheng Chi, Yan Luo, Xuemei Ding, Baoluo Ma","doi":"10.1109/ICCC56324.2022.10065849","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065849","url":null,"abstract":"To align with the national climate pledges for long-term emissions goals, each industry needs consider using advanced digital technologies to mitigate carbon emissions and make progress in low-carbon transformation. The textile and apparel industry plays a very important role in economy and is one of the main contributors of greehouse gases emissions. This paper aims to show the feasibility of using blockchain technology for the carbon footprint of products accounting and trace. A framework of blockchain-based application of carbon footprint of products in textile industry has been proposed, in which a permissioned blockchchain infrastructure is integrated with enterprise information systems, automatic data collection, aggregation and reliable data sharing between multiple stakeholders is fulfilled. A case study of carbon footprint of a silk product has been conducted and the results show that the printing and dyeing process, and the silk reeling process are the most two contributors for carbon emissions, accounting for 44% of the total.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121881434","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065909
Yunheng Luo, Jianshan Peng
As the core component of web browser, JavaScript engine has always been concerned about its security. Current state-of-the-art fuzzers for JavaScript engines mainly focus on generating correct and effective testcases by extracting semantic information from the initial corpus. However, we found that the existing fuzzers did not pay attention to the impact of branch conditions in the process of extracting semantic information, which led to incorrect testcases. To address this challenge, we propose a path-sensitive semantic analysis technique and implement it in a fuzz testing tool termed WindCore. Compared with the existing fuzzers, WindCore can more fully extract the semantic information in the initial corpus and generate testcases with correct syntax and semantics. Experimental results show that WindCore can greatly improve the correct rate of testcases with only a negligible performance overhead.
{"title":"WindCore: Path-Sensitive Semantic Analysis Technique for JavaScript Testcase Generation","authors":"Yunheng Luo, Jianshan Peng","doi":"10.1109/ICCC56324.2022.10065909","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065909","url":null,"abstract":"As the core component of web browser, JavaScript engine has always been concerned about its security. Current state-of-the-art fuzzers for JavaScript engines mainly focus on generating correct and effective testcases by extracting semantic information from the initial corpus. However, we found that the existing fuzzers did not pay attention to the impact of branch conditions in the process of extracting semantic information, which led to incorrect testcases. To address this challenge, we propose a path-sensitive semantic analysis technique and implement it in a fuzz testing tool termed WindCore. Compared with the existing fuzzers, WindCore can more fully extract the semantic information in the initial corpus and generate testcases with correct syntax and semantics. Experimental results show that WindCore can greatly improve the correct rate of testcases with only a negligible performance overhead.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132187531","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065850
Dejun Lei
Contrastive learning has been highly successful with computer vision and natural language processing. It can effectively address the under-sample situation. Contrastive learning has also been successfully implemented in recommender systems. It can not only address the problem of the small number of samples but also improve the learning impact of long-tailed data. Recommender systems contain large amounts of graph data. Graph neural networks are good at learning graph node representations. Through the neighbor information in the graph, it is possible to understand the potential intention of the user. Contrastive learning mainly includes sequence-based and graph-based contrastive learning in recommender systems. Currently, the modeling of both sequence contrastive learning and graph comparison learning in recommender systems is based on the user's single intent. However, the user's behavior consists of multiple intents. This paper proposes a new method which is named MIGC for modeling of user's numerous intents. Graph contrastive learning is introduced into the recommendation system recall algorithm and User's multi-interest modeling. This approach not only learns multiple users' intents but also improves the representation of long-tail data. Firstly, we construct a bipartite graph from user-to-item behavior data. Secondly, the multi-intents of users are a model of the graph. Finally, vector representations of users and items are obtained through contrastive learning of graph neural networks for vector recall in recommender systems. The experiments in this paper used the public dataset MovieLens and the private dataset e-commerce. And both offline and online have achieved a certain improvement. This study aims to start a new approach to users' multi-intent recall.
{"title":"MIGC: Multi-intent Graph Contrastive Learning in Recommendation","authors":"Dejun Lei","doi":"10.1109/ICCC56324.2022.10065850","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065850","url":null,"abstract":"Contrastive learning has been highly successful with computer vision and natural language processing. It can effectively address the under-sample situation. Contrastive learning has also been successfully implemented in recommender systems. It can not only address the problem of the small number of samples but also improve the learning impact of long-tailed data. Recommender systems contain large amounts of graph data. Graph neural networks are good at learning graph node representations. Through the neighbor information in the graph, it is possible to understand the potential intention of the user. Contrastive learning mainly includes sequence-based and graph-based contrastive learning in recommender systems. Currently, the modeling of both sequence contrastive learning and graph comparison learning in recommender systems is based on the user's single intent. However, the user's behavior consists of multiple intents. This paper proposes a new method which is named MIGC for modeling of user's numerous intents. Graph contrastive learning is introduced into the recommendation system recall algorithm and User's multi-interest modeling. This approach not only learns multiple users' intents but also improves the representation of long-tail data. Firstly, we construct a bipartite graph from user-to-item behavior data. Secondly, the multi-intents of users are a model of the graph. Finally, vector representations of users and items are obtained through contrastive learning of graph neural networks for vector recall in recommender systems. The experiments in this paper used the public dataset MovieLens and the private dataset e-commerce. And both offline and online have achieved a certain improvement. This study aims to start a new approach to users' multi-intent recall.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134339128","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065727
Zeke Wu, Haifeng Shuai, R. Liu, K. Guo, Shibing Zhu
With the arrival of the sixth generation (6G) mo-bile communication technology, one of its signature features is satellite-terrestrial fusion communication. Hence, integrated satellite multiple terrestrial relay network (ISMTRN) has become an emerging network architecture that has recently attracted much attention from scholars. Nevertheless, because of the open nature of the wireless channel, it inevitably leads to security issues. Moreover, covert communication technology has been considered one of the most promising and effective methods for secure communication. This paper explores the ISMTRN's covert performance with partial relay selection scheme. Particularly, one scenario that when a satellite communicates with one user via relaying, the specific relay may transmit covert information to the user is considered. Furthermore, we derive the closed-form expression of probability of detection error (PDE) to evaluate covert performance. Lastly, by virtue of numerical simulation, the influence of related system parameters on covert performance is investigated.
{"title":"Performance Analysis of Covert Communication Based on Integrated Satellite Multiple Terrestrial Relay Networks","authors":"Zeke Wu, Haifeng Shuai, R. Liu, K. Guo, Shibing Zhu","doi":"10.1109/ICCC56324.2022.10065727","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065727","url":null,"abstract":"With the arrival of the sixth generation (6G) mo-bile communication technology, one of its signature features is satellite-terrestrial fusion communication. Hence, integrated satellite multiple terrestrial relay network (ISMTRN) has become an emerging network architecture that has recently attracted much attention from scholars. Nevertheless, because of the open nature of the wireless channel, it inevitably leads to security issues. Moreover, covert communication technology has been considered one of the most promising and effective methods for secure communication. This paper explores the ISMTRN's covert performance with partial relay selection scheme. Particularly, one scenario that when a satellite communicates with one user via relaying, the specific relay may transmit covert information to the user is considered. Furthermore, we derive the closed-form expression of probability of detection error (PDE) to evaluate covert performance. Lastly, by virtue of numerical simulation, the influence of related system parameters on covert performance is investigated.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130294977","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065728
Ziyun Fu, Haiyang Liu
The Viterbi algorithm is one of the most commonly used methods for decoding convolutional codes, which outputs a maximum-likelihood codeword for the input sequence. However, the complexity of the Viterbi algorithm is high when the constraint length is large. To address this issue, we propose a hybrid algorithm that contains at most two stages for decoding convolutional codes in this paper. In the first stage, the normalized min-sum algorithm (NMSA) with a small number of iterations is applied. If the output of the NMSA is not a codeword, the scarce-state-transition (SST) Viterbi algorithm is invoked for the second stage of decoding. We provide a method for constructing the input vector of the SST Viterbi algorithm, from which a truncating method is further presented for complexity reduction. Simulation results on two rate-l/2 convolutional codes show that the proposed hybrid algorithm has little performance degradation compared with the Viterbi algorithm. Meanwhile, the complexity of the proposed hybrid algorithm is reduced, especially in the high signal-to-noise ratio region.
{"title":"Low-Complexity Hybrid Algorithm for Decoding Convolutional Codes","authors":"Ziyun Fu, Haiyang Liu","doi":"10.1109/ICCC56324.2022.10065728","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065728","url":null,"abstract":"The Viterbi algorithm is one of the most commonly used methods for decoding convolutional codes, which outputs a maximum-likelihood codeword for the input sequence. However, the complexity of the Viterbi algorithm is high when the constraint length is large. To address this issue, we propose a hybrid algorithm that contains at most two stages for decoding convolutional codes in this paper. In the first stage, the normalized min-sum algorithm (NMSA) with a small number of iterations is applied. If the output of the NMSA is not a codeword, the scarce-state-transition (SST) Viterbi algorithm is invoked for the second stage of decoding. We provide a method for constructing the input vector of the SST Viterbi algorithm, from which a truncating method is further presented for complexity reduction. Simulation results on two rate-l/2 convolutional codes show that the proposed hybrid algorithm has little performance degradation compared with the Viterbi algorithm. Meanwhile, the complexity of the proposed hybrid algorithm is reduced, especially in the high signal-to-noise ratio region.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130298516","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 : 2022-12-09DOI: 10.1109/ICCC56324.2022.10065731
Lixiang Deng, Bing Han, Shuang Ren
Most existing point cloud upsampling approaches focus on exploiting dense ground truth point clouds as supervised information to upsample sparse point clouds. However, it is arduous to collect such a high-quality paired sparse-dense dataset for training. Therefore, this paper proposes a novel unsupervised point cloud upsampling network, called UPU-DGTNet, which incorporates dynamic graph convolutions into the hierarchical transformers to better encode local and global point features and generate dense and uniform point clouds without using ground truth point clouds. Specifically, we first propose a dynamic graph transformer (DG T) module as a feature extractor to encode multi-scale local and global point features. In addition, we develop a transformer shuffle (TS) module as an upsampler that leverages the shifted channel cross attention (SCCA) to further aggregate and refine the multi-scale point features. Finally, we introduce the farthest point sample (FPS) method into the reconstruction loss and join the uniform loss to train the network so that the output points could preserve original geometric structures and be distributed uniformly. Various experiments on synthetic and real-scanned datasets demonstrate that our method can achieve impressive results and even competitive performances against some supervised methods.
{"title":"UPU-DGTNet: Dynamic Graph Transformer Network for Unsupervised Point Cloud Upsampling","authors":"Lixiang Deng, Bing Han, Shuang Ren","doi":"10.1109/ICCC56324.2022.10065731","DOIUrl":"https://doi.org/10.1109/ICCC56324.2022.10065731","url":null,"abstract":"Most existing point cloud upsampling approaches focus on exploiting dense ground truth point clouds as supervised information to upsample sparse point clouds. However, it is arduous to collect such a high-quality paired sparse-dense dataset for training. Therefore, this paper proposes a novel unsupervised point cloud upsampling network, called UPU-DGTNet, which incorporates dynamic graph convolutions into the hierarchical transformers to better encode local and global point features and generate dense and uniform point clouds without using ground truth point clouds. Specifically, we first propose a dynamic graph transformer (DG T) module as a feature extractor to encode multi-scale local and global point features. In addition, we develop a transformer shuffle (TS) module as an upsampler that leverages the shifted channel cross attention (SCCA) to further aggregate and refine the multi-scale point features. Finally, we introduce the farthest point sample (FPS) method into the reconstruction loss and join the uniform loss to train the network so that the output points could preserve original geometric structures and be distributed uniformly. Various experiments on synthetic and real-scanned datasets demonstrate that our method can achieve impressive results and even competitive performances against some supervised methods.","PeriodicalId":263098,"journal":{"name":"2022 IEEE 8th International Conference on Computer and Communications (ICCC)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115596303","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：