Pub Date : 2022-08-31DOI: 10.1109/CSI54720.2022.9924105
P. Radhika Dileep, L. R. Deepthi
Networks can be used to represent a variety of real world complex interacting systems in which vertices represents interacting entities and a network link represents a connection between two nodes or entities. Citation graphs are widely utilized in a variety of graph mining situations like citation recommendation and locating research hotspots. Link prediction is considered as a significant task in data and graph mining and deals with prediction of the future or missing network links based on the given network knowledge. In this research, the problem of prediction of links in weighted citation network is addressed and also we compare how much weighing the network can improve the link prediction accuracy. Normally link prediction problems consider only the existence of links. This might lead to a less accurate prediction as it will not give the strength of the relationship between the two entities. In this study, we analyzed the Search Path Count method, which is used to assign weights to the citation links. So rather than just considering the presence of the links, two weighted path methods using Search Path Count weights are proposed in this research for link prediction. Experiments on real citation dataset show that using the Search Path Count weights to evaluate the relevance of the edges in citation networks improves the accuracy of link prediction systems.
{"title":"Analysis of Link Prediction Methods in Weighted and Unweighted Citation Network","authors":"P. Radhika Dileep, L. R. Deepthi","doi":"10.1109/CSI54720.2022.9924105","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924105","url":null,"abstract":"Networks can be used to represent a variety of real world complex interacting systems in which vertices represents interacting entities and a network link represents a connection between two nodes or entities. Citation graphs are widely utilized in a variety of graph mining situations like citation recommendation and locating research hotspots. Link prediction is considered as a significant task in data and graph mining and deals with prediction of the future or missing network links based on the given network knowledge. In this research, the problem of prediction of links in weighted citation network is addressed and also we compare how much weighing the network can improve the link prediction accuracy. Normally link prediction problems consider only the existence of links. This might lead to a less accurate prediction as it will not give the strength of the relationship between the two entities. In this study, we analyzed the Search Path Count method, which is used to assign weights to the citation links. So rather than just considering the presence of the links, two weighted path methods using Search Path Count weights are proposed in this research for link prediction. Experiments on real citation dataset show that using the Search Path Count weights to evaluate the relevance of the edges in citation networks improves the accuracy of link prediction systems.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122637854","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9924005
Parvathy S.R., Deepak Jayan P., Nimmy Pathrose, Rajesh K.R., Lekshmy Janardhanan R, James Varghese, V. S., Nimmy Mathew, Sujith B. Kallara
Red Palm Weevil infestation affecting palm trees is a major problem faced by coconut farmers, causing severe economic losses to palm cultivators worldwide. This infestation is fatal to the trees and can easily spread to nearby palms affecting a large number of trees, if left undetected. Detection of infestation at an early stage is critical for timely action to save palm trees. The paper presents a system for early detection of Red Palm Weevil infestation, by providing a smart, non-invasive and portable solution to the problem. The system based on an accelerometer sensor, analyses the signals from the palm and checks out for the unique features present in the signals generated by the pest. Audio, visual and SMS warnings are generated by the system on detecting infestation. The system also wirelessly communicates the infestation status to a remote database, which stores and maintains the historical data of palm trees monitored using this device, for facilitating pest control and management.
{"title":"Red Palm Weevil Detection System for Early Warning and Mitigation of Crop Loss","authors":"Parvathy S.R., Deepak Jayan P., Nimmy Pathrose, Rajesh K.R., Lekshmy Janardhanan R, James Varghese, V. S., Nimmy Mathew, Sujith B. Kallara","doi":"10.1109/CSI54720.2022.9924005","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924005","url":null,"abstract":"Red Palm Weevil infestation affecting palm trees is a major problem faced by coconut farmers, causing severe economic losses to palm cultivators worldwide. This infestation is fatal to the trees and can easily spread to nearby palms affecting a large number of trees, if left undetected. Detection of infestation at an early stage is critical for timely action to save palm trees. The paper presents a system for early detection of Red Palm Weevil infestation, by providing a smart, non-invasive and portable solution to the problem. The system based on an accelerometer sensor, analyses the signals from the palm and checks out for the unique features present in the signals generated by the pest. Audio, visual and SMS warnings are generated by the system on detecting infestation. The system also wirelessly communicates the infestation status to a remote database, which stores and maintains the historical data of palm trees monitored using this device, for facilitating pest control and management.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122935529","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9924077
K. Remya Revi, Meera Mary Isaac, R. Antony, M. Wilscy
Advancements in AI techniques like Generative Adversarial Network (GAN) facilitate the creation of realistic-looking fake face images and these images are used to create fake profiles on various social media platforms. In this work, we develop deep learning-based binary classification models to distinguish GAN-generated fake face images from camera-captured real face images. The classification models are developed by fine-tuning three lightweight state-of-the-art pre-trained Convolutional Neural Networks (CNNs) - GoogLeNet, ResNet-18, and MobileNet-v2 -using the transfer learning approach. In this method, instead of RGB images, joint color texture feature maps of the images obtained using Opponent Color-Local Binary Pattern (OC-LBP) are used as input to the CNN. For the experimental analysis, we use datasets that contain fake face images generated by Progressive Growing GAN (PGGAN) and Style-based GAN (StyleGAN2), and camera-captured real face images from CelebFaces Attributes- High Quality (CelebA-HQ) and Flickr Faces High Quality (FFHQ) datasets. The proposed method shows remarkable performance in terms of test accuracy, generalization capability, and robustness against JPEG compression. Also, the method exhibits excellent performance when compared with state-of-the-art methods.
生成对抗网络(GAN)等人工智能技术的进步有助于创建逼真的假面部图像，这些图像用于在各种社交媒体平台上创建假个人资料。在这项工作中，我们开发了基于深度学习的二分类模型，以区分gan生成的假人脸图像和相机捕获的真实人脸图像。分类模型是通过使用迁移学习方法微调三个轻量级的最先进的预训练卷积神经网络(cnn)——GoogLeNet、ResNet-18和MobileNet-v2——开发的。该方法使用对手颜色局部二值模式(OC-LBP)获得的图像的联合颜色纹理特征图作为CNN的输入，而不是RGB图像。为了进行实验分析，我们使用了包含由Progressive Growing GAN (PGGAN)和Style-based GAN (StyleGAN2)生成的假人脸图像的数据集，以及从CelebA-HQ和Flickr Faces High Quality (FFHQ)数据集捕获的真实人脸图像。该方法在测试精度、泛化能力和对JPEG压缩的鲁棒性方面表现出显著的性能。此外，与最先进的方法相比，该方法表现出优异的性能。
{"title":"GAN-generated Fake Face Image Detection using Opponent Color Local Binary Pattern and Deep Learning Technique","authors":"K. Remya Revi, Meera Mary Isaac, R. Antony, M. Wilscy","doi":"10.1109/CSI54720.2022.9924077","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924077","url":null,"abstract":"Advancements in AI techniques like Generative Adversarial Network (GAN) facilitate the creation of realistic-looking fake face images and these images are used to create fake profiles on various social media platforms. In this work, we develop deep learning-based binary classification models to distinguish GAN-generated fake face images from camera-captured real face images. The classification models are developed by fine-tuning three lightweight state-of-the-art pre-trained Convolutional Neural Networks (CNNs) - GoogLeNet, ResNet-18, and MobileNet-v2 -using the transfer learning approach. In this method, instead of RGB images, joint color texture feature maps of the images obtained using Opponent Color-Local Binary Pattern (OC-LBP) are used as input to the CNN. For the experimental analysis, we use datasets that contain fake face images generated by Progressive Growing GAN (PGGAN) and Style-based GAN (StyleGAN2), and camera-captured real face images from CelebFaces Attributes- High Quality (CelebA-HQ) and Flickr Faces High Quality (FFHQ) datasets. The proposed method shows remarkable performance in terms of test accuracy, generalization capability, and robustness against JPEG compression. Also, the method exhibits excellent performance when compared with state-of-the-art methods.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126822273","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9924135
Ahmed Abul Hasanaath, AbuMuhammad Moinuddeen, Nazeeruddin Mohammad, M. Khan, Ahmed A. Hussain
Continuous and real-time monitoring of road quality conditions is essential for the maintenance of roads and to ensure the safety of drivers and their vehicles. However, the continuous monitoring of thousands of kilometers of roads and highways is a very tedious, time-consuming, error-prone, and expensive operation. A deep learning based approach that can automatically classify the road condition can help tremendously in cutting down the time, effort, accuracy, and cost for monitoring and maintenance of vast road infrastructure. This paper proposes a mechanism to continuously monitor deteriorating road conditions at the city or municipality level in real time and classify them into four different categories (good, medium, bad and unpaved) using custom-built and transfer learning from pre-trained deep learning models (VGG16 and MobileNetV2). The dataset is collected from different roads in the Kingdom of Saudi Arabia. The dataset is composed of close-up road images taken in real time (while driving the car) at regular intervals using an Android App. In the data capture model, the Android App helps to easily tag (label) the captured images for model training purposes. In the classifier mode, the Android app uses the developed deep learning model to classify the captured image and then transmits the medium, bad or unpaved road condition to the central server along with longitude and latitude information to update the centralized map of the city (or municipality). The proposed approach provides an accuracy of 98.6 % to classify the road condition based on images captured during real time driving of the vehicle.
{"title":"Continuous and Realtime Road Condition Assessment Using Deep Learning","authors":"Ahmed Abul Hasanaath, AbuMuhammad Moinuddeen, Nazeeruddin Mohammad, M. Khan, Ahmed A. Hussain","doi":"10.1109/CSI54720.2022.9924135","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924135","url":null,"abstract":"Continuous and real-time monitoring of road quality conditions is essential for the maintenance of roads and to ensure the safety of drivers and their vehicles. However, the continuous monitoring of thousands of kilometers of roads and highways is a very tedious, time-consuming, error-prone, and expensive operation. A deep learning based approach that can automatically classify the road condition can help tremendously in cutting down the time, effort, accuracy, and cost for monitoring and maintenance of vast road infrastructure. This paper proposes a mechanism to continuously monitor deteriorating road conditions at the city or municipality level in real time and classify them into four different categories (good, medium, bad and unpaved) using custom-built and transfer learning from pre-trained deep learning models (VGG16 and MobileNetV2). The dataset is collected from different roads in the Kingdom of Saudi Arabia. The dataset is composed of close-up road images taken in real time (while driving the car) at regular intervals using an Android App. In the data capture model, the Android App helps to easily tag (label) the captured images for model training purposes. In the classifier mode, the Android app uses the developed deep learning model to classify the captured image and then transmits the medium, bad or unpaved road condition to the central server along with longitude and latitude information to update the centralized map of the city (or municipality). The proposed approach provides an accuracy of 98.6 % to classify the road condition based on images captured during real time driving of the vehicle.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128970948","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9924065
Aastha Dak, Asfia Urooj, R. Radhakrishnan
This work addresses the problem of tracking and interception of a ballistic target having spiralling motion on re-entry. Interception is achieved by an interceptor missile which collects the required measurements using an inbuilt seeker, such that accurate estimates for target states are generated. The usual assumption that these measurements are corrupted by Gaussian noise is revisited, as significant outliers are observed in radar measurements. Since the conventional estimators tend to diverge in the presence of measurement outliers, this work propose an accurate and robust estimation algorithm by incorporating the maximum correntropy (MC) criterion. Hence, a Cauchy kernel based MC unscented Kalman filter (CM-UKF) is proposed for accurate state estimation. Also, proportional navigation guidance (PNG) law is implemented such that a possible interception is realized. The estimation accuracy of CM-UKF along with the PNG law is compared with that of the traditional UKF and Gaussian kernel based MC UKF (MC-UKF), by evaluating the average miss-distance and root mean square error (RMSE) in states.
{"title":"Estimation and Interception of a Spiralling Target on Reentry in the Presence of non-Gaussian Measurement Noise","authors":"Aastha Dak, Asfia Urooj, R. Radhakrishnan","doi":"10.1109/CSI54720.2022.9924065","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924065","url":null,"abstract":"This work addresses the problem of tracking and interception of a ballistic target having spiralling motion on re-entry. Interception is achieved by an interceptor missile which collects the required measurements using an inbuilt seeker, such that accurate estimates for target states are generated. The usual assumption that these measurements are corrupted by Gaussian noise is revisited, as significant outliers are observed in radar measurements. Since the conventional estimators tend to diverge in the presence of measurement outliers, this work propose an accurate and robust estimation algorithm by incorporating the maximum correntropy (MC) criterion. Hence, a Cauchy kernel based MC unscented Kalman filter (CM-UKF) is proposed for accurate state estimation. Also, proportional navigation guidance (PNG) law is implemented such that a possible interception is realized. The estimation accuracy of CM-UKF along with the PNG law is compared with that of the traditional UKF and Gaussian kernel based MC UKF (MC-UKF), by evaluating the average miss-distance and root mean square error (RMSE) in states.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114889192","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9924042
Trupti Pravin Loya, Ragini Dadgal
Background-Range of motion (ROM) is a module of evaluation of the body that is utilized in protocols of identification and rehabilitation. Goniometer is the device which are routinely used to assess musculoskeletal range of motion. Clinicians now have access to smartphone goniometer software. This study proposes a novel gadget that measures different joints ROM applying a smartphone goniometer. Objective- To test the reliability, relevance, use of smartphone goniometer application for ROM measurement at different joints. Sources of information-Four sources of information (Pubmed, Scopus, Web of Science) have been looked from 2012 to 2022. Review method-Studies on the reliability, validity and use of smartphone goniometer application were included. High quality experiment trails were chosen for the study. Result-98 articles were extracted; 4 articles were included in study which emphasized the use of Smartphone goniometer application (SGA) for ROM measurement in different joints. Studies shows inconsistent results on use of smartphone goniometer application on different joints. Conclusion- The measurement produced with the evaluated SGA are accurate as obtained with the UG. Therefore, it's a valid instrument for measuring range of motion for different joints.
活动范围(ROM)是身体评估的一个模块，用于识别和康复方案。测角仪是一种常规用于评估肌肉骨骼活动范围的设备。临床医生现在可以使用智能手机测角仪软件。本研究提出一种利用智能手机测角仪测量不同关节ROM的新装置。目的:测试智能手机测角仪应用于不同关节ROM测量的可靠性、相关性。信息来源从2012年到2022年，四种信息来源(Pubmed, Scopus, Web of Science)已经被研究过。综述方法:对智能手机测角仪应用程序的信度、效度和使用情况进行了研究。本研究选择了高质量的实验轨迹。结果:共提取文献98篇;本研究纳入4篇文章，强调使用智能手机测角仪应用程序(SGA)测量不同关节的ROM。研究表明，在不同关节上使用智能手机测角仪应用程序的结果不一致。结论-使用评估的SGA产生的测量结果与使用UG获得的结果一样准确。因此，它是测量不同关节活动范围的有效工具。
{"title":"Use of Smartphone Goniometer Application for Range of Motion Measurement in Different Joints: Review Article","authors":"Trupti Pravin Loya, Ragini Dadgal","doi":"10.1109/CSI54720.2022.9924042","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924042","url":null,"abstract":"Background-Range of motion (ROM) is a module of evaluation of the body that is utilized in protocols of identification and rehabilitation. Goniometer is the device which are routinely used to assess musculoskeletal range of motion. Clinicians now have access to smartphone goniometer software. This study proposes a novel gadget that measures different joints ROM applying a smartphone goniometer. Objective- To test the reliability, relevance, use of smartphone goniometer application for ROM measurement at different joints. Sources of information-Four sources of information (Pubmed, Scopus, Web of Science) have been looked from 2012 to 2022. Review method-Studies on the reliability, validity and use of smartphone goniometer application were included. High quality experiment trails were chosen for the study. Result-98 articles were extracted; 4 articles were included in study which emphasized the use of Smartphone goniometer application (SGA) for ROM measurement in different joints. Studies shows inconsistent results on use of smartphone goniometer application on different joints. Conclusion- The measurement produced with the evaluated SGA are accurate as obtained with the UG. Therefore, it's a valid instrument for measuring range of motion for different joints.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"114 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131728176","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9924016
Aishvarya Garg, S. Nigam, R. Singh
Human activity recognition is a wide research area of computer vision that finds applications in smart surveillance system, healthcare, and human robotic interactions. Nowadays, deep learning methods have achieved more interest due to its ability of executing feature extraction and classification steps simultaneously. In this paper, we have focused on the vision based human activity recognition using deep learning algorithms. Long short term memory (LSTM) is a special form of recurrent neural networks (RNN), specifically designed for long term data dependencies. Also it is a known fact that among deep learning algorithms, convolutional neural networks (CNN) have earned high performance in image classification. To overcome the limitation of LSTM in case of classification of static images, a hybrid CNN-LSTM model is proposed in which features are firstly extracted through CNN and then feed to LSTM as a sequence by the means of time distributed layer. This model is utilized for classifying six activities from two datasets which have shown the accuracy of 96.24% and 93.39% on KTH and Weizmann datasets, respectively. We have also implemented the CNN and LSTM models separately on these datasets with same parameters as used in hybrid model to study their impact on accuracy and loss.
{"title":"Vision based Human Activity Recognition using Hybrid Deep Learning","authors":"Aishvarya Garg, S. Nigam, R. Singh","doi":"10.1109/CSI54720.2022.9924016","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924016","url":null,"abstract":"Human activity recognition is a wide research area of computer vision that finds applications in smart surveillance system, healthcare, and human robotic interactions. Nowadays, deep learning methods have achieved more interest due to its ability of executing feature extraction and classification steps simultaneously. In this paper, we have focused on the vision based human activity recognition using deep learning algorithms. Long short term memory (LSTM) is a special form of recurrent neural networks (RNN), specifically designed for long term data dependencies. Also it is a known fact that among deep learning algorithms, convolutional neural networks (CNN) have earned high performance in image classification. To overcome the limitation of LSTM in case of classification of static images, a hybrid CNN-LSTM model is proposed in which features are firstly extracted through CNN and then feed to LSTM as a sequence by the means of time distributed layer. This model is utilized for classifying six activities from two datasets which have shown the accuracy of 96.24% and 93.39% on KTH and Weizmann datasets, respectively. We have also implemented the CNN and LSTM models separately on these datasets with same parameters as used in hybrid model to study their impact on accuracy and loss.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129018300","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9924106
Nivea Kesav, Jibukumar M.G
The Covid-19 virus, which initially originated in Wuhan, China, was declared a pandemic by the World Health Organization on March 11, 2020. Since then, it has had a tremendous impact on human health and the World economy. Rapid identification and treatment of the disease have been a prime concern. Analysis of Radiographic Chest X-ray images has become an effective way to determine the disease and its severity. This paper proposes a low complex methodology that uses Convolutional Neural Networks (CNN) for classifying three types of X-ray images, Covid-19, Healthy and Viral Pneumonia. The architecture consists of two channels: the main channel with four convolutional layers with increasing order of filter size and a side channel with two convolutional layers of the same filter size. The architecture performs well with an overall accuracy of 95.24% and with only 89,41,783 parameters. It has been compared with different deep CNN s and several state-of-the-art works of literature.
{"title":"Improved Bi-Channel CNN For Covid-19 Diagnosis","authors":"Nivea Kesav, Jibukumar M.G","doi":"10.1109/CSI54720.2022.9924106","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9924106","url":null,"abstract":"The Covid-19 virus, which initially originated in Wuhan, China, was declared a pandemic by the World Health Organization on March 11, 2020. Since then, it has had a tremendous impact on human health and the World economy. Rapid identification and treatment of the disease have been a prime concern. Analysis of Radiographic Chest X-ray images has become an effective way to determine the disease and its severity. This paper proposes a low complex methodology that uses Convolutional Neural Networks (CNN) for classifying three types of X-ray images, Covid-19, Healthy and Viral Pneumonia. The architecture consists of two channels: the main channel with four convolutional layers with increasing order of filter size and a side channel with two convolutional layers of the same filter size. The architecture performs well with an overall accuracy of 95.24% and with only 89,41,783 parameters. It has been compared with different deep CNN s and several state-of-the-art works of literature.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115554181","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9923996
Murali Krishna Madugula, Santosh Kumar Majhi, Nibedan Panda
The Subset-Sum Problem (SSP) ensures a significant role in various practical applications, which include cryptography and coding theory owing to the importance in the functionality of some of the public key cryptography systems. Consider the set S of n real numbers, where the 2n - 1 diverse subsets are presented without including the empty set. The SSP is defined as the determination of N subsets, where the summation of elements in the subset needs to be N the smallest over all the possible subsets. This problem was involved in diverse applications in operations research and practice. But, the problem is very complex in computation. Hence, this paper aims to solve the SSP with a well-enabled meta-heuristic algorithm named Arithmetic Optimization Algorithm (AOA). Here, a novel optimization algorithm is developed for reducing the error among the target and attained a solution, and also to solve the SSA issue. At last, the simulation analysis reveals that the suggested AOA can ensure optimal results when using the benchmark data.
子集和问题（Subset-Sum Problem，SSP）在各种实际应用中发挥着重要作用，其中包括密码学和编码理论，因为它在一些公钥密码学系统的功能中具有重要作用。考虑由 n 个实数组成的集合 S，其中有 2n - 1 个不同的子集，但不包括空集。SSP 的定义是确定 N 个子集，其中子集中元素的和必须是所有可能子集中最小的 N 个。这个问题在运筹学研究和实践中有多种应用。但是，这个问题的计算非常复杂。因此，本文旨在使用一种名为算术优化算法（AOA）的功能强大的元启发式算法来解决 SSP 问题。本文开发了一种新颖的优化算法，以减少目标之间的误差并获得一个解决方案，同时解决 SSA 问题。最后，仿真分析表明，建议的 AOA 在使用基准数据时能确保获得最佳结果。
{"title":"An Efficient Arithmetic Optimization Algorithm for Solving Subset-sum Problem","authors":"Murali Krishna Madugula, Santosh Kumar Majhi, Nibedan Panda","doi":"10.1109/CSI54720.2022.9923996","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9923996","url":null,"abstract":"The Subset-Sum Problem (SSP) ensures a significant role in various practical applications, which include cryptography and coding theory owing to the importance in the functionality of some of the public key cryptography systems. Consider the set S of n real numbers, where the 2n - 1 diverse subsets are presented without including the empty set. The SSP is defined as the determination of N subsets, where the summation of elements in the subset needs to be N the smallest over all the possible subsets. This problem was involved in diverse applications in operations research and practice. But, the problem is very complex in computation. Hence, this paper aims to solve the SSP with a well-enabled meta-heuristic algorithm named Arithmetic Optimization Algorithm (AOA). Here, a novel optimization algorithm is developed for reducing the error among the target and attained a solution, and also to solve the SSA issue. At last, the simulation analysis reveals that the suggested AOA can ensure optimal results when using the benchmark data.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128023701","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 : 2022-08-31DOI: 10.1109/CSI54720.2022.9923997
Vivek Rahinj, Rashmika K. Patole, S. Metkar
Audio authentication is the primary task in an audio forensics scenario in which audio tampering detection is one of the objectives. In this paper, we offer a fresh approach to audio tampering detection using supervised learning and active learning methods. The present techniques are based on supervised learning, and they require a massive amount of labeled data for classification. There is very little availability of standard data. The paper provides a comparative study of supervised and active learning approaches. The work uses unlabeled dataset for classification which is the primary focus in any active learning method. The proposed work uses less than 1-sec audio files for copy and move tampering. Result gives 92.78% accuracy for supervised learning using stft whereas for active learning it gives 87.38%. Active learning reduces the cost of annotation as we do not have to label all the data.
{"title":"Active Learning Based Audio Tampering Detection","authors":"Vivek Rahinj, Rashmika K. Patole, S. Metkar","doi":"10.1109/CSI54720.2022.9923997","DOIUrl":"https://doi.org/10.1109/CSI54720.2022.9923997","url":null,"abstract":"Audio authentication is the primary task in an audio forensics scenario in which audio tampering detection is one of the objectives. In this paper, we offer a fresh approach to audio tampering detection using supervised learning and active learning methods. The present techniques are based on supervised learning, and they require a massive amount of labeled data for classification. There is very little availability of standard data. The paper provides a comparative study of supervised and active learning approaches. The work uses unlabeled dataset for classification which is the primary focus in any active learning method. The proposed work uses less than 1-sec audio files for copy and move tampering. Result gives 92.78% accuracy for supervised learning using stft whereas for active learning it gives 87.38%. Active learning reduces the cost of annotation as we do not have to label all the data.","PeriodicalId":221137,"journal":{"name":"2022 International Conference on Connected Systems & Intelligence (CSI)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125132228","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：