Xinjian Wei, Kaidi Wang, Guangxu Li, Hyoungseop Kim
In order to get with the environmental changes of battlefield quickly, the military camouflage should be changeable. If the camouflage patterns of the clothes and vehicles like tanks are different from the environment, it's very easy for cameras of enemies to find. As we all know that the same patterns is used in the most of current military all over the world. In this paper, we propose a novel feature-extraction method from an image using convolutional neural networks. Then the pattern will be combined with the environmental style pattern. The composite image is mapped onto the surface of the actual 3D clothes and vehicles finally. In this paper, the Eye-Movement equipment is applied to evaluate the results for better comparison. We can produce the proper pattern even the different and complicated environment.
{"title":"An Automatic Design of Camouflage Patterns Based on CNNs","authors":"Xinjian Wei, Kaidi Wang, Guangxu Li, Hyoungseop Kim","doi":"10.1145/3404555.3404637","DOIUrl":"https://doi.org/10.1145/3404555.3404637","url":null,"abstract":"In order to get with the environmental changes of battlefield quickly, the military camouflage should be changeable. If the camouflage patterns of the clothes and vehicles like tanks are different from the environment, it's very easy for cameras of enemies to find. As we all know that the same patterns is used in the most of current military all over the world. In this paper, we propose a novel feature-extraction method from an image using convolutional neural networks. Then the pattern will be combined with the environmental style pattern. The composite image is mapped onto the surface of the actual 3D clothes and vehicles finally. In this paper, the Eye-Movement equipment is applied to evaluate the results for better comparison. We can produce the proper pattern even the different and complicated environment.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123223605","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 development of unmanned aerial vehicle (UAV) technology, more and more unmanned aerial vehicles have entered various fields, including the aspect of moving target detection. Due to the characteristics of the drone, the captured video does not have a fixed background, which makes the traditional moving target detection algorithm unusable. Herein, this paper proposes a self-matching detection algorithm based on Gaussian mixture model and CSIFT. This algorithm is applied to unmanned On board, under the background of moving targets, the target can be identified when there is relative movement between the target and the background.
{"title":"Self-Matching Moving Target Detection Algorithm Based On Gaussian Mixture Model And CSIFT","authors":"Qiang Wu, Jiaxiang Zhao, Xin Zheng","doi":"10.1145/3404555.3404648","DOIUrl":"https://doi.org/10.1145/3404555.3404648","url":null,"abstract":"With the development of unmanned aerial vehicle (UAV) technology, more and more unmanned aerial vehicles have entered various fields, including the aspect of moving target detection. Due to the characteristics of the drone, the captured video does not have a fixed background, which makes the traditional moving target detection algorithm unusable. Herein, this paper proposes a self-matching detection algorithm based on Gaussian mixture model and CSIFT. This algorithm is applied to unmanned On board, under the background of moving targets, the target can be identified when there is relative movement between the target and the background.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124949733","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}
Malicious encrypted traffic poses great threat to cyber security owing to encryption and the ability to bypass traditional traffic detection schemes. Malicious encrypted traffic identification is a challenging task and has attracted researchers' attention nowadays. Existing research way mainly extracts various statistical features of data-flow, which relies artificial experience heavily. To round the above problem. a fingerprint enhancement and second-order Markov chain based scheme is proposed in this paper, obtaining features more easily. Fingerprint enhancement is done to replace SSL fingerprint by refining data-flow's behavior. Then enhanced fingerprint is fed to second-order Markov chain to obtain dominating feature for identification model. To our best knowledge, this paper is the first one focusing on using fingerprint and second order Markov chain to simplify feature extraction. Finally, the proposed scheme is verified based on public dataset Stratosphere IPS.
{"title":"A Fingerprint Enhancement and Second-Order Markov Chain Based Malicious Encrypted Traffic Identification Scheme","authors":"Daichong Chao","doi":"10.1145/3404555.3404590","DOIUrl":"https://doi.org/10.1145/3404555.3404590","url":null,"abstract":"Malicious encrypted traffic poses great threat to cyber security owing to encryption and the ability to bypass traditional traffic detection schemes. Malicious encrypted traffic identification is a challenging task and has attracted researchers' attention nowadays. Existing research way mainly extracts various statistical features of data-flow, which relies artificial experience heavily. To round the above problem. a fingerprint enhancement and second-order Markov chain based scheme is proposed in this paper, obtaining features more easily. Fingerprint enhancement is done to replace SSL fingerprint by refining data-flow's behavior. Then enhanced fingerprint is fed to second-order Markov chain to obtain dominating feature for identification model. To our best knowledge, this paper is the first one focusing on using fingerprint and second order Markov chain to simplify feature extraction. Finally, the proposed scheme is verified based on public dataset Stratosphere IPS.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129631450","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}
Aiming at the task offloading in edge computing, this paper proposes a task offloading algorithm that utilizes the cache function of the edge server. When a task is uploaded to the edge server, it is determined whether this type of task is uploaded to the edge server for the first time. If it is determined to be yes, then the corresponding task content is cached on this server, and each task is set to have been cached the same. The execution delay and energy consumption of computing nodes of type content will be reduced. When a task that has cached the same type of content is uploaded to the edge server, the pre-set evaluation parameters take into account factors such as delay and energy consumption to calculate the optimal processing position for the task, so that each uploaded task can be executed in the least costly manner, thereby achieving the purpose of reducing the operating cost of the entire system. We use iFogSim simulator for simulation experiments. Simulation results show that this strategy can effectively reduce the latency and energy consumption of edge systems. In the ideal case, the task latency and energy consumption are reduced by 15% and 18% respectively compared to the original algorithm.
{"title":"Offloading Strategy for Edge Computing Tasks Based on Cache Mechanism","authors":"Juan Fang, Wenzheng Zeng","doi":"10.1145/3404555.3404575","DOIUrl":"https://doi.org/10.1145/3404555.3404575","url":null,"abstract":"Aiming at the task offloading in edge computing, this paper proposes a task offloading algorithm that utilizes the cache function of the edge server. When a task is uploaded to the edge server, it is determined whether this type of task is uploaded to the edge server for the first time. If it is determined to be yes, then the corresponding task content is cached on this server, and each task is set to have been cached the same. The execution delay and energy consumption of computing nodes of type content will be reduced. When a task that has cached the same type of content is uploaded to the edge server, the pre-set evaluation parameters take into account factors such as delay and energy consumption to calculate the optimal processing position for the task, so that each uploaded task can be executed in the least costly manner, thereby achieving the purpose of reducing the operating cost of the entire system. We use iFogSim simulator for simulation experiments. Simulation results show that this strategy can effectively reduce the latency and energy consumption of edge systems. In the ideal case, the task latency and energy consumption are reduced by 15% and 18% respectively compared to the original algorithm.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129193363","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 rapid development of mobile internet technology, mobile terminal APP is closely related to the daily life of contemporary university students. This paper analyzes the algorithm of location data by obtaining the location data samples of university student terminal APP and defines the positions in semantics by obtaining the trajectory route of students' behavior. Analyze the daily behavior habits, social relationship, personal interests and hobbies and so on of university students by characteristics of university students' daily behaviors and the data of their moving tracks. In order to build a computational model for similarity of university students' user behavior to analyze the behavior of university students and induce four user types of university students with similar behaviors according to the classification of user behavior of university students by similarity calculation model. In order to carry out corresponding learning career planning guidance for different types of university students to establish a prediction model for the growth of university students which has been verified that it has a practical guiding significance by tests.
{"title":"Research on Prediction Model of College Students' Growth Based on Mobile Location Data Mining: Take Hunan Mass Media College as an example","authors":"Yanshu Liu, Can Yi","doi":"10.1145/3404555.3404580","DOIUrl":"https://doi.org/10.1145/3404555.3404580","url":null,"abstract":"With the rapid development of mobile internet technology, mobile terminal APP is closely related to the daily life of contemporary university students. This paper analyzes the algorithm of location data by obtaining the location data samples of university student terminal APP and defines the positions in semantics by obtaining the trajectory route of students' behavior. Analyze the daily behavior habits, social relationship, personal interests and hobbies and so on of university students by characteristics of university students' daily behaviors and the data of their moving tracks. In order to build a computational model for similarity of university students' user behavior to analyze the behavior of university students and induce four user types of university students with similar behaviors according to the classification of user behavior of university students by similarity calculation model. In order to carry out corresponding learning career planning guidance for different types of university students to establish a prediction model for the growth of university students which has been verified that it has a practical guiding significance by tests.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130009189","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}
In modern cities, numerous urban perception devices collect and release urban data all the time, but urban data may become abnormal due to environmental interference or artificial tampering. In view of the problem that urban data will face data anomalies, this paper designs a distributed gauss membership anomaly data filtering algorithm, and defines a set of extraction protocols suitable for heterogeneous data. Simulation results show that this algorithm can filter abnormal data in real time, improve the efficiency of urban computing and reduce the cost of network.
{"title":"A Distributed Anomaly Filtering Algorithm for Heterogeneous Data Based on City Computing","authors":"Shiwei Wang, Yangyang Li, Xiaobin Xu, Guijie Yue","doi":"10.1145/3404555.3404636","DOIUrl":"https://doi.org/10.1145/3404555.3404636","url":null,"abstract":"In modern cities, numerous urban perception devices collect and release urban data all the time, but urban data may become abnormal due to environmental interference or artificial tampering. In view of the problem that urban data will face data anomalies, this paper designs a distributed gauss membership anomaly data filtering algorithm, and defines a set of extraction protocols suitable for heterogeneous data. Simulation results show that this algorithm can filter abnormal data in real time, improve the efficiency of urban computing and reduce the cost of network.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126388519","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}
In view of the problem that the number of clusters need to be set manually, it is difficult to process the multi-dimensional data effectively, and the clustering results are not described effectively when the multi-dimensional data need to be clustered. This paper proposes a method of adaptive spatial clustering and its cloud model representation for the multi-dimensional data. This method can be used to cluster multi-dimensional spatial data, form qualitative description of clustering results, and realize the reconstruction and verification of qualitative description features. Through simulation experiments, this method can cluster data adaptively without the need to set the number of clusters. At the same time, it has a good ability to abstract and reconstruct digital features.
{"title":"Adaptive Spatial Clustering for Multi-Dimensional Data and Its Cloud Model Representation","authors":"Bin Gao, Xinhai Zhang, Xiaobin Xu, Yifeng Liu","doi":"10.1145/3404555.3404634","DOIUrl":"https://doi.org/10.1145/3404555.3404634","url":null,"abstract":"In view of the problem that the number of clusters need to be set manually, it is difficult to process the multi-dimensional data effectively, and the clustering results are not described effectively when the multi-dimensional data need to be clustered. This paper proposes a method of adaptive spatial clustering and its cloud model representation for the multi-dimensional data. This method can be used to cluster multi-dimensional spatial data, form qualitative description of clustering results, and realize the reconstruction and verification of qualitative description features. Through simulation experiments, this method can cluster data adaptively without the need to set the number of clusters. At the same time, it has a good ability to abstract and reconstruct digital features.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126398498","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}
Traditionally, the dice loss compares the similarity of boundaries between ground truths and predictions. However, the result can be unauthentic when it comes to the situation that both ground truth and predictions are too small. The focal Tversky loss is proposed to address the imbalance between positive and negative samples as well as to contribute a better trade-off between precision and recall. In this paper, we introduce a novel loss function named 'Maven Loss' by considering 'specificity' to handle the issue of data disequilibrium and to help achieve weighing both abilities to correctly segment lesion and non-lesion areas. To evaluate our loss function, we also propose an AGW-Net based on the attention U-Net and W-Net by injecting self-reinforced skip connections. Experiment on ISIC 2018 dataset in which lesions occupy 21.4% on average of the whole images shows that maven loss function and the new network architecture improved IOU and F1-score by 4.9% and 3% compared to the standard attention U-Net, respectively.
{"title":"Maven Loss with AGW-Net for Biomedical Image Segmentation","authors":"Yuze Li, Kaijun Wang, Hehui Gu","doi":"10.1145/3404555.3404561","DOIUrl":"https://doi.org/10.1145/3404555.3404561","url":null,"abstract":"Traditionally, the dice loss compares the similarity of boundaries between ground truths and predictions. However, the result can be unauthentic when it comes to the situation that both ground truth and predictions are too small. The focal Tversky loss is proposed to address the imbalance between positive and negative samples as well as to contribute a better trade-off between precision and recall. In this paper, we introduce a novel loss function named 'Maven Loss' by considering 'specificity' to handle the issue of data disequilibrium and to help achieve weighing both abilities to correctly segment lesion and non-lesion areas. To evaluate our loss function, we also propose an AGW-Net based on the attention U-Net and W-Net by injecting self-reinforced skip connections. Experiment on ISIC 2018 dataset in which lesions occupy 21.4% on average of the whole images shows that maven loss function and the new network architecture improved IOU and F1-score by 4.9% and 3% compared to the standard attention U-Net, respectively.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132713591","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}
Xin Tao, Didem Gürdür Broo, Martin Törngren, De-Jiu Chen
Cyber-Physical Systems (CPS) are a result of highly cross-disciplinary processes and are evolving to perform increasingly challenging tasks in dynamically changing environments. This leads to an increasing CPS complexity and therefore the management of uncertainty to ensure the trustworthiness of these systems is needed. Our paper focuses on uncertainty management (UM) both in general and more specifically in the context of CPS situation awareness (SA). The motivation behind this is the important role of SA and its many inherent uncertainties. To this end, firstly, a literature review is conducted to acquire the state of the art of UM. Later, we present findings and observations from the literature review, with two main challenges identified - inconsistent understanding and terminology among a multitude of uncertainty perspectives, and a lack of collaboration among different communities. On this basis, lastly, two case studies are conducted to exemplify the challenges and provide brief ideas on how to deal with them. The whole investigation in the paper suggests an urgent strengthening of common understanding through enhanced collaboration and regulations.
{"title":"Uncertainty Management in Situation Awareness for Cyber-Physical Systems: State of the Art and Challenge","authors":"Xin Tao, Didem Gürdür Broo, Martin Törngren, De-Jiu Chen","doi":"10.1145/3404555.3404558","DOIUrl":"https://doi.org/10.1145/3404555.3404558","url":null,"abstract":"Cyber-Physical Systems (CPS) are a result of highly cross-disciplinary processes and are evolving to perform increasingly challenging tasks in dynamically changing environments. This leads to an increasing CPS complexity and therefore the management of uncertainty to ensure the trustworthiness of these systems is needed. Our paper focuses on uncertainty management (UM) both in general and more specifically in the context of CPS situation awareness (SA). The motivation behind this is the important role of SA and its many inherent uncertainties. To this end, firstly, a literature review is conducted to acquire the state of the art of UM. Later, we present findings and observations from the literature review, with two main challenges identified - inconsistent understanding and terminology among a multitude of uncertainty perspectives, and a lack of collaboration among different communities. On this basis, lastly, two case studies are conducted to exemplify the challenges and provide brief ideas on how to deal with them. The whole investigation in the paper suggests an urgent strengthening of common understanding through enhanced collaboration and regulations.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"96 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133674486","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}
Voice conversion towards a specific speaker requires a large number of target speaker's utterances, which is expensive in practice. This paper proposes a speaker-adaptive voice conversion (SAVC) system, which accomplishes voice conversion towards arbitrary speakers with limited data. First, a multi-speaker voice conversion (MSVC) model is trained to learn the shared information between speakers and build a speaker latent space. Second, utterances of a new target speaker are used to fine tune the MSVC model aiming to learn the voice of the target speaker. In the two steps, phonetic posteriorgrams (PPGs), a speaker-independent linguistic feature, and speaker embeddings such as i-vector or x-vector are encoded to train the model. In order to achieve better results, two different adaptive approaches are explored: adaptation on the whole MSVC model or additional linear-hidden layers (AHL). As the results show, both adaptive approaches significantly outperform the MSVC model without adaptation. Besides, the whole adapted model based on x-vector gets a higher similarity to target speaker within 10 utterances.
{"title":"Voice Conversion towards Arbitrary Speakers With Limited Data","authors":"Ying Zhang, Wenjun Zhang, Dandan Song","doi":"10.1145/3404555.3404627","DOIUrl":"https://doi.org/10.1145/3404555.3404627","url":null,"abstract":"Voice conversion towards a specific speaker requires a large number of target speaker's utterances, which is expensive in practice. This paper proposes a speaker-adaptive voice conversion (SAVC) system, which accomplishes voice conversion towards arbitrary speakers with limited data. First, a multi-speaker voice conversion (MSVC) model is trained to learn the shared information between speakers and build a speaker latent space. Second, utterances of a new target speaker are used to fine tune the MSVC model aiming to learn the voice of the target speaker. In the two steps, phonetic posteriorgrams (PPGs), a speaker-independent linguistic feature, and speaker embeddings such as i-vector or x-vector are encoded to train the model. In order to achieve better results, two different adaptive approaches are explored: adaptation on the whole MSVC model or additional linear-hidden layers (AHL). As the results show, both adaptive approaches significantly outperform the MSVC model without adaptation. Besides, the whole adapted model based on x-vector gets a higher similarity to target speaker within 10 utterances.","PeriodicalId":220526,"journal":{"name":"Proceedings of the 2020 6th International Conference on Computing and Artificial Intelligence","volume":"225 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133846013","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：