The formal verification of the operating systems in power monitoring system is an important means to ensure the security of the operating system in power monitoring system. This paper introduces the verification principles and framework of formal verification of operating systems in power monitoring system, the languages and tools used in formal verification, and some classic projects of formal verification of operating systems. Through the introduction of the related content of the formalization of these operating systems, some ideas and future development trends of the formal verification of the current operating systems are explained. It has completed the verification process, beginning with weak type safety and progressing to functional correctness, proof of the high-level abstract protocol, and modification of the low-level code. These gain from the constant advancement and refinement of tools and technologies for formal verification of operating systems, but it is also subject to formal verification tools and techniques, and cannot genuinely go towards the last practical link of production. The automated research on formal verification tools and technologies will continue to be a significant advance in operating system formal verification.
{"title":"Survey of the Formal Verification of Operating Systems in Power Monitoring System","authors":"Kangle Yang, Jianye Yu, Xinshen Wei, Feng You, Haidong Huang, Xuesong Huo","doi":"10.1145/3609703.3609714","DOIUrl":"https://doi.org/10.1145/3609703.3609714","url":null,"abstract":"The formal verification of the operating systems in power monitoring system is an important means to ensure the security of the operating system in power monitoring system. This paper introduces the verification principles and framework of formal verification of operating systems in power monitoring system, the languages and tools used in formal verification, and some classic projects of formal verification of operating systems. Through the introduction of the related content of the formalization of these operating systems, some ideas and future development trends of the formal verification of the current operating systems are explained. It has completed the verification process, beginning with weak type safety and progressing to functional correctness, proof of the high-level abstract protocol, and modification of the low-level code. These gain from the constant advancement and refinement of tools and technologies for formal verification of operating systems, but it is also subject to formal verification tools and techniques, and cannot genuinely go towards the last practical link of production. The automated research on formal verification tools and technologies will continue to be a significant advance in operating system formal verification.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"586 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116176002","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}
Reinforcement learning algorithms have been used to discover the strategies in game theory. This study investigates whether Q learning, one of the classic reinforcement learning methods, is capable of training bargaining players via self-play, a training paradigm used by AlphaGo, to maximum their profit. We also compare our empirical results with the known theoretic solutions and perform an comprehensive analysis upon their differences. To accomplish these, we come up with two policy updating methods used in the training process, namely alternate update and simultaneous update, which are tailored for two players who propose offers and counter-offers in an alternating manner under a time constraint enforced by the discount factors. Our experimental results have demonstrated that the values of the discount factor actually have tangible impact on how far the bargaining outcomes deviate from the game theoretic solutions.
{"title":"Policy Updating Methods of Q Learning for Two Player Bargaining Game","authors":"Jianing Xu, Bei Zhou, Nanlin Jin","doi":"10.1145/3609703.3609722","DOIUrl":"https://doi.org/10.1145/3609703.3609722","url":null,"abstract":"Reinforcement learning algorithms have been used to discover the strategies in game theory. This study investigates whether Q learning, one of the classic reinforcement learning methods, is capable of training bargaining players via self-play, a training paradigm used by AlphaGo, to maximum their profit. We also compare our empirical results with the known theoretic solutions and perform an comprehensive analysis upon their differences. To accomplish these, we come up with two policy updating methods used in the training process, namely alternate update and simultaneous update, which are tailored for two players who propose offers and counter-offers in an alternating manner under a time constraint enforced by the discount factors. Our experimental results have demonstrated that the values of the discount factor actually have tangible impact on how far the bargaining outcomes deviate from the game theoretic solutions.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121897271","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}
The multi-area economic dispatch (MAED) is a hot and vital research topic for energy saving and emission reduction. Multi-areal economic dispatch refers to the most economical distribution of load requirement among the output units under the premise of satisfying the physical and operational constraints of multiple areas. Each area is connected by a transmission line. In this paper, an improved algorithm (SA-TLBO), which uses adaptive teaching factor to replace the teaching factor in the original teaching-learning based optimization, is developed. Since, adaptive teaching factor can achieve a good balance between convergence speed and search ability, thus improving the overall performance of the algorithm. The method is tested on a system with ten areas, and each area has a 130-unit system. Compared with other two improved strategies and conventional algorithms, the proposed SA-TLBO is shown to yield better solutions for multi-area economic dispatch problems.
{"title":"An Improved Self-Adaptive Teaching-learning Based Optimization for Multi-area Economic Dispatch","authors":"Qun Niu, Gui Xu, L. Tang","doi":"10.1145/3609703.3609716","DOIUrl":"https://doi.org/10.1145/3609703.3609716","url":null,"abstract":"The multi-area economic dispatch (MAED) is a hot and vital research topic for energy saving and emission reduction. Multi-areal economic dispatch refers to the most economical distribution of load requirement among the output units under the premise of satisfying the physical and operational constraints of multiple areas. Each area is connected by a transmission line. In this paper, an improved algorithm (SA-TLBO), which uses adaptive teaching factor to replace the teaching factor in the original teaching-learning based optimization, is developed. Since, adaptive teaching factor can achieve a good balance between convergence speed and search ability, thus improving the overall performance of the algorithm. The method is tested on a system with ten areas, and each area has a 130-unit system. Compared with other two improved strategies and conventional algorithms, the proposed SA-TLBO is shown to yield better solutions for multi-area economic dispatch problems.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130817602","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}
Traditional art exhibitions are usually dominated by relatively static displays such as text, pictures and common multimedia technology. Subject to technical limitations, the exhibition means are relatively simple and the content is relatively thin, which cannot fully meet the exhibition needs of the organizers, nor can it mobilize the enthusiasm of the visitors, and fails to fully show the communication of the exhibition. Therefore, an object detection model based on You Only Look Once(YOLO) network is proposed in this paper to recognize and track the spheres made by felt process. First, the YOLO network was pre-trained using the open source data set, and then the pre-training model was fine-tuned according to the felt sphere image training set. Before fine tuning, the k-means clustering algorithm was used to cluster the marking information of the sphere training set made by felt process. Secondly, for the display of the effect after recognition, OpenCV image processing is used for image special effect processing of the specific recognition area. Through the experimental results, the object detection based on YOLO network proposed in this paper can reach 80.95% in detection accuracy mAP@0.5:0.95 and detection speed up to 20ms, showing excellent performance in detection accuracy and detection speed. It can fit the background interactive display effect of felt art well.
传统的艺术展览通常以文字、图片和常见的多媒体技术等相对静态的展示为主。受技术限制，参展手段比较单一，内容比较单薄，既不能充分满足主办方的参展需求，也不能调动观众的积极性，未能充分表现出展会的交流性。为此，本文提出了一种基于YOLO (You Only Look Once)网络的目标检测模型，用于对毛毡加工的球体进行识别和跟踪。首先利用开源数据集对YOLO网络进行预训练，然后根据毛毡球图像训练集对预训练模型进行微调。在微调之前，使用k-means聚类算法对毡制球训练集的标记信息进行聚类。其次，对于识别后效果的显示，采用OpenCV图像处理对特定识别区域进行图像特效处理。通过实验结果，本文提出的基于YOLO网络的目标检测，检测精度可达80.95% mAP@0.5:0.95，检测速度可达20ms，在检测精度和检测速度上均表现出优异的性能。它能很好地贴合毛毡艺术的背景交互展示效果。
{"title":"Three-Dimensional Sphere Recognition and Tracking Based on YOLO","authors":"Luying Li, Wenjun Huang","doi":"10.1145/3609703.3609706","DOIUrl":"https://doi.org/10.1145/3609703.3609706","url":null,"abstract":"Traditional art exhibitions are usually dominated by relatively static displays such as text, pictures and common multimedia technology. Subject to technical limitations, the exhibition means are relatively simple and the content is relatively thin, which cannot fully meet the exhibition needs of the organizers, nor can it mobilize the enthusiasm of the visitors, and fails to fully show the communication of the exhibition. Therefore, an object detection model based on You Only Look Once(YOLO) network is proposed in this paper to recognize and track the spheres made by felt process. First, the YOLO network was pre-trained using the open source data set, and then the pre-training model was fine-tuned according to the felt sphere image training set. Before fine tuning, the k-means clustering algorithm was used to cluster the marking information of the sphere training set made by felt process. Secondly, for the display of the effect after recognition, OpenCV image processing is used for image special effect processing of the specific recognition area. Through the experimental results, the object detection based on YOLO network proposed in this paper can reach 80.95% in detection accuracy mAP@0.5:0.95 and detection speed up to 20ms, showing excellent performance in detection accuracy and detection speed. It can fit the background interactive display effect of felt art well.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"22 6S 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122810921","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}
Transformer models have shown remarkable effectiveness in capturing long-range dependencies and extracting features for single image super-resolution. However, their deployment on edge devices is hindered by their high computational complexity. To address this challenge, we propose Inception Swin Transformer (IST), a novel model that leverages frequency domain separation to reduce redundant computations.In IST, we exploit the strengths of both CNN-based networks and Transformer variants to handle high-frequency and low-frequency features, respectively. By dynamically utilizing frequency factors to separate feature maps, IST ensures that different components are processed appropriately. Additionally, IST maintains a balanced trade-off between model speed and performance by gradually reducing the proportion of high-frequency components.Our experiments demonstrate that IST effectively reduces the FLOPs while preserving high performance. The combination of Transformers’ accuracy and CNN variants’ efficiency enables IST to significantly reduce computational strain without compromising quality. Comparative analysis reveals that IST outperforms other models, achieving superior results with less FLOPs.
{"title":"Frequency-Split Inception Transformer for Image Super-Resolution","authors":"Wei Xu","doi":"10.1145/3609703.3609708","DOIUrl":"https://doi.org/10.1145/3609703.3609708","url":null,"abstract":"Transformer models have shown remarkable effectiveness in capturing long-range dependencies and extracting features for single image super-resolution. However, their deployment on edge devices is hindered by their high computational complexity. To address this challenge, we propose Inception Swin Transformer (IST), a novel model that leverages frequency domain separation to reduce redundant computations.In IST, we exploit the strengths of both CNN-based networks and Transformer variants to handle high-frequency and low-frequency features, respectively. By dynamically utilizing frequency factors to separate feature maps, IST ensures that different components are processed appropriately. Additionally, IST maintains a balanced trade-off between model speed and performance by gradually reducing the proportion of high-frequency components.Our experiments demonstrate that IST effectively reduces the FLOPs while preserving high performance. The combination of Transformers’ accuracy and CNN variants’ efficiency enables IST to significantly reduce computational strain without compromising quality. Comparative analysis reveals that IST outperforms other models, achieving superior results with less FLOPs.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124346982","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 continuous development of social economy, infrastructure such as highways and bridges continues to develop, and overloaded vehicles are repeatedly prohibited. Therefore, many places have set up "height limit gantry frames" at the entrances of bridges and overpasses, artificially limiting the types and sizes of vehicles, increasing the safety of bridge structures and roads, but also bringing many safety hazards. In recent years, accidents of ultra-high vehicles colliding with "height limit gantry" have occurred frequently, causing serious personnel and property losses. Therefore, it is particularly important to identify and prompt super high vehicles in advance to avoid such accidents. To address such issues, a height limit information early warning system based on V2X multi-sensor information fusion has been developed. Using LiDAR to identify the height of vehicles, using cameras to identify vehicle license plates, and then conducting information fusion, combined with the above information, warning dangerous vehicles through various channels such as V2X or roadside LED screens.
{"title":"An early warning system for height limit based on multi-sensor information fusion","authors":"Chengjun Feng, Chao Wu, Yifan Wu, Dongyi He, Peng Zhou","doi":"10.1145/3609703.3609718","DOIUrl":"https://doi.org/10.1145/3609703.3609718","url":null,"abstract":"With the continuous development of social economy, infrastructure such as highways and bridges continues to develop, and overloaded vehicles are repeatedly prohibited. Therefore, many places have set up \"height limit gantry frames\" at the entrances of bridges and overpasses, artificially limiting the types and sizes of vehicles, increasing the safety of bridge structures and roads, but also bringing many safety hazards. In recent years, accidents of ultra-high vehicles colliding with \"height limit gantry\" have occurred frequently, causing serious personnel and property losses. Therefore, it is particularly important to identify and prompt super high vehicles in advance to avoid such accidents. To address such issues, a height limit information early warning system based on V2X multi-sensor information fusion has been developed. Using LiDAR to identify the height of vehicles, using cameras to identify vehicle license plates, and then conducting information fusion, combined with the above information, warning dangerous vehicles through various channels such as V2X or roadside LED screens.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126464104","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}
At present, data center security and energy consumption have been continuously concerned and discussed. There are some new technology to reduce energy consumption of data center, but few studies focus on make full use of the resources of the existing data center during routine maintenance to solve urgent problems such as low interest rate of distributed power supply resources, mismatch of power supply resources and space resources in cabinet, and unequal distribution of AC power supply systems. This paper focus on problems existing in data center, design a general distribution scheme to recommend network equipment optimal cabinet and location based on entropy weight method ,which combined with entropy weight method to optimize important attributes such as cabinet power utilization, cabinet space utilization, cabinet resource imbalance, and three-phase imbalance, recommend network equipment optimal cabinet and location totally depending on objective data, avoid the problem maintenance staff lack of experience or inconsiderate to make wrong decision effectively, achieve full utilization of data center power system resources and operation optimization. The scheme is verified to be effective, has certain guiding significance for network equipment location selected management in data center.
{"title":"Network equipment recommended placement based on entropy weight method and improved ideal point method distribution scheme design","authors":"Size Liu, Zhenxing Qi, Xinpei Liu, Fangke Lu","doi":"10.1145/3609703.3609720","DOIUrl":"https://doi.org/10.1145/3609703.3609720","url":null,"abstract":"At present, data center security and energy consumption have been continuously concerned and discussed. There are some new technology to reduce energy consumption of data center, but few studies focus on make full use of the resources of the existing data center during routine maintenance to solve urgent problems such as low interest rate of distributed power supply resources, mismatch of power supply resources and space resources in cabinet, and unequal distribution of AC power supply systems. This paper focus on problems existing in data center, design a general distribution scheme to recommend network equipment optimal cabinet and location based on entropy weight method ,which combined with entropy weight method to optimize important attributes such as cabinet power utilization, cabinet space utilization, cabinet resource imbalance, and three-phase imbalance, recommend network equipment optimal cabinet and location totally depending on objective data, avoid the problem maintenance staff lack of experience or inconsiderate to make wrong decision effectively, achieve full utilization of data center power system resources and operation optimization. The scheme is verified to be effective, has certain guiding significance for network equipment location selected management in data center.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133646489","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}
Maddineni Bhargava, K. Vijayan, Oshin Anand, Gaurav Raina
The use of multilingual models for natural language processing is becoming increasingly popular in industrial and business applications, particularly in multilingual societies. In this study, we investigate the transfer learning capabilities of multilingual language models like mBERT and XLM-R across several Indian languages. We study the performance characteristics of a classifier model with mBERT/XLM-R as the front-end, which is trained only in one language for two tasks: text categorization of news articles and sentiment analysis of product reviews. News articles, on the same event but in different languages, are representative of what may be termed as ‘inherently parallel’ data; i.e. data that exhibits similar content across multiple languages, though not necessarily in parallel sentences. Other examples of such data would be customer inquiries/reviews about the same product, social media activity pertaining to the same topic, etcetera. After training in one language, we study the performance characteristics of this classifier model when applied to other languages. Our experiments reveal that by exploiting the inherently parallel nature of the data, XLM-R performs remarkably well when adapted for any Indian language dataset. Further, our study reveals the importance of simultaneously fine-tuning multilingual models with in-domain data from one language in order to express their cross-lingual and domain transfer learning abilities together.
{"title":"Exploration of transfer learning capability of multilingual models for text classification","authors":"Maddineni Bhargava, K. Vijayan, Oshin Anand, Gaurav Raina","doi":"10.1145/3609703.3609711","DOIUrl":"https://doi.org/10.1145/3609703.3609711","url":null,"abstract":"The use of multilingual models for natural language processing is becoming increasingly popular in industrial and business applications, particularly in multilingual societies. In this study, we investigate the transfer learning capabilities of multilingual language models like mBERT and XLM-R across several Indian languages. We study the performance characteristics of a classifier model with mBERT/XLM-R as the front-end, which is trained only in one language for two tasks: text categorization of news articles and sentiment analysis of product reviews. News articles, on the same event but in different languages, are representative of what may be termed as ‘inherently parallel’ data; i.e. data that exhibits similar content across multiple languages, though not necessarily in parallel sentences. Other examples of such data would be customer inquiries/reviews about the same product, social media activity pertaining to the same topic, etcetera. After training in one language, we study the performance characteristics of this classifier model when applied to other languages. Our experiments reveal that by exploiting the inherently parallel nature of the data, XLM-R performs remarkably well when adapted for any Indian language dataset. Further, our study reveals the importance of simultaneously fine-tuning multilingual models with in-domain data from one language in order to express their cross-lingual and domain transfer learning abilities together.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114725072","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}
Lane detection has been one of the most important functions in the autonomous driving perception module. Most of the current research require complex post-processing and curve fitting processes before they can be used by subsequent regulation modules. In this paper, we propose the LLFormer algorithm combining CNN and Transformer structure, which is the first attempt to perform end-to-end lane detection based on laser point cloud and output its cubic polynomial coefficients. In addition, this paper modifies the structure of the conventional transformer and proposes the Generating Lane Query (GLQ) module. The output of encoder is plugged into GLQ for initialization of lane query in decoder, preserving the uniqueness of each frame of point cloud data. We test the performance of the proposed algorithm in the public dataset K-Lane, and the results show that the accuracy of the proposed LLFormer is close to the existing SOTA algorithm. The number of model parameters of LLFormer is only 9.01M, and the amount of operations is only 0.19GFLOPs, which are 1/26 and 1/2937 of the existing SOTA algorithm, respectively. The frequency of inference calculation is 35.9FPS, which can fully meet the real-time requirements for industrial deployment.
{"title":"LLFormer: An Efficient and Real-time LiDAR Lane Detection Method based on Transformer","authors":"Haoxiang Jie, Xinyi Zuo, Jian Gao, W. Liu, Jun-Dong Hu, Shuai Cheng","doi":"10.1145/3609703.3609707","DOIUrl":"https://doi.org/10.1145/3609703.3609707","url":null,"abstract":"Lane detection has been one of the most important functions in the autonomous driving perception module. Most of the current research require complex post-processing and curve fitting processes before they can be used by subsequent regulation modules. In this paper, we propose the LLFormer algorithm combining CNN and Transformer structure, which is the first attempt to perform end-to-end lane detection based on laser point cloud and output its cubic polynomial coefficients. In addition, this paper modifies the structure of the conventional transformer and proposes the Generating Lane Query (GLQ) module. The output of encoder is plugged into GLQ for initialization of lane query in decoder, preserving the uniqueness of each frame of point cloud data. We test the performance of the proposed algorithm in the public dataset K-Lane, and the results show that the accuracy of the proposed LLFormer is close to the existing SOTA algorithm. The number of model parameters of LLFormer is only 9.01M, and the amount of operations is only 0.19GFLOPs, which are 1/26 and 1/2937 of the existing SOTA algorithm, respectively. The frequency of inference calculation is 35.9FPS, which can fully meet the real-time requirements for industrial deployment.","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"71 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116041605","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}
{"title":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","authors":"","doi":"10.1145/3609703","DOIUrl":"https://doi.org/10.1145/3609703","url":null,"abstract":"","PeriodicalId":101485,"journal":{"name":"Proceedings of the 2023 5th International Conference on Pattern Recognition and Intelligent Systems","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125308843","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：