Pub Date : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319494
Ty V. Nguyen, Incheon Paik
This paper addresses the plant disease detection and classification using Deep Learning approach. In particular, we propose a novel model using the Triplet Loss together with the fine-tuned pre-trained MobileNet model to extract good features, classify, and detect diseases of plants from the open-source PlantVillage dataset. Using our proposed model, the achievable results are 99.92%, which outperforms the existing models using the same dataset. Furthermore, our proposed model can support the large-scale agricultural sector, which plays an important role in ensuring food security during the current COVID-19 crisis.
{"title":"Feature Extraction with Triplet Loss to Classify Disease on Leaf Data","authors":"Ty V. Nguyen, Incheon Paik","doi":"10.1109/iCAST51195.2020.9319494","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319494","url":null,"abstract":"This paper addresses the plant disease detection and classification using Deep Learning approach. In particular, we propose a novel model using the Triplet Loss together with the fine-tuned pre-trained MobileNet model to extract good features, classify, and detect diseases of plants from the open-source PlantVillage dataset. Using our proposed model, the achievable results are 99.92%, which outperforms the existing models using the same dataset. Furthermore, our proposed model can support the large-scale agricultural sector, which plays an important role in ensuring food security during the current COVID-19 crisis.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131393882","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319482
Praveen K. Parashiva, A. P. Vinod
Error-Related Potential (Errp) is the bioelectric potential elicited in human brain as a result of the cognitive state of awareness when an error is perceived. Identifying ErrP from a single trial electroencephalogram (EEG) data can be used in taking corrective actions to fix the error or as a learning strategy in Brain Computer Interface (BCI) systems. The ErrP dataset recorded using EEG will contain both erroneous and correct actions. A classifier such as the Artificial Neural Network (ANN) can be trained to identify the erroneous versus correct action from a single trial EEG data. However, the classifier will have large number of parameters to be learned, and typically, the ErrP dataset is unbalanced with smaller number of erroneous trials. Therefore, the trained classifier may not generalize the data well. To classify the ErrP with better accuracy, an ANN architecture is proposed in this work. Learning the parameters of the ANN is carried out in two stages (Stage-1 and Stage-2) in the proposed method. The first stage of learning will have relatively large feature samples collected from several subjects. The first stage learning is aimed to capture the global characteristics of the ErrP. In the second stage, the pre-trained ANN classifier from the first stage is tuned for each subject. The ErrP dataset has two sessions dataset recorded from six subjects and the Stage-1 and Stage-2 training models are cross-validated. The overall classification accuracy achieved after cross-validation is 74.78 ± 3.43% and 86.03 ± 1.02% for erroneous and correct trials respectively. The improvement in the classification accuracy achieved is 12.67% and 15.51% for erroneous and correct trials respectively compared with the existing statistical classifier method. The method proposed shows an efficient way to train ANN classifier to achieve higher classification accuracy for unbalanced and smaller dataset such as ErrP.
错误相关电位(error - related Potential, Errp)是人类在感知到错误时，由于认知意识状态而在大脑中引发的生物电电位。从单个试验脑电图(EEG)数据中识别ErrP可用于采取纠正措施以修复错误或作为脑机接口(BCI)系统中的学习策略。使用EEG记录的ErrP数据集将包含错误和正确的操作。像人工神经网络(ANN)这样的分类器可以被训练来从单个试验脑电图数据中识别错误和正确的动作。然而，分类器将有大量的参数需要学习，通常，ErrP数据集是不平衡的，错误试验的数量较少。因此，训练好的分类器可能不能很好地泛化数据。为了更好地对ErrP进行分类，本文提出了一种人工神经网络体系结构。在本文提出的方法中，人工神经网络的参数学习分为两个阶段(阶段1和阶段2)进行。学习的第一阶段将从几个科目中收集相对较大的特征样本。第一阶段学习的目的是捕捉ErrP的全局特征。在第二阶段，针对每个主题对第一阶段预训练的ANN分类器进行调整。ErrP数据集有两个会话数据集，记录了来自六个受试者的数据集，并且交叉验证了第一阶段和第二阶段的训练模型。交叉验证后，错误试验和正确试验的总体分类准确率分别为74.78±3.43%和86.03±1.02%。与现有的统计分类器方法相比，该方法的误试和正确率分别提高了12.67%和15.51%。本文提出的方法是一种有效的训练ANN分类器的方法，可以在ErrP等不平衡和较小的数据集上实现更高的分类精度。
{"title":"Improving Classification Accuracy of Detecting Error-Related Potentials using Two-stage Trained Neural Network Classifier","authors":"Praveen K. Parashiva, A. P. Vinod","doi":"10.1109/iCAST51195.2020.9319482","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319482","url":null,"abstract":"Error-Related Potential (Errp) is the bioelectric potential elicited in human brain as a result of the cognitive state of awareness when an error is perceived. Identifying ErrP from a single trial electroencephalogram (EEG) data can be used in taking corrective actions to fix the error or as a learning strategy in Brain Computer Interface (BCI) systems. The ErrP dataset recorded using EEG will contain both erroneous and correct actions. A classifier such as the Artificial Neural Network (ANN) can be trained to identify the erroneous versus correct action from a single trial EEG data. However, the classifier will have large number of parameters to be learned, and typically, the ErrP dataset is unbalanced with smaller number of erroneous trials. Therefore, the trained classifier may not generalize the data well. To classify the ErrP with better accuracy, an ANN architecture is proposed in this work. Learning the parameters of the ANN is carried out in two stages (Stage-1 and Stage-2) in the proposed method. The first stage of learning will have relatively large feature samples collected from several subjects. The first stage learning is aimed to capture the global characteristics of the ErrP. In the second stage, the pre-trained ANN classifier from the first stage is tuned for each subject. The ErrP dataset has two sessions dataset recorded from six subjects and the Stage-1 and Stage-2 training models are cross-validated. The overall classification accuracy achieved after cross-validation is 74.78 ± 3.43% and 86.03 ± 1.02% for erroneous and correct trials respectively. The improvement in the classification accuracy achieved is 12.67% and 15.51% for erroneous and correct trials respectively compared with the existing statistical classifier method. The method proposed shows an efficient way to train ANN classifier to achieve higher classification accuracy for unbalanced and smaller dataset such as ErrP.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130768454","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319472
Weilun Wang, G. Chakraborty, B. Chakraborty
With the recent development of algorithm for computer-aided diagnosis (CAD) system, detection of pulmonary nodules from computed tomography (CT) imaging data with high accuracy is becoming possible. Existing CAD system is able to automatically output the location of a nodule with its confidence. It helps the radiologist to save time for nodule detection work. However, not all nodules will develop into lung cancer. Depending on its grade of malignancy, the probability of developing into lung cancer is different. Evaluating the grade of malignancy of pulmonary nodule is performed by doctors and highly depends on personal experience. In order to further automate the process of lung cancer prognosis, a system that accurately evaluates the grade of malignancy of a pulmonary nodule is needed. It will be helpful to re-evaluate the detected nodules and provide proper suggestion for therapeutic method. There are two types of tasks for malignancy classification (1) to classify a sample into benign or malignant (2) to classify a sample into malignancy grades (from grade-1 to grade-5). Many researches have achieved a high accuracy for task-1, but the results on task-2 are still poor. In this work, we present a 3D Multi-scale DenseNet to classify the grade of malignancy of pulmonary nodules. Through the observation of CT image data, we found that for some small nodules it is impossible to extract their morphological features due to their small size. Our idea is to convert the original CT image into three different scales (Multi-scale) and input them into three parallel 3D densely-connected convolutional network (DenseN et) blocks. Finally, the extracted features from the last layer of the three networks are concatenated to classify the grade of malignancy. In this way, the morphological features of small nodules can be better obtained without affecting the feature extraction of large nodules. In this study, 1882 samples from the dataset of Lung Image Database Consortium (LID C) are used for training and testing. Overall, we achieved 68.5 % accuracy for the task of malignancy grades classification.
{"title":"3D Multi-scale DenseNet for Malignancy Grade Classification of Pulmonary Nodules","authors":"Weilun Wang, G. Chakraborty, B. Chakraborty","doi":"10.1109/iCAST51195.2020.9319472","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319472","url":null,"abstract":"With the recent development of algorithm for computer-aided diagnosis (CAD) system, detection of pulmonary nodules from computed tomography (CT) imaging data with high accuracy is becoming possible. Existing CAD system is able to automatically output the location of a nodule with its confidence. It helps the radiologist to save time for nodule detection work. However, not all nodules will develop into lung cancer. Depending on its grade of malignancy, the probability of developing into lung cancer is different. Evaluating the grade of malignancy of pulmonary nodule is performed by doctors and highly depends on personal experience. In order to further automate the process of lung cancer prognosis, a system that accurately evaluates the grade of malignancy of a pulmonary nodule is needed. It will be helpful to re-evaluate the detected nodules and provide proper suggestion for therapeutic method. There are two types of tasks for malignancy classification (1) to classify a sample into benign or malignant (2) to classify a sample into malignancy grades (from grade-1 to grade-5). Many researches have achieved a high accuracy for task-1, but the results on task-2 are still poor. In this work, we present a 3D Multi-scale DenseNet to classify the grade of malignancy of pulmonary nodules. Through the observation of CT image data, we found that for some small nodules it is impossible to extract their morphological features due to their small size. Our idea is to convert the original CT image into three different scales (Multi-scale) and input them into three parallel 3D densely-connected convolutional network (DenseN et) blocks. Finally, the extracted features from the last layer of the three networks are concatenated to classify the grade of malignancy. In this way, the morphological features of small nodules can be better obtained without affecting the feature extraction of large nodules. In this study, 1882 samples from the dataset of Lung Image Database Consortium (LID C) are used for training and testing. Overall, we achieved 68.5 % accuracy for the task of malignancy grades classification.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"11 9","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120936402","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319468
Qianmin Chen, Eric Rigall, Xianglong Wang, H. Fan, Junyu Dong
We present a neural network to detect playing cards in real poker scenes through a camera, where the playing card area represents only 0.7% of the shot table area. In the acquired images, the suits of cards are fuzzy and difficult to identify, even to the naked eye. Because of the relatively few pixels corresponding to the cards, traditional image processing and pattern recognition methods struggle to detect them. Therefore, we use deep learning methods to detect, which have shown to be easy-to-use, faster and more accurate in a broad range of computer vision applications over the years. Inspired by the sandglass block, we improved the current state-of-the-art neural network architecture for object detection, EfficientDet, to retain more features. Experiments have been conducted to evaluate the performance of our improved EfficientDet model and showed that it achieved considerable performance improvement compared with the other deep learning models.
{"title":"Poker Watcher: Playing Card Detection Based on EfficientDet and Sandglass Block","authors":"Qianmin Chen, Eric Rigall, Xianglong Wang, H. Fan, Junyu Dong","doi":"10.1109/iCAST51195.2020.9319468","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319468","url":null,"abstract":"We present a neural network to detect playing cards in real poker scenes through a camera, where the playing card area represents only 0.7% of the shot table area. In the acquired images, the suits of cards are fuzzy and difficult to identify, even to the naked eye. Because of the relatively few pixels corresponding to the cards, traditional image processing and pattern recognition methods struggle to detect them. Therefore, we use deep learning methods to detect, which have shown to be easy-to-use, faster and more accurate in a broad range of computer vision applications over the years. Inspired by the sandglass block, we improved the current state-of-the-art neural network architecture for object detection, EfficientDet, to retain more features. Experiments have been conducted to evaluate the performance of our improved EfficientDet model and showed that it achieved considerable performance improvement compared with the other deep learning models.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121723512","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319484
Huitao Wang, Kai Su, I. M. Chowdhury, Qiangfu Zhao, Yoichi Tomioka
On-road risk detection and alert system is a crucial and important task in our day to day life. Deep Learning approaches have got much attention in solving this noble task. In this paper, we have performed a comparative study on two recent architectures that handle the on-road risk detection task, which are Block-Wise Detection and Modular Selective Network (MS-Net). In the Block-Wise Detection, we have used the VGG19, VGG19-BN, and ResNet family as the backbone network. On the other hand, for MS-Net we have used the ResNet-44 as the router and ResNet-101 as the expert network. In this experiment, we evaluate our model on an “on-road risk detection dataset”, which was created by our research group using an RGB-D sensor mounted on a senior car. On this dataset, we can achieve an accuracy of 89.40% for MS-Net. For the Block-Wise Detection model, we can achieve an accuracy of 90.51% if we use ResNet-50 as the backbone network. However, if we choose the network models used in MS-Net, we can double the inference speed. Thus, compared with Block-Wise Detection, we think the overall performance of MS-NET is better, and is potentially more useful for driving assistance of elderly drivers.
{"title":"Comparison Between Block-Wise Detection and A Modular Selective Approach","authors":"Huitao Wang, Kai Su, I. M. Chowdhury, Qiangfu Zhao, Yoichi Tomioka","doi":"10.1109/iCAST51195.2020.9319484","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319484","url":null,"abstract":"On-road risk detection and alert system is a crucial and important task in our day to day life. Deep Learning approaches have got much attention in solving this noble task. In this paper, we have performed a comparative study on two recent architectures that handle the on-road risk detection task, which are Block-Wise Detection and Modular Selective Network (MS-Net). In the Block-Wise Detection, we have used the VGG19, VGG19-BN, and ResNet family as the backbone network. On the other hand, for MS-Net we have used the ResNet-44 as the router and ResNet-101 as the expert network. In this experiment, we evaluate our model on an “on-road risk detection dataset”, which was created by our research group using an RGB-D sensor mounted on a senior car. On this dataset, we can achieve an accuracy of 89.40% for MS-Net. For the Block-Wise Detection model, we can achieve an accuracy of 90.51% if we use ResNet-50 as the backbone network. However, if we choose the network models used in MS-Net, we can double the inference speed. Thus, compared with Block-Wise Detection, we think the overall performance of MS-NET is better, and is potentially more useful for driving assistance of elderly drivers.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122258560","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319469
Kai Su, Huitao Wang, I. M. Chowdhury, Qiangfu Zhao, Yoichi Tomioka
In this paper, we study real-time object detection based on cell-wise segmentation. Existing object detection methods usually focus on detecting interesting object's positions and sizes and demand expensive computing resources. This process makes it difficult to achieve high-speed and high-precision detection with low-cost devices. We propose a method called You Only Look at Interested Cells or in-short YOLIC to solve the problem by focusing on predefined interested cells (i.e., subregions) in an image. A key challenge here is how to predict the object types contained in all interested cells efficiently, all at once. Instead of using multiple predictors for all interested cells, we use only one deep learner to classify all interested cells. In other words, YOLIC applies the concept of multi-label classification for object detection. YOLIC can use exiting classification models without any structural change. The main point is to define a proper loss function for training. Using on-road risk detection as a test case, we confirmed that YOLIC is significantly faster and accurate than YOLO-v3 in terms of FPS and F1-score.
本文研究了基于单元分割的实时目标检测。现有的目标检测方法通常集中于检测感兴趣的目标的位置和大小，需要耗费昂贵的计算资源。这一过程使得用低成本的设备实现高速、高精度的检测变得困难。我们提出了一种名为You Only Look at Interested Cells(简称YOLIC)的方法，通过关注图像中预定义的感兴趣的细胞(即子区域)来解决这个问题。这里的一个关键挑战是如何一次有效地预测所有感兴趣的单元格中包含的对象类型。我们只使用一个深度学习器对所有感兴趣的细胞进行分类，而不是对所有感兴趣的细胞使用多个预测器。换句话说，YOLIC将多标签分类的概念应用于目标检测。YOLIC可以在不改变结构的情况下使用现有的分类模型。重点是定义一个合适的训练损失函数。以道路风险检测为例，我们证实YOLIC在FPS和f1分数方面明显比YOLO-v3更快、更准确。
{"title":"You Only Look at Interested Cells: Real-Time Object Detection Based on Cell-Wise Segmentation","authors":"Kai Su, Huitao Wang, I. M. Chowdhury, Qiangfu Zhao, Yoichi Tomioka","doi":"10.1109/iCAST51195.2020.9319469","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319469","url":null,"abstract":"In this paper, we study real-time object detection based on cell-wise segmentation. Existing object detection methods usually focus on detecting interesting object's positions and sizes and demand expensive computing resources. This process makes it difficult to achieve high-speed and high-precision detection with low-cost devices. We propose a method called You Only Look at Interested Cells or in-short YOLIC to solve the problem by focusing on predefined interested cells (i.e., subregions) in an image. A key challenge here is how to predict the object types contained in all interested cells efficiently, all at once. Instead of using multiple predictors for all interested cells, we use only one deep learner to classify all interested cells. In other words, YOLIC applies the concept of multi-label classification for object detection. YOLIC can use exiting classification models without any structural change. The main point is to define a proper loss function for training. Using on-road risk detection as a test case, we confirmed that YOLIC is significantly faster and accurate than YOLO-v3 in terms of FPS and F1-score.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128076651","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319407
Tsukasa Ueno, Qiangfu Zhao, Shota Nakada
Many methods for product defect detection have been proposed in the literature. The methods can be roughly divided into two categories, namely conventional statistical methods and machine learning-based ones. Especially for image-based defect detection, deep learning is known as the state-of-the-art. For product defect detection, the main issue is to reduce the false negative error rate (FNER) to almost zero, while keeping a relatively low false positive error rate (FPER). We can reduce the errors by introducing a rejection mechanism, but this approach may reject too many products for manual re-checking. In this study, we found that extremely low FNER can be achieved if we combine several techniques in using deep learning. In this paper, we introduce the techniques briefly, and provide experimental results to show how these techniques affect the performance for defect detection.
{"title":"Deep Learning-Based Industry Product Defect Detection with Low False Negative Error Tolerance","authors":"Tsukasa Ueno, Qiangfu Zhao, Shota Nakada","doi":"10.1109/iCAST51195.2020.9319407","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319407","url":null,"abstract":"Many methods for product defect detection have been proposed in the literature. The methods can be roughly divided into two categories, namely conventional statistical methods and machine learning-based ones. Especially for image-based defect detection, deep learning is known as the state-of-the-art. For product defect detection, the main issue is to reduce the false negative error rate (FNER) to almost zero, while keeping a relatively low false positive error rate (FPER). We can reduce the errors by introducing a rejection mechanism, but this approach may reject too many products for manual re-checking. In this study, we found that extremely low FNER can be achieved if we combine several techniques in using deep learning. In this paper, we introduce the techniques briefly, and provide experimental results to show how these techniques affect the performance for defect detection.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133080780","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}
As social media has become more widely used, fake news has become an increasingly serious problem. The representative countermeasures against fake news are fake news detection and automated fact-checking. However, these countermeasures are not sufficient because people using social media tend to ignore facts that contradict their current beliefs. Therefore, developing effective countermeasures requires understanding the nature of fake news dissemination. Previous models related to this aim have been proposed for describing and analyzing opinion dissemination among people. However, these models are not adequate because they are based on the assumptions that ignore the presence of fake. That is, they assume that people believe their friends equally without doubting and that reliability among people does not change. In this paper, we propose a model that can better describe the opinion dissemination in the presence of fake news. In our model, each person updates the reliability of and doubt about his or her friends and exchanges opinions among each other. Applying the proposed model to artificial and real-world social networks, we found three clues to analyze the nature of fake news dissemination: 1) people can less accurately perceive that fake news is fake than they can perceive that real news is real. 2) it takes much more time for people to perceive fake news to be fake than to perceive real news to be real. 3) the results of findings 1 and 2 concerning fake news are because people become skeptical about friends in the presence of fake news and therefore people do not update opinions much.
{"title":"A Fake News Dissemination Model Based on Updating Reliability and Doubt among Individuals","authors":"Kento Yoshikawa, Takumi Awa, Risa Kusano, Hiroyuki Sato, Masatsugu Ichino, H. Yoshiura","doi":"10.1109/iCAST51195.2020.9319485","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319485","url":null,"abstract":"As social media has become more widely used, fake news has become an increasingly serious problem. The representative countermeasures against fake news are fake news detection and automated fact-checking. However, these countermeasures are not sufficient because people using social media tend to ignore facts that contradict their current beliefs. Therefore, developing effective countermeasures requires understanding the nature of fake news dissemination. Previous models related to this aim have been proposed for describing and analyzing opinion dissemination among people. However, these models are not adequate because they are based on the assumptions that ignore the presence of fake. That is, they assume that people believe their friends equally without doubting and that reliability among people does not change. In this paper, we propose a model that can better describe the opinion dissemination in the presence of fake news. In our model, each person updates the reliability of and doubt about his or her friends and exchanges opinions among each other. Applying the proposed model to artificial and real-world social networks, we found three clues to analyze the nature of fake news dissemination: 1) people can less accurately perceive that fake news is fake than they can perceive that real news is real. 2) it takes much more time for people to perceive fake news to be fake than to perceive real news to be real. 3) the results of findings 1 and 2 concerning fake news are because people become skeptical about friends in the presence of fake news and therefore people do not update opinions much.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123878786","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319474
Chenxin Jia, Ying Cao, Jian Yang, Y. Rao, H. Fan, Wenlin Yao
In order to obtain more accurate pose estimation, the visual-inertial odometry (VIO) system with outlier rejection and loop closure is proposed in this paper. Considering that feature matching is an important part in the front-end of the VIO system, its accuracy will affect the performance of the entire system. So we introduce an outlier rejection method of the grid-based motion statistics (GMS) algorithm to the VIO system. And for more robust feature correspondence and better camera pose estimation, we propose an improved GMS method to eliminate the mismatched points. Besides, we adopt the loop closure strategy to correct the cumulative error of the VIO system. Finally, we estimate the camera pose, velocity and IMU bias simultaneously by minimizing the loss function which contains reprojection error and IMU error. A large number of experiments on EuRoC demonstrate that the proposed method outperforms the advanced VIO system ROVIO and is comparable to the state-of-the-art VIO system OKVIS.
{"title":"Improving Visual- Inertial Odometry with Robust Outlier Rejection and Loop Closure","authors":"Chenxin Jia, Ying Cao, Jian Yang, Y. Rao, H. Fan, Wenlin Yao","doi":"10.1109/iCAST51195.2020.9319474","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319474","url":null,"abstract":"In order to obtain more accurate pose estimation, the visual-inertial odometry (VIO) system with outlier rejection and loop closure is proposed in this paper. Considering that feature matching is an important part in the front-end of the VIO system, its accuracy will affect the performance of the entire system. So we introduce an outlier rejection method of the grid-based motion statistics (GMS) algorithm to the VIO system. And for more robust feature correspondence and better camera pose estimation, we propose an improved GMS method to eliminate the mismatched points. Besides, we adopt the loop closure strategy to correct the cumulative error of the VIO system. Finally, we estimate the camera pose, velocity and IMU bias simultaneously by minimizing the loss function which contains reprojection error and IMU error. A large number of experiments on EuRoC demonstrate that the proposed method outperforms the advanced VIO system ROVIO and is comparable to the state-of-the-art VIO system OKVIS.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127081897","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 : 2020-12-07DOI: 10.1109/iCAST51195.2020.9319480
T. Hashimoto, Kilho Shin, D. Shepard, T. Kuboyama
This paper presents an analysis of an Indonesian gender equality survey: in 2019, we conducted a survey of attitudes about gender roles in Indonesia and obtained data from 122 individuals. The obtained data were analyzed using our original clustering method (UFVS, Unsupervised Feature Value Selection) to form clusters. The extracted features characterized the clusters and helped to analyze the attitudes of Indonesians towards gender equality. This method allowed the respondents to be grouped by features and each group characteristics could be easily identified. It facilitated the understanding of the survey data.
{"title":"Indonesian Gender Equality Survey Analysis Using Feature Selection Based Clustering","authors":"T. Hashimoto, Kilho Shin, D. Shepard, T. Kuboyama","doi":"10.1109/iCAST51195.2020.9319480","DOIUrl":"https://doi.org/10.1109/iCAST51195.2020.9319480","url":null,"abstract":"This paper presents an analysis of an Indonesian gender equality survey: in 2019, we conducted a survey of attitudes about gender roles in Indonesia and obtained data from 122 individuals. The obtained data were analyzed using our original clustering method (UFVS, Unsupervised Feature Value Selection) to form clusters. The extracted features characterized the clusters and helped to analyze the attitudes of Indonesians towards gender equality. This method allowed the respondents to be grouped by features and each group characteristics could be easily identified. It facilitated the understanding of the survey data.","PeriodicalId":212570,"journal":{"name":"2020 11th International Conference on Awareness Science and Technology (iCAST)","volume":"194 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123737523","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：