Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136047
Pankaj Jha, Anubhav Kumar, N. Sharma, Divya Saxena
a compact dual-band textile antenna is designed for wearable 5G and Wi-Fi 6E communication. The full ground of antenna is modified using two circular slots with an appropriate radius for improving the impedance matching towards the desired band. The designed vase-shaped ground structure with an elliptical-shaped radiator is obtaining the high-speed Wi-Fi 6E band in the antenna. Further, a CSRR is accomplished in the radiator which perturbs the surface wave effectively and the antenna achieves a lower operating band in the allotted 5G spectrum. The CSRR also improves the total gain of the antenna in the lower operating band which makes it more advantageous in wearable wireless communication. 10 dB IBW of dual band antenna is varying between 3.4 - 4.1 GHz and 6.2 – 7.1 GHz. To confirm the on-body radiation effect, the specific absorption ratio (SAR) is extracted from three layered human phantom models and the values of obtained SAR are quite healthy.
{"title":"CSRR Loaded Compact Textile Antenna with Defected Ground for Wearable 5G and Wi-Fi 6E Applications","authors":"Pankaj Jha, Anubhav Kumar, N. Sharma, Divya Saxena","doi":"10.1109/PCEMS58491.2023.10136047","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136047","url":null,"abstract":"a compact dual-band textile antenna is designed for wearable 5G and Wi-Fi 6E communication. The full ground of antenna is modified using two circular slots with an appropriate radius for improving the impedance matching towards the desired band. The designed vase-shaped ground structure with an elliptical-shaped radiator is obtaining the high-speed Wi-Fi 6E band in the antenna. Further, a CSRR is accomplished in the radiator which perturbs the surface wave effectively and the antenna achieves a lower operating band in the allotted 5G spectrum. The CSRR also improves the total gain of the antenna in the lower operating band which makes it more advantageous in wearable wireless communication. 10 dB IBW of dual band antenna is varying between 3.4 - 4.1 GHz and 6.2 – 7.1 GHz. To confirm the on-body radiation effect, the specific absorption ratio (SAR) is extracted from three layered human phantom models and the values of obtained SAR are quite healthy.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130506953","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136100
Nachiket Sawwalakhe, Yash Jain, Shubham Patil, Ajay Jondhale, A. Keskar
More than 40% of the world’s population relies heavily on agriculture for their livelihood. In recent years, there has been an increased interest in the development of autonomous vehicles, specifically farming robots, to improve agricultural processes and yield. Traditional farming methods, such as the manual spreading of fertiliser and carrying heavy pesticide cans, can be labour-intensive and time-consuming for farmers. To address these challenges, we present a self-contained, intelligent farming robot that can handle these tasks quickly and precisely to help with these problems. Without the need for human assistance, this robot can navigate rocky agricultural terrain, ensuring that crops are properly fertilised and cared for. The proposed robot has six wheels and a rocker-bogie design, and it is run by a Raspberry Pi microcontroller. The bot’s main purpose is to move autonomously within a farm in a structured manner, such as a straight row, and to fertilise and apply pesticides to plants as needed. The bot also utilises computer vision technology to perform these tasks effectively. Farmers can increase their operational efficiency and save time and resources by implementing the suggested bot. Since there is no longer a need for manual labour, farmers are less physically taxed, and their farms are more profitable overall. The design and implementation of this autonomous plantation layout and fertiliser bot offer a promising solution to advance the agriculture industry.
{"title":"Design and Implementation of an Autonomous Plantation Layout Follower Fertilizer Bot for Advanced Farming","authors":"Nachiket Sawwalakhe, Yash Jain, Shubham Patil, Ajay Jondhale, A. Keskar","doi":"10.1109/PCEMS58491.2023.10136100","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136100","url":null,"abstract":"More than 40% of the world’s population relies heavily on agriculture for their livelihood. In recent years, there has been an increased interest in the development of autonomous vehicles, specifically farming robots, to improve agricultural processes and yield. Traditional farming methods, such as the manual spreading of fertiliser and carrying heavy pesticide cans, can be labour-intensive and time-consuming for farmers. To address these challenges, we present a self-contained, intelligent farming robot that can handle these tasks quickly and precisely to help with these problems. Without the need for human assistance, this robot can navigate rocky agricultural terrain, ensuring that crops are properly fertilised and cared for. The proposed robot has six wheels and a rocker-bogie design, and it is run by a Raspberry Pi microcontroller. The bot’s main purpose is to move autonomously within a farm in a structured manner, such as a straight row, and to fertilise and apply pesticides to plants as needed. The bot also utilises computer vision technology to perform these tasks effectively. Farmers can increase their operational efficiency and save time and resources by implementing the suggested bot. Since there is no longer a need for manual labour, farmers are less physically taxed, and their farms are more profitable overall. The design and implementation of this autonomous plantation layout and fertiliser bot offer a promising solution to advance the agriculture industry.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116253120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136054
Swapnil P. Badar, K. Khanchandani, Pravin Wankhede
The information sent by the transmitter to the receiver through the channel may be user information or control information. For error-free communication, errorcorrecting codes are needed to detect and correct errors. ECCs like low-density parity check (LDPC) and polar code are selected for channel data and channel control coding, respectively, for 5G wireless communication. Compared to LDPC and turbo codes, Polar code has the ability to use all channel capacity. The Successive Cancellation decoder is a basic polar decoder, which has longer latency due to its sequential nature. A polar decoder with special nodes is proposed in this paper. This fast polar decoder makes the decoding operation faster. The polar decoder is designed using special nodes–Rate-zero, Rate-one, Single Parity Check, and Repetition nodes. These special nodes are generated from the proposed node generator circuit. VLSI architectures of special nodes and fast polar decoder are generated by the Xilinx platform, which is shown in this paper
{"title":"Fast Polar Decoder Implementation using Special Nodes","authors":"Swapnil P. Badar, K. Khanchandani, Pravin Wankhede","doi":"10.1109/PCEMS58491.2023.10136054","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136054","url":null,"abstract":"The information sent by the transmitter to the receiver through the channel may be user information or control information. For error-free communication, errorcorrecting codes are needed to detect and correct errors. ECCs like low-density parity check (LDPC) and polar code are selected for channel data and channel control coding, respectively, for 5G wireless communication. Compared to LDPC and turbo codes, Polar code has the ability to use all channel capacity. The Successive Cancellation decoder is a basic polar decoder, which has longer latency due to its sequential nature. A polar decoder with special nodes is proposed in this paper. This fast polar decoder makes the decoding operation faster. The polar decoder is designed using special nodes–Rate-zero, Rate-one, Single Parity Check, and Repetition nodes. These special nodes are generated from the proposed node generator circuit. VLSI architectures of special nodes and fast polar decoder are generated by the Xilinx platform, which is shown in this paper","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"89 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116302804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136062
K.Bhanu Prathap, G.Divya Swaroop, B.Praveen Kumar, V. Kamble, Mayuri A Parate
Normal people can readily connect and communicate with one another, however, those with hearing and speech impairments find it difficult to converse with normal-hearing people without the assistance of a translator. The only way deaf and dumb people can communicate is through Sign Language. Indian Sign Language has its own grammar, syntax, vocabulary, and unique language features. We propose two methods, namely Bidirectional LSTM and BERT Transformer to address the problem of sign language translation. The proposed work is validated on standard datasets and provides promising results. The INCLUDE-50 dataset is used to validate the performance of the proposed algorithm. The deep neural network is evaluated using a combination of approaches for augmentation of the data, features extraction using the mediapipe.On the Dataset INCLUDE 50 the best performing model obtained an accuracy of 89.5%. This model employs a feature extractor that has been pre-trained, as well as an encoder and a decoder.
正常人可以很容易地相互联系和交流，然而，那些有听力和语言障碍的人发现，如果没有翻译的帮助，很难与听力正常的人交谈。聋哑人交流的唯一方式是通过手语。印度手语有自己的语法、句法、词汇和独特的语言特征。我们提出了两种方法，即双向LSTM和BERT转换器来解决手语翻译问题。所提出的工作在标准数据集上进行了验证，并提供了有希望的结果。使用INCLUDE-50数据集验证了所提出算法的性能。深度神经网络的评估使用了数据增强和mediapipe特征提取的组合方法。在Dataset INCLUDE 50上，表现最好的模型获得了89.5%的准确率。该模型采用了一个预训练的特征提取器，以及一个编码器和一个解码器。
{"title":"ISLR: Indian Sign Language Recognition","authors":"K.Bhanu Prathap, G.Divya Swaroop, B.Praveen Kumar, V. Kamble, Mayuri A Parate","doi":"10.1109/PCEMS58491.2023.10136062","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136062","url":null,"abstract":"Normal people can readily connect and communicate with one another, however, those with hearing and speech impairments find it difficult to converse with normal-hearing people without the assistance of a translator. The only way deaf and dumb people can communicate is through Sign Language. Indian Sign Language has its own grammar, syntax, vocabulary, and unique language features. We propose two methods, namely Bidirectional LSTM and BERT Transformer to address the problem of sign language translation. The proposed work is validated on standard datasets and provides promising results. The INCLUDE-50 dataset is used to validate the performance of the proposed algorithm. The deep neural network is evaluated using a combination of approaches for augmentation of the data, features extraction using the mediapipe.On the Dataset INCLUDE 50 the best performing model obtained an accuracy of 89.5%. This model employs a feature extractor that has been pre-trained, as well as an encoder and a decoder.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"568 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120875810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136055
Neha Arora, K. K. Gola, S. Gulati, P. Chutani
what used to be an annual or bi-annual phenomenon of toys/sports/utility item purchase nearly two decades back, is a weekly/bi-weekly/daily transaction now-adays. Toys purchase by kids in a very frequent transaction that happens almost every alternate day in a big segment of society and thus produce high volumes of data. Consequently, there is rising scope to apply data mining methods to obtain toys/items_of_interest buying patterns amongst kids. In the present piece of research, we have applied Apriori algorithm to perform data mining using the data collected through a Google form after circulating children’s (age group 5-9) acquaintance of toys, across the country; the survey got carried out through students of two engineering colleges where diverse group of students from different parts of the country are studying. Nine association rules were achieved after applying Apriori Algorithm on the data set of the Toys/Sports items thus formed. Further, accuracy of framed rules has also been manually validated by the store owner; Beyblades and Carom are the most preferred toys/sports items; whereas Bicycle and Bat-Ball falls at second position in the list. The results provide very useful association amongst toys/Sports.
{"title":"Survey, Analysis and Association Rules derivation using Apriori Method for buying preference amongst kids of age-group 5 to 9 in India","authors":"Neha Arora, K. K. Gola, S. Gulati, P. Chutani","doi":"10.1109/PCEMS58491.2023.10136055","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136055","url":null,"abstract":"what used to be an annual or bi-annual phenomenon of toys/sports/utility item purchase nearly two decades back, is a weekly/bi-weekly/daily transaction now-adays. Toys purchase by kids in a very frequent transaction that happens almost every alternate day in a big segment of society and thus produce high volumes of data. Consequently, there is rising scope to apply data mining methods to obtain toys/items_of_interest buying patterns amongst kids. In the present piece of research, we have applied Apriori algorithm to perform data mining using the data collected through a Google form after circulating children’s (age group 5-9) acquaintance of toys, across the country; the survey got carried out through students of two engineering colleges where diverse group of students from different parts of the country are studying. Nine association rules were achieved after applying Apriori Algorithm on the data set of the Toys/Sports items thus formed. Further, accuracy of framed rules has also been manually validated by the store owner; Beyblades and Carom are the most preferred toys/sports items; whereas Bicycle and Bat-Ball falls at second position in the list. The results provide very useful association amongst toys/Sports.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115190886","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}
Customers are becoming more concerned about the QoS (Quality of Service) that businesses can give them in today’s modern world. Because the services offered by many vendors are not very distinguishable, businesses have increased rivalry to maintain and improve their quality of service. Customer Relationship Management (CRM) systems are used to provide businesses with the capacity to boost their profitability by acquiring new consumers, establishing a continuous relationship with existing customers, as well as keeping more of their existing customers. Machine Learning (ML) algorithms are used in CRM (Customer Relationship Management) systems to evaluate personal & behavioral data from clients. This gives a company a competitive edge by improving the percentage of consumers they keep as clients. This research attempts to evaluate and assess the performance of several machine learning (ML) approaches to solve the subscriber prediction issue in email marketing. Different analytical machine learning methods that belong to diverse types of learning are selected for this work, especially classification and regressor techniques. Models were used on the dataset of emails that comprises twenty-three features. The experimental outcome demonstrates that RF (Random Forest) & Adaboost outperform all other machine learning methods with an almost similar accuracy of 95%. KNN and ensemble approach achieved the highest 89.8% and 91% R2 scores. The comparison found that the ensemble approach outperforms state-of-arts machine learning methods regarding accuracy, error value, and R2 score.
{"title":"A Novel Approach for Enhancing Customer Retention Using Machine Learning Techniques in Email Marketing Application","authors":"Dharmveer Yadav, Jagriti Singh, Priti Verma, Vikram Rajpoot, Gunjan Chhabra","doi":"10.1109/PCEMS58491.2023.10136072","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136072","url":null,"abstract":"Customers are becoming more concerned about the QoS (Quality of Service) that businesses can give them in today’s modern world. Because the services offered by many vendors are not very distinguishable, businesses have increased rivalry to maintain and improve their quality of service. Customer Relationship Management (CRM) systems are used to provide businesses with the capacity to boost their profitability by acquiring new consumers, establishing a continuous relationship with existing customers, as well as keeping more of their existing customers. Machine Learning (ML) algorithms are used in CRM (Customer Relationship Management) systems to evaluate personal & behavioral data from clients. This gives a company a competitive edge by improving the percentage of consumers they keep as clients. This research attempts to evaluate and assess the performance of several machine learning (ML) approaches to solve the subscriber prediction issue in email marketing. Different analytical machine learning methods that belong to diverse types of learning are selected for this work, especially classification and regressor techniques. Models were used on the dataset of emails that comprises twenty-three features. The experimental outcome demonstrates that RF (Random Forest) & Adaboost outperform all other machine learning methods with an almost similar accuracy of 95%. KNN and ensemble approach achieved the highest 89.8% and 91% R2 scores. The comparison found that the ensemble approach outperforms state-of-arts machine learning methods regarding accuracy, error value, and R2 score.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"2021 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128062687","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136118
Anamika Bisane, Shivanand Chandravanshi, P. Thakre, Purab Kesharwani, Atiya Khan
Digital reviews now have a significant impact on how consumers communicate globally and how they make purchases. When a buyer looks at the product’s ratings and reviews, they are frequently confused by the sheer volume of them. In the proposed study, product reviews are classified into positive and negative sentiments using the VADER (Valence Aware Dictionary for Sentiment Reasoning), a machine learning model that classifies reviews into positive and negative sentiments based on attributes discovered by the model that is used in the proposed work to categories product reviews into positive and negative categories. We provide the consumer with a graph of the number of good and negative reviews for the product they are interested in, as well as the total positive and negative review polarity for the item. To save clients’ time, a graphical representation of the analysis is also given.
{"title":"A Comprehensive Product Review System for Improved Customer Satisfaction","authors":"Anamika Bisane, Shivanand Chandravanshi, P. Thakre, Purab Kesharwani, Atiya Khan","doi":"10.1109/PCEMS58491.2023.10136118","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136118","url":null,"abstract":"Digital reviews now have a significant impact on how consumers communicate globally and how they make purchases. When a buyer looks at the product’s ratings and reviews, they are frequently confused by the sheer volume of them. In the proposed study, product reviews are classified into positive and negative sentiments using the VADER (Valence Aware Dictionary for Sentiment Reasoning), a machine learning model that classifies reviews into positive and negative sentiments based on attributes discovered by the model that is used in the proposed work to categories product reviews into positive and negative categories. We provide the consumer with a graph of the number of good and negative reviews for the product they are interested in, as well as the total positive and negative review polarity for the item. To save clients’ time, a graphical representation of the analysis is also given.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116559425","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136110
Kazi Newaj Faisal, R. Sharma
The Wigner-Ville distribution (WVD) is a widely used tool in the time-frequency analysis of non-stationary signals. However, the presence of false-terms in WVD for multicomponent signals can limit its applicability and interpretation. Various kernel and window-based smoothing methods have been used to remove false-terms from WVD, but they often come at the cost of reduced time-frequency resolution of autoterms. This paper proposes a novel sliding time and frequency windowing-based technique for removing false-terms from WVD, which aims to overcome the limitations of kernel-based methods. The proposed method segments a multi-component signal using overlapping windows in time and frequency domains successively and the WVD of each windowed signal is computed. The WVDs of all windowed signals are added together to obtain the falseterm free WVD. Energy scaling is also applied to minimize the effect of overlapping windows. Performance of the proposed method is evaluated for different multi-component synthetic signals and a natural ECG signal using various performance measures. The simulation results demonstrate that the proposed method can effectively remove false-terms from the WVD with improved auto-term enhancement and time-frequency resolution. Results from the proposed method are also compared with different kernel and window-based smoothing methods to show its superiority over these methods.
{"title":"Suppression of False-terms in Wigner-Ville Distribution using Time and Frequency Windowing","authors":"Kazi Newaj Faisal, R. Sharma","doi":"10.1109/PCEMS58491.2023.10136110","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136110","url":null,"abstract":"The Wigner-Ville distribution (WVD) is a widely used tool in the time-frequency analysis of non-stationary signals. However, the presence of false-terms in WVD for multicomponent signals can limit its applicability and interpretation. Various kernel and window-based smoothing methods have been used to remove false-terms from WVD, but they often come at the cost of reduced time-frequency resolution of autoterms. This paper proposes a novel sliding time and frequency windowing-based technique for removing false-terms from WVD, which aims to overcome the limitations of kernel-based methods. The proposed method segments a multi-component signal using overlapping windows in time and frequency domains successively and the WVD of each windowed signal is computed. The WVDs of all windowed signals are added together to obtain the falseterm free WVD. Energy scaling is also applied to minimize the effect of overlapping windows. Performance of the proposed method is evaluated for different multi-component synthetic signals and a natural ECG signal using various performance measures. The simulation results demonstrate that the proposed method can effectively remove false-terms from the WVD with improved auto-term enhancement and time-frequency resolution. Results from the proposed method are also compared with different kernel and window-based smoothing methods to show its superiority over these methods.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"85 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132779037","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136073
Sarvat Ali, Shital A. Raut
High blood glucose levels cause lesions on the retina of the eye, resulting in a degenerative condition known as diabetic retinopathy (DR), which impacts vision and can cause irreversible vision loss. The most common cause of blindness in diabetic people is thought to be diabetic retinopathy. Early diagnosis of diabetic retinopathy is essential to efficiently maintaining the patient’s vision. We attempted to give first-hand verification to this fundamental problem of DR detection to save time, money and efforts of ophthalmologists. The latter also proved to be more challenging, especially early on in the disease, when disease characteristics are less obvious in the fundus images. Deep learning algorithms and machine learning-based medical image analysis have aided in the early identification of diabetic retinopathy along with the evaluation of retinal fundus images. This paper attempts to preprocess and binary classify fundus images from the famous Aptos dataset using finetuned ResNet50 as well as features extraction from ResNet50 and later classifying using machine learning models. We have achieved an accuracy of 0.9802, an AUC score of 0.9937, F1 score of 0.9870, a precision of 0.9890, a recall as 0.9845 and kappa score of 0.9481 on the evaluation data by fine-tuning of ResNet50.
{"title":"Detection of Diabetic Retinopathy from fundus images using Resnet50","authors":"Sarvat Ali, Shital A. Raut","doi":"10.1109/PCEMS58491.2023.10136073","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136073","url":null,"abstract":"High blood glucose levels cause lesions on the retina of the eye, resulting in a degenerative condition known as diabetic retinopathy (DR), which impacts vision and can cause irreversible vision loss. The most common cause of blindness in diabetic people is thought to be diabetic retinopathy. Early diagnosis of diabetic retinopathy is essential to efficiently maintaining the patient’s vision. We attempted to give first-hand verification to this fundamental problem of DR detection to save time, money and efforts of ophthalmologists. The latter also proved to be more challenging, especially early on in the disease, when disease characteristics are less obvious in the fundus images. Deep learning algorithms and machine learning-based medical image analysis have aided in the early identification of diabetic retinopathy along with the evaluation of retinal fundus images. This paper attempts to preprocess and binary classify fundus images from the famous Aptos dataset using finetuned ResNet50 as well as features extraction from ResNet50 and later classifying using machine learning models. We have achieved an accuracy of 0.9802, an AUC score of 0.9937, F1 score of 0.9870, a precision of 0.9890, a recall as 0.9845 and kappa score of 0.9481 on the evaluation data by fine-tuning of ResNet50.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133878679","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 idea behind a Denial of Service(DoS) attack is to overload or flood the system or the network with systems that the system becomes incapacitated. A Distributed Denial of Service(DDoS) attack is a similar attack with multiple systems attacking one victim. In this paper we discuss the methods to detect these attacks in a working system using mathematical and entropy based techniques. The proposed mathematical model uses both the mean and standard deviation as thresholds for classification as they work better when the data is unsymmetrical like a real working system’s network data. The proposed entropy model uses a combination of Shannon’s entropy and the mathematical threshold. This model takes care of the anomalous non-attack cases like a ping to a blocked IP address or rejected packets.
{"title":"DoS and DDoS attack detection using Mathematical and Entropy Methods","authors":"Sumedha Janani Siriyapuraju, V. Gowri, Srilikhita Balla, Mukesh Kumar Vanika, Abhay Gandhi","doi":"10.1109/PCEMS58491.2023.10136042","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136042","url":null,"abstract":"The idea behind a Denial of Service(DoS) attack is to overload or flood the system or the network with systems that the system becomes incapacitated. A Distributed Denial of Service(DDoS) attack is a similar attack with multiple systems attacking one victim. In this paper we discuss the methods to detect these attacks in a working system using mathematical and entropy based techniques. The proposed mathematical model uses both the mean and standard deviation as thresholds for classification as they work better when the data is unsymmetrical like a real working system’s network data. The proposed entropy model uses a combination of Shannon’s entropy and the mathematical threshold. This model takes care of the anomalous non-attack cases like a ping to a blocked IP address or rejected packets.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126534415","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：