This paper presents a new font shape style transfer technique that employs a generative adversarial network (GAN) and skeleton-based input feature maps to modify a target text to match a target font shape while retaining the original text content. Our GAN model is modified from a Shape-Matching GAN which utilizes a StyleNet generator and a PatchGAN discriminator. Rather than using a base-font character images as input to the generator like other existing font transfer models, we utilize the proposed skeleton-based features as input. The experimental results show that our model can produce the unseen characters in the desired font style better than an existing method.
{"title":"Skeleton-based Generative Adversarial Networks for Font Shape Style Transfer: Learning text style from some characters and transferring the style to any unseen characters","authors":"Thanaphon Thanusan, K. Patanukhom","doi":"10.1145/3596286.3596288","DOIUrl":"https://doi.org/10.1145/3596286.3596288","url":null,"abstract":"This paper presents a new font shape style transfer technique that employs a generative adversarial network (GAN) and skeleton-based input feature maps to modify a target text to match a target font shape while retaining the original text content. Our GAN model is modified from a Shape-Matching GAN which utilizes a StyleNet generator and a PatchGAN discriminator. Rather than using a base-font character images as input to the generator like other existing font transfer models, we utilize the proposed skeleton-based features as input. The experimental results show that our model can produce the unseen characters in the desired font style better than an existing method.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121346836","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}
abstract. Major social events, e.g., civil unrests, generally impact both social stability and civil life. Therefore, anticipating the occurrence of concerned social events in advance is of great significance to decision makers. By mining previous indicators of the event type of interest from open-source data, we can make inference on whether a particular one of that type will occur sometime in the future. In recent years, this kind of data-driven approaches have been proposed to predict social events. However, there are still some challenges remaining to be addressed: (I) Modeling previous feature for a particular event based on limited and obtainable data source. (II) Mining temporal dependences between complicated information in different periods. (III) Explaining prediction results from a reasonable perspective. To cope with these research issues, we proposed a hierarchical attention-based feature learning framework for interpretable social event prediction. We model the evolution processes prior to the onset of an event of interest using a sequence of temporal event graphs. Then, we employ the GNN (Graph Neural Network) approach for graph mining and the attention mechanism on multi-level data for feature learning. For model explanation, an importance evaluation indicator is proposed to identify influential factors of distinct feature levels leading to the event occurrence from the past. Additionally, we conduct experiments on four real-world datasets to verify the proposed method. The results indicate that it outperforms other baseline models on protest prediction tasks.
{"title":"Hierarchical Attention based Feature Learning for Interpretable Social Event Prediction","authors":"Yinsen Wang, Xin Zhang, Yan Pan, Zexin Fu","doi":"10.1145/3596286.3596298","DOIUrl":"https://doi.org/10.1145/3596286.3596298","url":null,"abstract":"abstract. Major social events, e.g., civil unrests, generally impact both social stability and civil life. Therefore, anticipating the occurrence of concerned social events in advance is of great significance to decision makers. By mining previous indicators of the event type of interest from open-source data, we can make inference on whether a particular one of that type will occur sometime in the future. In recent years, this kind of data-driven approaches have been proposed to predict social events. However, there are still some challenges remaining to be addressed: (I) Modeling previous feature for a particular event based on limited and obtainable data source. (II) Mining temporal dependences between complicated information in different periods. (III) Explaining prediction results from a reasonable perspective. To cope with these research issues, we proposed a hierarchical attention-based feature learning framework for interpretable social event prediction. We model the evolution processes prior to the onset of an event of interest using a sequence of temporal event graphs. Then, we employ the GNN (Graph Neural Network) approach for graph mining and the attention mechanism on multi-level data for feature learning. For model explanation, an importance evaluation indicator is proposed to identify influential factors of distinct feature levels leading to the event occurrence from the past. Additionally, we conduct experiments on four real-world datasets to verify the proposed method. The results indicate that it outperforms other baseline models on protest prediction tasks.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127751878","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 exponential growth of online customer reviews has created challenges for potential buyers to filter and identify helpful reviews, directly affecting their shopping experience. Accurate prediction of review helpfulness can improve the selection and presentation of valuable reviews, leading to a better user experience and more informed purchasing decisions. To address the limitations of traditional machine learning methods that rely on handcrafted features and fail to capture semantic context, this paper presents a comparative analysis of existing deep learning models to predict the helpfulness of online reviews. Our study employs larger and more diverse datasets from three popular e-commerce platforms: TripAdvisor, Amazon, and Yelp, and compares multiple deep learning models, including Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and DistilBert, to identify the most accurate and effective predictions. Additionally, the study compares the deep learning models to the traditional machine learning algorithm XGBoost. Understanding the benefits and limitations of each model can lead to improved model selection and optimization, resulting in more accurate and efficient predictions for a wide range of applications. The results show that CNN consistently outperforms the other deep learning models and XGBoost regarding Mean Squared Error (MSE) and training time across all datasets.
{"title":"Comparative Analysis of Deep Learning Models for Predicting Online Review Helpfulness","authors":"Sirinda Palahan","doi":"10.1145/3596286.3596300","DOIUrl":"https://doi.org/10.1145/3596286.3596300","url":null,"abstract":"The exponential growth of online customer reviews has created challenges for potential buyers to filter and identify helpful reviews, directly affecting their shopping experience. Accurate prediction of review helpfulness can improve the selection and presentation of valuable reviews, leading to a better user experience and more informed purchasing decisions. To address the limitations of traditional machine learning methods that rely on handcrafted features and fail to capture semantic context, this paper presents a comparative analysis of existing deep learning models to predict the helpfulness of online reviews. Our study employs larger and more diverse datasets from three popular e-commerce platforms: TripAdvisor, Amazon, and Yelp, and compares multiple deep learning models, including Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and DistilBert, to identify the most accurate and effective predictions. Additionally, the study compares the deep learning models to the traditional machine learning algorithm XGBoost. Understanding the benefits and limitations of each model can lead to improved model selection and optimization, resulting in more accurate and efficient predictions for a wide range of applications. The results show that CNN consistently outperforms the other deep learning models and XGBoost regarding Mean Squared Error (MSE) and training time across all datasets.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132698846","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}
Architectural floor plans play an important role in sharing the building information among engineers, designers, and clients. Automatic floor plan analysis can help in improving work efficiency and accuracy. Object detection and recognition are critical in understanding and analyzing a floor plan document. However, few research works have been conducted to date for automatic object detection in architectural floor plans. In this paper, a convolutional neural network, namely ArchNet, is proposed to detect various visual objects, such as door, window, and stairs. The ArchNet is a modified version of YOLO network, and consists of five modules: backbone, multiscale receptive fields, neck, head, and non-maximal suppression. In this paper, ArchNet is used to detect 13 object classes commonly found in architectural floor plans. Experimental results show that the proposed architecture can achieve a mean average precision of 75% which is superior compared to the state-of-the-art techniques.
{"title":"Deep Object Detection for Complex Architectural Floor Plans with Efficient Receptive Fields","authors":"Zhongguo Xu, N. Jha, Syed Mehadi, M. Mandal","doi":"10.1145/3596286.3596295","DOIUrl":"https://doi.org/10.1145/3596286.3596295","url":null,"abstract":"Architectural floor plans play an important role in sharing the building information among engineers, designers, and clients. Automatic floor plan analysis can help in improving work efficiency and accuracy. Object detection and recognition are critical in understanding and analyzing a floor plan document. However, few research works have been conducted to date for automatic object detection in architectural floor plans. In this paper, a convolutional neural network, namely ArchNet, is proposed to detect various visual objects, such as door, window, and stairs. The ArchNet is a modified version of YOLO network, and consists of five modules: backbone, multiscale receptive fields, neck, head, and non-maximal suppression. In this paper, ArchNet is used to detect 13 object classes commonly found in architectural floor plans. Experimental results show that the proposed architecture can achieve a mean average precision of 75% which is superior compared to the state-of-the-art techniques.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115089413","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, X-ray technology, B-ultrasound and magnetic resonance imaging technology have more or less defects in the detection of female breast cancer, so the early detection of breast cancer is still a very important challenge. Ultrasound tomography (UT) can solve these problems very well. This project mainly uses the TVAL3 algorithm to reconstruct the original image from the information collected by the UT system for clinical use. TVAL3 algorithm involves a large number of matrix-vector multiplications and transposed matrix-vector multiplications, which will consume a lot of time if traditional CPU methods are used. For the characteristics of matrix-vector multiplication, this project uses CUDA to call GPU for parallel computing. At the same time, in order to further increase the speed of the calculation, we put part of the unchanged content into the GPU in advance to reduce the time spent on the transfer process. The final speedups of 20x, 10x and 5x were achieved in matrix vector multiplication, transpose matrix vector multiplication and total time, respectively.
{"title":"3D Ultrasound Tomography Image Reconstruction Algorithm by GPU","authors":"Jiaduo Gong","doi":"10.1145/3596286.3596290","DOIUrl":"https://doi.org/10.1145/3596286.3596290","url":null,"abstract":"At present, X-ray technology, B-ultrasound and magnetic resonance imaging technology have more or less defects in the detection of female breast cancer, so the early detection of breast cancer is still a very important challenge. Ultrasound tomography (UT) can solve these problems very well. This project mainly uses the TVAL3 algorithm to reconstruct the original image from the information collected by the UT system for clinical use. TVAL3 algorithm involves a large number of matrix-vector multiplications and transposed matrix-vector multiplications, which will consume a lot of time if traditional CPU methods are used. For the characteristics of matrix-vector multiplication, this project uses CUDA to call GPU for parallel computing. At the same time, in order to further increase the speed of the calculation, we put part of the unchanged content into the GPU in advance to reduce the time spent on the transfer process. The final speedups of 20x, 10x and 5x were achieved in matrix vector multiplication, transpose matrix vector multiplication and total time, respectively.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126636439","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}
Abstract: With the increasing popularity of photovoltaic power generation, the demand for photovoltaic panel defect detection in the industry is also increasing. Deep learning can automatically extract individual photovoltaic panels from images or videos, and perform the defect detection task on it. Aiming at the problem of low detection accuracy of existing deep learning-based photovoltaic panel defect detection methods, an improved Mask R-CNN photovoltaic panel defect detection algorithm is proposed. To improve the training performance, the feature pyramid (FPN) structure is improved, and the cascade network based on attention guidance is adopted to fuse more features and prevent the loss of shallow semantic information to a certain extent. Secondly, Group Normalization (GN) is used to replace Batch Normalization (BN) in the traditional high-performance deep neural network models. The quality of the self-made dataset is improved by Mosaic data enhancement to prevent accuracy loss due to insufficient sample size in the dataset. The effectiveness of the algorithm is verified by the self-made dataset and the public COCO2017 dataset. The improved Mask R-CNN algorithm has a detection accuracy of more than 89% on the self-made photovoltaic panel dataset and 44.6% bounding box average precision (APbbox) and 41.5% mask average precision (APmask) on the COCO2017 dataset, which is 6.4% and 5.8% higher than the original Mask R-CNN algorithm respectively. Finally, to comprehensively analyze the detection performance of the improved algorithm in photovoltaic panel defect detection tasks, the common deep learning-based defect detection algorithms for photovoltaic panel defect detection are summarized. Based on this, a comparison and summary of the improved algorithm in this paper are conducted.
{"title":"Improved Mask R-CNN Network Method for PV Panel Defect Detection","authors":"Wangwang Yang, Z. Deng, Enwen Hu, Yao Zhang","doi":"10.1145/3596286.3596287","DOIUrl":"https://doi.org/10.1145/3596286.3596287","url":null,"abstract":"Abstract: With the increasing popularity of photovoltaic power generation, the demand for photovoltaic panel defect detection in the industry is also increasing. Deep learning can automatically extract individual photovoltaic panels from images or videos, and perform the defect detection task on it. Aiming at the problem of low detection accuracy of existing deep learning-based photovoltaic panel defect detection methods, an improved Mask R-CNN photovoltaic panel defect detection algorithm is proposed. To improve the training performance, the feature pyramid (FPN) structure is improved, and the cascade network based on attention guidance is adopted to fuse more features and prevent the loss of shallow semantic information to a certain extent. Secondly, Group Normalization (GN) is used to replace Batch Normalization (BN) in the traditional high-performance deep neural network models. The quality of the self-made dataset is improved by Mosaic data enhancement to prevent accuracy loss due to insufficient sample size in the dataset. The effectiveness of the algorithm is verified by the self-made dataset and the public COCO2017 dataset. The improved Mask R-CNN algorithm has a detection accuracy of more than 89% on the self-made photovoltaic panel dataset and 44.6% bounding box average precision (APbbox) and 41.5% mask average precision (APmask) on the COCO2017 dataset, which is 6.4% and 5.8% higher than the original Mask R-CNN algorithm respectively. Finally, to comprehensively analyze the detection performance of the improved algorithm in photovoltaic panel defect detection tasks, the common deep learning-based defect detection algorithms for photovoltaic panel defect detection are summarized. Based on this, a comparison and summary of the improved algorithm in this paper are conducted.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"90 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122535391","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}
Efficient detection of dim-small targets with high accuracy is a difficult task in the field of infrared target tracking since the tiny size of small infrared targets significantly reduces the accuracy of conventional models. To address this issue, this paper improves YOLOv7 so that it can be applied to the detection of infrared dim-small targets. Initially, an enhanced MPConv-based pooling structure is proposed, which reduces the high false detection rate caused by white point noise. Then, a CBAM attention module is added to the backbone structure, which employs both spatial and channel attention to preserve more of the original characteristics of infrared faint targets. Finally, the EIOU loss is utilized in the Head module to increase the speed of model convergence. Experiments reveal that the improved algorithm achieves a model mAP of 70.8% on the dim-small target dataset, which represents a 3.4% improvement over YOLOv7 and outperforms other conventional algorithms.
{"title":"An Infrared Dim-small Target Detection Method Based on Improved YOLOv7","authors":"Yujie Zheng, Yuyong Cui, Xinyi Gao","doi":"10.1145/3596286.3596289","DOIUrl":"https://doi.org/10.1145/3596286.3596289","url":null,"abstract":"Efficient detection of dim-small targets with high accuracy is a difficult task in the field of infrared target tracking since the tiny size of small infrared targets significantly reduces the accuracy of conventional models. To address this issue, this paper improves YOLOv7 so that it can be applied to the detection of infrared dim-small targets. Initially, an enhanced MPConv-based pooling structure is proposed, which reduces the high false detection rate caused by white point noise. Then, a CBAM attention module is added to the backbone structure, which employs both spatial and channel attention to preserve more of the original characteristics of infrared faint targets. Finally, the EIOU loss is utilized in the Head module to increase the speed of model convergence. Experiments reveal that the improved algorithm achieves a model mAP of 70.8% on the dim-small target dataset, which represents a 3.4% improvement over YOLOv7 and outperforms other conventional algorithms.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123439136","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}
Chen De, Yanrui Dong, Zhou Jun Xiong, Wang Hai, Du Yi Xian, Li Shi Peng
For the battery cell production lines using automatic defect detection equipment, it becomes necessary to control the fluctuation of the yield rate of the production line, and adapt to the difference among the cell replacements and the incoming material processes of each batch. So the defect detection algorithm needs to adjust the detection criteria according to the target preferential rate range, which is a new challenge for artificial intelligent manufacturing in the battery industry. Taking the dent defect on the side of the battery cell as an example, based on both the traditional algorithm and the deep learning algorithm, two kinds of self-adaptive adjustment method of the defect detection criteria are proposed. The traditional algorithm employs the linear interpolation method to classify good and bad products based on depth and area information, and calculates the optimal critical values of depth and area that meet the target yield rate. The deep learning algorithm combines the convolutional neural network and the support vector machine classifiers, with the histogram of oriented gradient feature as the classifier input, so as to classify different degrees of defective products. The test results show that the detection criteria can be adjusted flexibly and automatically for the side dent defects, which could realize the self-adaption of algorithm to the target yield rate, save the manual operation time, and improve the production efficiency.
{"title":"Self-adaptive Methods with Flexible Detection Criteria of the Battery Cell Dent Defect","authors":"Chen De, Yanrui Dong, Zhou Jun Xiong, Wang Hai, Du Yi Xian, Li Shi Peng","doi":"10.1145/3596286.3596296","DOIUrl":"https://doi.org/10.1145/3596286.3596296","url":null,"abstract":"For the battery cell production lines using automatic defect detection equipment, it becomes necessary to control the fluctuation of the yield rate of the production line, and adapt to the difference among the cell replacements and the incoming material processes of each batch. So the defect detection algorithm needs to adjust the detection criteria according to the target preferential rate range, which is a new challenge for artificial intelligent manufacturing in the battery industry. Taking the dent defect on the side of the battery cell as an example, based on both the traditional algorithm and the deep learning algorithm, two kinds of self-adaptive adjustment method of the defect detection criteria are proposed. The traditional algorithm employs the linear interpolation method to classify good and bad products based on depth and area information, and calculates the optimal critical values of depth and area that meet the target yield rate. The deep learning algorithm combines the convolutional neural network and the support vector machine classifiers, with the histogram of oriented gradient feature as the classifier input, so as to classify different degrees of defective products. The test results show that the detection criteria can be adjusted flexibly and automatically for the side dent defects, which could realize the self-adaption of algorithm to the target yield rate, save the manual operation time, and improve the production efficiency.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122266356","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}
T. Thumthawatworn, K. Nongpong, Pawut Satitsuksanoh
Over the last couple of years, the recent pandemic rendered work operations to be mobile and relied heavily on real-time and traffic-intensive applications such as online classrooms and meetings. Since our working life requires seamless mobility and stable wireless connectivity, heterogeneous wireless networks gain more attention as key infrastructures to fulfill communication needs. Intelligent handover decision deems necessary to select the appropriate wireless network. An intelligent mechanism such as fuzzy logic proves to enhance such decision-making. Different membership functions used in a fuzzy inference system contribute to different network selection performances. This work evaluates different fuzzy membership functions, hence appropriate membership functions can be used in the design of fuzzy-based network selection mechanisms for heterogeneous wireless networks.
{"title":"Analysis of Fuzzy Network Selection Algorithm for Heterogeneous Wireless Mobile Networks","authors":"T. Thumthawatworn, K. Nongpong, Pawut Satitsuksanoh","doi":"10.1145/3596286.3596299","DOIUrl":"https://doi.org/10.1145/3596286.3596299","url":null,"abstract":"Over the last couple of years, the recent pandemic rendered work operations to be mobile and relied heavily on real-time and traffic-intensive applications such as online classrooms and meetings. Since our working life requires seamless mobility and stable wireless connectivity, heterogeneous wireless networks gain more attention as key infrastructures to fulfill communication needs. Intelligent handover decision deems necessary to select the appropriate wireless network. An intelligent mechanism such as fuzzy logic proves to enhance such decision-making. Different membership functions used in a fuzzy inference system contribute to different network selection performances. This work evaluates different fuzzy membership functions, hence appropriate membership functions can be used in the design of fuzzy-based network selection mechanisms for heterogeneous wireless networks.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"138 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127448335","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}
Infrared thermal imaging is widely used in industrial inspection due to its advantages such as passive identification, non-contact detection, long detection distance and strong environmental adaptability. In power systems, infrared thermal imaging can be used to carry out live detection of power equipment to prevent or examine potential risk and threats. This paper provides a fault detection method for power equipment through the visible light-infrared image fusion technology. The information of infrared image is collected through infrared thermal imager, and the infrared image is preprocessed. The scale invariant feature transform (SIFT) feature point detection algorithm is used to extract the difference between visible light image and infrared image. The feature points are screened and registered by random sample consensus (RANSAC) algorithm to realize the fusion of the visible light image and the infrared image of the power equipment, so as to detect the working status of the power equipment and accurately locate the fault source when a fault occurs.
{"title":"Application of visible light-infrared image fusion technology in power system fault detection","authors":"Sichao Chen, Yang Luo, Jianbo Yin, Guohua Zhou, Dilong Shen, Liang Shen","doi":"10.1145/3596286.3596294","DOIUrl":"https://doi.org/10.1145/3596286.3596294","url":null,"abstract":"Infrared thermal imaging is widely used in industrial inspection due to its advantages such as passive identification, non-contact detection, long detection distance and strong environmental adaptability. In power systems, infrared thermal imaging can be used to carry out live detection of power equipment to prevent or examine potential risk and threats. This paper provides a fault detection method for power equipment through the visible light-infrared image fusion technology. The information of infrared image is collected through infrared thermal imager, and the infrared image is preprocessed. The scale invariant feature transform (SIFT) feature point detection algorithm is used to extract the difference between visible light image and infrared image. The feature points are screened and registered by random sample consensus (RANSAC) algorithm to realize the fusion of the visible light image and the infrared image of the power equipment, so as to detect the working status of the power equipment and accurately locate the fault source when a fault occurs.","PeriodicalId":208318,"journal":{"name":"Proceedings of the 2023 Asia Conference on Computer Vision, Image Processing and Pattern Recognition","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133685103","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：