Pub Date : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135228
Anqi Zhao
Adversarial examples are images that can be easily misclassified by deep learning models when small, imperceptible changes are added to them. They pose a security concern for the use of DNNs in practical applications because of their transferability. One mainstream method for black-box targeted attacks is feature space attacks. They mainly focus on iterative attacks and have generally attempted to modify the intermediate feature of source category similar to the feature of target category. In this paper, we explore that certain characteristics of features, such as feature style statistics, might be able to better represent the features themselves, and potentially lead to better transferability compared to simply aligning the source and the target features directly. Based on this, we propose a method which use the relationship between the style statistics of intermediate features and its corresponding category to boost black-box targeted transferability in generative attacks. We demonstrate the effectiveness of this method through experimental results.
{"title":"Boosting black-box targeted transferability from feature statistic perspective","authors":"Anqi Zhao","doi":"10.1109/ICCECE58074.2023.10135228","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135228","url":null,"abstract":"Adversarial examples are images that can be easily misclassified by deep learning models when small, imperceptible changes are added to them. They pose a security concern for the use of DNNs in practical applications because of their transferability. One mainstream method for black-box targeted attacks is feature space attacks. They mainly focus on iterative attacks and have generally attempted to modify the intermediate feature of source category similar to the feature of target category. In this paper, we explore that certain characteristics of features, such as feature style statistics, might be able to better represent the features themselves, and potentially lead to better transferability compared to simply aligning the source and the target features directly. Based on this, we propose a method which use the relationship between the style statistics of intermediate features and its corresponding category to boost black-box targeted transferability in generative attacks. We demonstrate the effectiveness of this method through experimental results.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"113 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124907531","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 science and technology, the electronic control unit (ECU) equipment has gradually become complex, and the security of the controller area network (CAN) bus protocol as the standard protocol for ECU communication has also been higher. In recent years, researchers have conducted extensive research on the security of the CAN bus, such as the anti-counterfeiting attack mechanism based on intrusion detection technology and the authentication code technology, which has improved the security of the CAN bus. However, directly modifying the CAN protocol itself based on intrusion detection technology will cause insufficient backward compatibility and a large number of devices cannot be reused; The authentication code-based approach cannot trace the origin of counterfeit devices and forensic evidence of attack devices, so further improvements are needed. In view of the above two problems, this paper proposes a CAN terminal anti-counterfeiting attack method based on physical fingerprint, which calculates the autocorrelation difference between the collected original signal and the calculated reconstructed signal, and obtains the physical fingerprint of the device as the basis for CAN terminal identity authentication, thereby achieving anti-counterfeiting attack. In this paper, 30 CAN terminals are used to experimentally verify the method in this paper, collect and transmit random data and transfer fixed data CAN frames to form two sets of data sets, and train and test using linear discriminant models, and the final model can identify the device identity rate on the two data sets up to 99%.
{"title":"Anti-counterfeiting Method of CAN Terminal Based on Device Physical Fingerprint","authors":"Xiaozhong Hu, A. Hu, Jiabao Yu, Yanjun Ding, Hongxing Hu, Peng Guo","doi":"10.1109/ICCECE58074.2023.10135443","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135443","url":null,"abstract":"With the development of science and technology, the electronic control unit (ECU) equipment has gradually become complex, and the security of the controller area network (CAN) bus protocol as the standard protocol for ECU communication has also been higher. In recent years, researchers have conducted extensive research on the security of the CAN bus, such as the anti-counterfeiting attack mechanism based on intrusion detection technology and the authentication code technology, which has improved the security of the CAN bus. However, directly modifying the CAN protocol itself based on intrusion detection technology will cause insufficient backward compatibility and a large number of devices cannot be reused; The authentication code-based approach cannot trace the origin of counterfeit devices and forensic evidence of attack devices, so further improvements are needed. In view of the above two problems, this paper proposes a CAN terminal anti-counterfeiting attack method based on physical fingerprint, which calculates the autocorrelation difference between the collected original signal and the calculated reconstructed signal, and obtains the physical fingerprint of the device as the basis for CAN terminal identity authentication, thereby achieving anti-counterfeiting attack. In this paper, 30 CAN terminals are used to experimentally verify the method in this paper, collect and transmit random data and transfer fixed data CAN frames to form two sets of data sets, and train and test using linear discriminant models, and the final model can identify the device identity rate on the two data sets up to 99%.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128627105","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 the Visible Light Communication (VLC) system, the light emitting diode (LED) has a double function of illumination and communication. Aiming at the different distributions of luminous power and information access, a light source layout based on genetic algorithm is proposed. A $5mathrm{m}times 5mathrm{m}times 3mathrm{m}$ indoor VLC system model is established, and four LEDs are used as light sources to analyze the relationship between different light source layouts and the received light power distribution. Genetic algorithm is used to optimize the distance between LED light sources in a square room to seek the best light source layout scheme. The results show that the value of peak power deviation (PDP) is 1.34 compared with the traditional layout scheme, which can minimize the variation of light power on the receiving plane and improve signal transmission performance if this system is also used to achieved broad information access.
{"title":"A Light Source Layout Scheme for Indoor VLC System Application Using Genetic Algorithm","authors":"Renjie Zuo, An-rong Wang, Yufeng Shao, Chong Li, Peng-Ying Chen, Yanlin Li, Shuanfan Liu, J. Yuan","doi":"10.1109/ICCECE58074.2023.10135339","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135339","url":null,"abstract":"In the Visible Light Communication (VLC) system, the light emitting diode (LED) has a double function of illumination and communication. Aiming at the different distributions of luminous power and information access, a light source layout based on genetic algorithm is proposed. A $5mathrm{m}times 5mathrm{m}times 3mathrm{m}$ indoor VLC system model is established, and four LEDs are used as light sources to analyze the relationship between different light source layouts and the received light power distribution. Genetic algorithm is used to optimize the distance between LED light sources in a square room to seek the best light source layout scheme. The results show that the value of peak power deviation (PDP) is 1.34 compared with the traditional layout scheme, which can minimize the variation of light power on the receiving plane and improve signal transmission performance if this system is also used to achieved broad information access.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126469267","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135241
Caihong Wang
IPTV (Internet Protocol Television) provides interactive function, which greatly improves subjective experience of users. However, IP network is a best-effort network and does not guarantee a reliable connection, which makes video transmission based on IP network prone to problems. This paper focuses on the relationship between packet loss and video perception quality in IPTV transmission. Through subjective evaluation experiments, the influence of different packet loss rate and packet loss distributions (burst loss and random loss) are quantitatively studied. Experimental results show: When packet loss rate is the same, video with burst loss has a better quality than one with random loss. The proportion of burst loss in total packet loss increases, and subjective perception quality increases exponentially. Based on experimental results, this paper also designs an objective evaluation model of video quality based on packet loss rate and packet loss distribution.
{"title":"IPTV Video Perception Quality Based on Packet Loss Distribution","authors":"Caihong Wang","doi":"10.1109/ICCECE58074.2023.10135241","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135241","url":null,"abstract":"IPTV (Internet Protocol Television) provides interactive function, which greatly improves subjective experience of users. However, IP network is a best-effort network and does not guarantee a reliable connection, which makes video transmission based on IP network prone to problems. This paper focuses on the relationship between packet loss and video perception quality in IPTV transmission. Through subjective evaluation experiments, the influence of different packet loss rate and packet loss distributions (burst loss and random loss) are quantitatively studied. Experimental results show: When packet loss rate is the same, video with burst loss has a better quality than one with random loss. The proportion of burst loss in total packet loss increases, and subjective perception quality increases exponentially. Based on experimental results, this paper also designs an objective evaluation model of video quality based on packet loss rate and packet loss distribution.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"22 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127011568","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135296
Qiongmin Zhang, Ying Long, Hongshun Cai
As a degenerative neurodegenerative disease, early detection and intervention are still the key to prevent the aggravation of Alzheimer's disease (AD). Computer aided diagnosis based on Magnetic Resonance Imaging (MRI) can provide assistance for early detection of AD. In the application of limited MRI data, a multi-view lightweight neural network is proposed to effectively identify AD by reducing the amount of model computation and parameters and making full use of the image information. In the proposed method, we design an optimized lightweight neural network for feature extraction on limited MRI data. Attention mechanism is adopted to fuse multi-slice features and ensure the consistency of results. For full use of MRI data, personalized weights are learned and applied to fuse the multi-view features of transverse plane, sagittal plane and coronal plane. The experimental results indicate that compare with state-of-the-art lightweight methods, the proposed multi-view method achieves good balance on parameters (3.75M) and FLOPs (4.45G). Meanwhile, the classification performances of the proposed multi-view method are better in the three AD classification tasks. Through the construction of the model, it is expected to provide a feasible reference lightweight network for further AD detection.
{"title":"Classification of Alzheimer's Disease Based on Multi-view Lightweight Neural Network","authors":"Qiongmin Zhang, Ying Long, Hongshun Cai","doi":"10.1109/ICCECE58074.2023.10135296","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135296","url":null,"abstract":"As a degenerative neurodegenerative disease, early detection and intervention are still the key to prevent the aggravation of Alzheimer's disease (AD). Computer aided diagnosis based on Magnetic Resonance Imaging (MRI) can provide assistance for early detection of AD. In the application of limited MRI data, a multi-view lightweight neural network is proposed to effectively identify AD by reducing the amount of model computation and parameters and making full use of the image information. In the proposed method, we design an optimized lightweight neural network for feature extraction on limited MRI data. Attention mechanism is adopted to fuse multi-slice features and ensure the consistency of results. For full use of MRI data, personalized weights are learned and applied to fuse the multi-view features of transverse plane, sagittal plane and coronal plane. The experimental results indicate that compare with state-of-the-art lightweight methods, the proposed multi-view method achieves good balance on parameters (3.75M) and FLOPs (4.45G). Meanwhile, the classification performances of the proposed multi-view method are better in the three AD classification tasks. Through the construction of the model, it is expected to provide a feasible reference lightweight network for further AD detection.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127289348","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135411
Fengxiang Ni, Yufu Guo, Cunlong Zheng, D. Huang, Haoliang Sun, Y. Zhou
For the problems of large cost of manual troubleshooting and insufficient detail of operation and maintenance management for fault diagnosis in substation power's inner lines, the fault classification task is implemented for line faults in substations based on Optimum-Path Forest (OPF) algorithm using Python and Scikit-learn platform, and the classification accuracy of the algorithm is experimented on public fault dataset. The algorithm achieves a classification accuracy of 95.07%, which is better than traditional methods such as plain Bayes, random forest, and SVM. The experiments show that the algorithm can reduce the difficulty of the operation and inspection personnel to carry out detection, enhance the reliability of the substation line protection system, and provide a reliable power supply for the smart substation.
{"title":"Fault diagnosis of substation power's inner line based on Optimum-Path forest algorithm","authors":"Fengxiang Ni, Yufu Guo, Cunlong Zheng, D. Huang, Haoliang Sun, Y. Zhou","doi":"10.1109/ICCECE58074.2023.10135411","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135411","url":null,"abstract":"For the problems of large cost of manual troubleshooting and insufficient detail of operation and maintenance management for fault diagnosis in substation power's inner lines, the fault classification task is implemented for line faults in substations based on Optimum-Path Forest (OPF) algorithm using Python and Scikit-learn platform, and the classification accuracy of the algorithm is experimented on public fault dataset. The algorithm achieves a classification accuracy of 95.07%, which is better than traditional methods such as plain Bayes, random forest, and SVM. The experiments show that the algorithm can reduce the difficulty of the operation and inspection personnel to carry out detection, enhance the reliability of the substation line protection system, and provide a reliable power supply for the smart substation.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127730929","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135486
Juxin Yin, Lei Pan
Aiming at the shortcoming of existing pattern methods in handling massive data, a parallel hypernym relation extraction method on Spark is proposed. To improve the extraction accuracy, Combined with Spark's RDD programming model, an improved credibility algorithm(ppmit) is designed to identify the inverse hypernym relation; to address the data skew problem when calculating the credibility value of the hypernym relation, a data partitioning strategy - PID(Partition Index Dividing) algorithm is proposed to calculate the partition balance, add partition index to the overflow data, redivide the data, and ensure the partition data balance, thereby reducing the calculation time. Experiments conducted on the Chinese Wikipedia dataset show that the proposed method can guarantee the extraction accuracy and effectively improve the operational efficiency of the pattern extraction method.
针对现有模式方法在处理海量数据时存在的不足，提出了一种基于Spark的并行超词关系提取方法。为了提高提取精度，结合Spark的RDD编程模型，设计了一种改进的可信度算法(ppmit)来识别倒接词关系;为了解决计算上词关系可信度值时的数据倾斜问题，提出了一种数据分区策略——PID(Partition Index divide)算法，计算分区平衡，对溢出数据添加分区索引，重新划分数据，保证分区数据平衡，从而减少了计算时间。在中文维基百科数据集上进行的实验表明，该方法能够保证模式提取的准确性，有效提高模式提取方法的运行效率。
{"title":"Parallel hypernym relation extraction based on partition index dividing","authors":"Juxin Yin, Lei Pan","doi":"10.1109/ICCECE58074.2023.10135486","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135486","url":null,"abstract":"Aiming at the shortcoming of existing pattern methods in handling massive data, a parallel hypernym relation extraction method on Spark is proposed. To improve the extraction accuracy, Combined with Spark's RDD programming model, an improved credibility algorithm(ppmit) is designed to identify the inverse hypernym relation; to address the data skew problem when calculating the credibility value of the hypernym relation, a data partitioning strategy - PID(Partition Index Dividing) algorithm is proposed to calculate the partition balance, add partition index to the overflow data, redivide the data, and ensure the partition data balance, thereby reducing the calculation time. Experiments conducted on the Chinese Wikipedia dataset show that the proposed method can guarantee the extraction accuracy and effectively improve the operational efficiency of the pattern extraction method.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122569974","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135226
Dixing Zhou, Jing Zhou, Ou Ruan
To overcome the issues such as privacy leakage and Health Authorities passive tracking in the current contact tracing scheme, this paper proposes an active privacy-preserving contact tracing scheme for infectious diseases in cloud environments. In this scheme, Telecom Center timely collects participants' information, so that Health Authorities can proactively track the contacts of infectious diseases when needed, and participants needn't to install any additional Apps; Pseudorandom number generator is used to blind the participants' location information to protect their privacy; ElGamal encryption algorithm and its variants are introduced to encrypt the participants' identities, so that the Health Authorities can correctly obtain the information of contacts without revealing anything of other participants. Theoretical analysis and experimental simulation show that the scheme is secure and effective, and it also has good computation performance in large-scale data scenarios.
{"title":"Active Privacy-Preserving Contact Tracing Scheme for Infectious Diseases in Cloud Environment","authors":"Dixing Zhou, Jing Zhou, Ou Ruan","doi":"10.1109/ICCECE58074.2023.10135226","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135226","url":null,"abstract":"To overcome the issues such as privacy leakage and Health Authorities passive tracking in the current contact tracing scheme, this paper proposes an active privacy-preserving contact tracing scheme for infectious diseases in cloud environments. In this scheme, Telecom Center timely collects participants' information, so that Health Authorities can proactively track the contacts of infectious diseases when needed, and participants needn't to install any additional Apps; Pseudorandom number generator is used to blind the participants' location information to protect their privacy; ElGamal encryption algorithm and its variants are introduced to encrypt the participants' identities, so that the Health Authorities can correctly obtain the information of contacts without revealing anything of other participants. Theoretical analysis and experimental simulation show that the scheme is secure and effective, and it also has good computation performance in large-scale data scenarios.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127035345","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135274
Jiapei Zhang, Jingbing Wu, Hanxi Wang
Using a computer to splice images with overlapping parts is the current mainstream method for 3D panorama construction. However, the quality of the spliced sequence images cannot be guaranteed due to the uneven quality of the resulting panorama. Aiming at quality problems such as inconsistent brightness and contrast of stitched sequence images and easy noise generation during the shooting process, a batch quality improvement process for multiple sequence images required for 3D panorama construction is proposed, which is different from single image processing; firstly, the median value is used The filtering method is used to denoise multiple sequential images, secondly to make the gray range of multiple images consistent, and finally to improve the contrast of the sequential images by using histogram equalization. Experimental results show that this method has a good effect on improving the quality of 3D panorama stitching sequence images.
{"title":"Quality Analysis and Improvement of 3D Panoramic Stitching Sequence Image","authors":"Jiapei Zhang, Jingbing Wu, Hanxi Wang","doi":"10.1109/ICCECE58074.2023.10135274","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135274","url":null,"abstract":"Using a computer to splice images with overlapping parts is the current mainstream method for 3D panorama construction. However, the quality of the spliced sequence images cannot be guaranteed due to the uneven quality of the resulting panorama. Aiming at quality problems such as inconsistent brightness and contrast of stitched sequence images and easy noise generation during the shooting process, a batch quality improvement process for multiple sequence images required for 3D panorama construction is proposed, which is different from single image processing; firstly, the median value is used The filtering method is used to denoise multiple sequential images, secondly to make the gray range of multiple images consistent, and finally to improve the contrast of the sequential images by using histogram equalization. Experimental results show that this method has a good effect on improving the quality of 3D panorama stitching sequence images.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127619363","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135229
Siyuan Shen, W. Zhang
With the rise of the Metaverse, the need for efficient modeling of avatars becomes increasingly urgent. Building virtual human models from human image datasets has been a hot topic in computer vision. We used the speech synthesis technology to complete the conversion from text to speech waveform, and used the speech-lip shape generation method to generate a real person image with audio and video synchronization, finally used the thin plate spline transformation method to drive the virtual human image, and synthesizes a virtual human with audio and video synchronization image. Experimental results show that this method can effectively solve the problem of text-driven avatar lip mismatch and text-driven avatar audio and video asynchronous problems, and can synthesize high-quality, high-fidelity, low-latency avatars.
{"title":"A method for synthesizing dynamic image of virtual human","authors":"Siyuan Shen, W. Zhang","doi":"10.1109/ICCECE58074.2023.10135229","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135229","url":null,"abstract":"With the rise of the Metaverse, the need for efficient modeling of avatars becomes increasingly urgent. Building virtual human models from human image datasets has been a hot topic in computer vision. We used the speech synthesis technology to complete the conversion from text to speech waveform, and used the speech-lip shape generation method to generate a real person image with audio and video synchronization, finally used the thin plate spline transformation method to drive the virtual human image, and synthesizes a virtual human with audio and video synchronization image. Experimental results show that this method can effectively solve the problem of text-driven avatar lip mismatch and text-driven avatar audio and video asynchronous problems, and can synthesize high-quality, high-fidelity, low-latency avatars.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132824552","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：