Pub Date : 2023-04-05DOI: 10.1109/PCEMS58491.2023.10136115
D. Agrawal, Harshvardhan Dave, Abhishek P. Shete, Spandana Pulimamidi, Snigdha Bhagat, Punitkumar Bhavsar
Sign language recognition through image processing presents challenges related to the requirement of real time applicability and high accuracy. Though previous work adopting methodologies from deep convolutional neural network architectures have shown to achieve good performance, they lack a consummate solution in terms of accuracy due to consideration of word based recognition. Recent development of Inception Network based architectures have shown promising classification accuracy with relatively less computational demand. Hence in this paper we propose a methodology that adopts Inception Network for the task of Sign Language Recognition. We considered the American Sign Language Recognition and Correction model. The correction and suggestion tools are implemented in the model to rectify any incorrect sign detection. The results from our approach achieves accuracy in the order of 99 percent.
{"title":"Improved American Sign Language Recognition and Correction Using Inception Network, MediaPipe and PyEnchant","authors":"D. Agrawal, Harshvardhan Dave, Abhishek P. Shete, Spandana Pulimamidi, Snigdha Bhagat, Punitkumar Bhavsar","doi":"10.1109/PCEMS58491.2023.10136115","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136115","url":null,"abstract":"Sign language recognition through image processing presents challenges related to the requirement of real time applicability and high accuracy. Though previous work adopting methodologies from deep convolutional neural network architectures have shown to achieve good performance, they lack a consummate solution in terms of accuracy due to consideration of word based recognition. Recent development of Inception Network based architectures have shown promising classification accuracy with relatively less computational demand. Hence in this paper we propose a methodology that adopts Inception Network for the task of Sign Language Recognition. We considered the American Sign Language Recognition and Correction model. The correction and suggestion tools are implemented in the model to rectify any incorrect sign detection. The results from our approach achieves accuracy in the order of 99 percent.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"20 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":"126488257","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/PCEMS58491.2023.10136039
Alekya Nyalapelli, Shubham Sharma, Pranjal Phadnis, Maithili Patil, A. Tandle
As the fifth generation (5G) of wireless communication rolls out worldwide, conceptualized use cases and disruptive industry solutions are being deployed to offer smooth, frictionless, and secure connectivity. The landscape of Artificial Intelligence (AI) and Machine Learning (ML) can be seen as potential drivers in the automation and optimization of network performances and management complexities. The shifting network behaviors and complicated modern applications present diverse network performance traffic, which can be exploited by service providers to deal with network demands and provide superior user experiences. The existing research can be divided into the following 5G research areas, which include network traffic, resource allocation, network slicing, mobility management, physical layer security, etc., to name a few. The primary objective of this paper is to provide a comprehensive perspective on the expanding diversity of viable ML-assisted solutions for tackling various 5G network-level issues. The paper concludes with an indepth investigation of the challenges and unexplored directions of future research pertaining to making 5G applications more reliable for future use cases.
{"title":"Recent Advancements in Applications of Artificial Intelligence and Machine Learning for 5G Technology: A Review","authors":"Alekya Nyalapelli, Shubham Sharma, Pranjal Phadnis, Maithili Patil, A. Tandle","doi":"10.1109/PCEMS58491.2023.10136039","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136039","url":null,"abstract":"As the fifth generation (5G) of wireless communication rolls out worldwide, conceptualized use cases and disruptive industry solutions are being deployed to offer smooth, frictionless, and secure connectivity. The landscape of Artificial Intelligence (AI) and Machine Learning (ML) can be seen as potential drivers in the automation and optimization of network performances and management complexities. The shifting network behaviors and complicated modern applications present diverse network performance traffic, which can be exploited by service providers to deal with network demands and provide superior user experiences. The existing research can be divided into the following 5G research areas, which include network traffic, resource allocation, network slicing, mobility management, physical layer security, etc., to name a few. The primary objective of this paper is to provide a comprehensive perspective on the expanding diversity of viable ML-assisted solutions for tackling various 5G network-level issues. The paper concludes with an indepth investigation of the challenges and unexplored directions of future research pertaining to making 5G applications more reliable for future use cases.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"11 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":"129885007","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/PCEMS58491.2023.10136050
Sanket Soni, A. Khurshid, Anushree Mrugank Minase, Ashlesha Bonkinpelliwar
Water quality prediction is a crucial process before any consumption of water. Prediction and modeling methods are used for pollutants in water to deal with water pollution control. This work involves the use of a random forest learning algorithm to quantitate BOD and COD using parameter tuning to establish the importance of input variables. It uses minimal sensed quantitative parameters such as Temperature, pH, DO, and Conductivity along with categorical parameters. The trained model shows excellent efficiency compared to other models and is validated using the laboratory test results with a maximum error of 10%. It is computationally low-cost, requires minimal parameters, and is pruned to integrate and implement in an IoT hardware system, reducing the cost of expensive sensors.
{"title":"A TinyML Approach for Quantification of BOD and COD in Water","authors":"Sanket Soni, A. Khurshid, Anushree Mrugank Minase, Ashlesha Bonkinpelliwar","doi":"10.1109/PCEMS58491.2023.10136050","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136050","url":null,"abstract":"Water quality prediction is a crucial process before any consumption of water. Prediction and modeling methods are used for pollutants in water to deal with water pollution control. This work involves the use of a random forest learning algorithm to quantitate BOD and COD using parameter tuning to establish the importance of input variables. It uses minimal sensed quantitative parameters such as Temperature, pH, DO, and Conductivity along with categorical parameters. The trained model shows excellent efficiency compared to other models and is validated using the laboratory test results with a maximum error of 10%. It is computationally low-cost, requires minimal parameters, and is pruned to integrate and implement in an IoT hardware system, reducing the cost of expensive sensors.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"2010 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":"125631361","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/PCEMS58491.2023.10136103
Praveen Pawar, Naveen Pawar, Aditya Trivedi
The benefits of 5G millimetre wave (mmWave) technology include high data throughput, very low latency, stable connectivity, and massive capacity. Furthermore, unmanned aerial vehicles (UAVs) are being used as flying base stations to provide a viable option for reliable and cost-effective wireless communication. The goal of this research is to integrate a UAVassisted wireless network with a 5G mmWave communication system and assess coverage performance. The signal between the UAV and the user is heavily distorted by tropical atmospheric parameters such as rain, trees, and fog. This paper investigated the effect of attenuation on user coverage performance in the mmWave frequency bands of 28 GHz and 60 GHz (recommended by the International Telecommunication Union (ITU) for use in 5G cellular services). First, various factors such as rain rate, frequency, and foliage depth are used to calculate the attenuation of rain and foliage. Next, the suggested system model is used to analyse the likelihood of ground users’ downlink coverage in terms of total attenuation, channel gain, channel noise, and interference. However, the impacts of rain and foliage attenuation in various tropical locales cannot be avoided, even given the small cell size of the system.
{"title":"How Atmospheric Attenuation affects the UAV Communication Network?","authors":"Praveen Pawar, Naveen Pawar, Aditya Trivedi","doi":"10.1109/PCEMS58491.2023.10136103","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136103","url":null,"abstract":"The benefits of 5G millimetre wave (mmWave) technology include high data throughput, very low latency, stable connectivity, and massive capacity. Furthermore, unmanned aerial vehicles (UAVs) are being used as flying base stations to provide a viable option for reliable and cost-effective wireless communication. The goal of this research is to integrate a UAVassisted wireless network with a 5G mmWave communication system and assess coverage performance. The signal between the UAV and the user is heavily distorted by tropical atmospheric parameters such as rain, trees, and fog. This paper investigated the effect of attenuation on user coverage performance in the mmWave frequency bands of 28 GHz and 60 GHz (recommended by the International Telecommunication Union (ITU) for use in 5G cellular services). First, various factors such as rain rate, frequency, and foliage depth are used to calculate the attenuation of rain and foliage. Next, the suggested system model is used to analyse the likelihood of ground users’ downlink coverage in terms of total attenuation, channel gain, channel noise, and interference. However, the impacts of rain and foliage attenuation in various tropical locales cannot be avoided, even given the small cell size of the system.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"528 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":"129356474","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/PCEMS58491.2023.10136079
A. Sreekumar, R. Lekshmi
The battery storage system is an inevitable component in an electric vehicle. A precise state of charge is critically notable to evaluate the level of battery aging and ensure electric vehicle reliability and security. Recently, data driven methods have procured much popularity in many research fields. Data-driven methods are found to be a promising approach to provide a high accuracy solution to battery state of charge estimation problem. Research works focus on the state of charge estimation under a fixed operating temperature. Apparently, the maximum deliverable charge of a battery degrades with charge-discharge cycle and temperature. Thus, it is important to consider the operating temperature as one of the input features. This paper identifies the best data driven model for state of charge estimation among linear regression, random forest, CatBoost and XGBoost models. The data driven models are developed via training, validating, and testing stages deploying Li-ion (LG 18650HG2) battery data set under -10°C, 0°C, 10°C and 25°C. The best model is identified using performance indices. The results show the superior performance of XGBoost model under all temperatures and best performance at 25°C with mean absolute error of 0.68%, mean square error of 0.01% and root mean square error of 1.10%, mean absolute percentage error of 1.78%, and R2 value of 99.86%, compared to linear regression, random forest, CatBoost models.
{"title":"Comparative Study of Data Driven Methods for State of Charge Estimation of Li-ion Battery","authors":"A. Sreekumar, R. Lekshmi","doi":"10.1109/PCEMS58491.2023.10136079","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136079","url":null,"abstract":"The battery storage system is an inevitable component in an electric vehicle. A precise state of charge is critically notable to evaluate the level of battery aging and ensure electric vehicle reliability and security. Recently, data driven methods have procured much popularity in many research fields. Data-driven methods are found to be a promising approach to provide a high accuracy solution to battery state of charge estimation problem. Research works focus on the state of charge estimation under a fixed operating temperature. Apparently, the maximum deliverable charge of a battery degrades with charge-discharge cycle and temperature. Thus, it is important to consider the operating temperature as one of the input features. This paper identifies the best data driven model for state of charge estimation among linear regression, random forest, CatBoost and XGBoost models. The data driven models are developed via training, validating, and testing stages deploying Li-ion (LG 18650HG2) battery data set under -10°C, 0°C, 10°C and 25°C. The best model is identified using performance indices. The results show the superior performance of XGBoost model under all temperatures and best performance at 25°C with mean absolute error of 0.68%, mean square error of 0.01% and root mean square error of 1.10%, mean absolute percentage error of 1.78%, and R2 value of 99.86%, compared to linear regression, random forest, CatBoost models.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"23 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":"127801823","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/PCEMS58491.2023.10136075
P. Patel, Ajay Shanmukh Goteti
Visible light communication (VLC) is emerging as a solution for wireless communication systems to overcome the crowded radio spectrum. VLC makes use of a large, uncontrolled free spectrum. A light emitting diode is utilized as the transmitter, while Avalanche photodiodes or PIN photodiodes are employed as the receiver. In this paper, a basic visible light communication model is studied, then two indoor visible light communication models in which the sources are positioned on the roof’s ceiling and the receivers are placed five meters away from the sources towards the ground are designed. optiSystems software was used to design visible light communication models, which include a multiple input and multiple output (MIMO) model. The Line Of Sight (LOS) channel was used as a channel between the transmitter and receiver, along with on off keying (OOK) as a modulation technique. The system’s bit error rates (BER), quality factor, and eye diagram are evaluated at different data rates.
{"title":"Performance Analysis of MultiUser MIMO Indoor Visible Light Communication Systems","authors":"P. Patel, Ajay Shanmukh Goteti","doi":"10.1109/PCEMS58491.2023.10136075","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136075","url":null,"abstract":"Visible light communication (VLC) is emerging as a solution for wireless communication systems to overcome the crowded radio spectrum. VLC makes use of a large, uncontrolled free spectrum. A light emitting diode is utilized as the transmitter, while Avalanche photodiodes or PIN photodiodes are employed as the receiver. In this paper, a basic visible light communication model is studied, then two indoor visible light communication models in which the sources are positioned on the roof’s ceiling and the receivers are placed five meters away from the sources towards the ground are designed. optiSystems software was used to design visible light communication models, which include a multiple input and multiple output (MIMO) model. The Line Of Sight (LOS) channel was used as a channel between the transmitter and receiver, along with on off keying (OOK) as a modulation technique. The system’s bit error rates (BER), quality factor, and eye diagram are evaluated at different data rates.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"31 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":"127887803","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/PCEMS58491.2023.10136082
Aveg Ajay Ganorkar, Anurag Khandelwal, Meeti Khendelwal, P. Selokar
Mathematics is widely used in engineering and education domains. The ubiquitous use of smartphones makes it easy for a user to digitally look for a math solution. Manually typing a mathematical expression is a cumbersome process; many a time, even basic mathematical expressions with a superscript and subscript get difficult to interpret. As a result, this paper describes a cross-platform mobile scanner application integrated with optical character recognition (OCR) to extract the math expression from a handwritten or printed image, as well as the final solution to the problem. The mathematical equation scanner proposed is an efficient and time effective solution for solving polynomial and calculus equations.
{"title":"Cross Platform Mobile Application for solving Calculus","authors":"Aveg Ajay Ganorkar, Anurag Khandelwal, Meeti Khendelwal, P. Selokar","doi":"10.1109/PCEMS58491.2023.10136082","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136082","url":null,"abstract":"Mathematics is widely used in engineering and education domains. The ubiquitous use of smartphones makes it easy for a user to digitally look for a math solution. Manually typing a mathematical expression is a cumbersome process; many a time, even basic mathematical expressions with a superscript and subscript get difficult to interpret. As a result, this paper describes a cross-platform mobile scanner application integrated with optical character recognition (OCR) to extract the math expression from a handwritten or printed image, as well as the final solution to the problem. The mathematical equation scanner proposed is an efficient and time effective solution for solving polynomial and calculus equations.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"39 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":"132508301","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/PCEMS58491.2023.10136084
R. Tholkappian, S. Sinha, B. Chinni, N. Rao, V. Dogra
Photoacoustic Imaging(PAI) is an emerging soft tissue imaging system that can be potentially used for the detection of thyroid cancer. Computer-Aided diagnosis tools help further enhance the detection process by assisting the radiologist in the elucidation of medical data. This study aimed to classify the malignant and non-malignant thyroid tissue using different machine learning algorithms applied to the multi-wavelength PA data obtained, generated by the excised thyroid specimens from actual thyroid cancer patients. An exhaustive comparative analysis among the performances of three machine learning algorithms, random forest, support vector machine, and artificial neural network was performed for classifying benign vs malignant thyroid as well as non-malignant vs malignant thyroid. While the random forest algorithm efficiently classified benign vs malignant thyroid with the highest accuracy than the other two algorithms, the support vector machine outperformed the other two algorithms in classifying non-malignant vs malignant with the highest specificity, the area under the receiver operating characteristics, and accuracy. This study shows that multiwavelength PA data can be used with suitable machine algorithms for efficient thyroid cancer detection.
{"title":"Computer-aided tissue characterization for detection of thyroid cancer using multi-wavelength photoacoustic imaging.","authors":"R. Tholkappian, S. Sinha, B. Chinni, N. Rao, V. Dogra","doi":"10.1109/PCEMS58491.2023.10136084","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136084","url":null,"abstract":"Photoacoustic Imaging(PAI) is an emerging soft tissue imaging system that can be potentially used for the detection of thyroid cancer. Computer-Aided diagnosis tools help further enhance the detection process by assisting the radiologist in the elucidation of medical data. This study aimed to classify the malignant and non-malignant thyroid tissue using different machine learning algorithms applied to the multi-wavelength PA data obtained, generated by the excised thyroid specimens from actual thyroid cancer patients. An exhaustive comparative analysis among the performances of three machine learning algorithms, random forest, support vector machine, and artificial neural network was performed for classifying benign vs malignant thyroid as well as non-malignant vs malignant thyroid. While the random forest algorithm efficiently classified benign vs malignant thyroid with the highest accuracy than the other two algorithms, the support vector machine outperformed the other two algorithms in classifying non-malignant vs malignant with the highest specificity, the area under the receiver operating characteristics, and accuracy. This study shows that multiwavelength PA data can be used with suitable machine algorithms for efficient thyroid cancer detection.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"24 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":"131929873","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/PCEMS58491.2023.10136043
Reema Gera, Kalyan Ram Ambati, Pallavi G. Chakole, Naveen Cheggoju, V. Kamble, V. Satpute
Video surveillance plays an important role to analyze any anomaly activity in the given premises. However, cameras can only capture the video information but cannot determine the type of activity on its own. Therefore, such systems require regular human intervention and monitoring. This requires a lot of time and manual efforts. This calls for the need of automatic human activity recognition (HAR) system. This is possible using latest technologies like computer vision and deep learning based systems. Recognizing human activities in videos is a challenging task in computer vision. The main function of intelligent video systems is to automatically identify and tag the actions performed by people in video sequences accurately. The objective of this research is to develop a model that can accurately recognize and classify human activities from video footage. The information captured by the cameras i.e., videos can be used to determine the type of activity using deep learning based networks. Such a network should be capable of classifying the videos using the available spatial and temporal information. In this paper, a framework is proposed where the data is pre-processed initially to reject redundant information. This data is fed then into deep network to predict the event. In this paper, for HAR, two different network models are presented based on the size of the sequence of frames. One network takes in just the most significant frame and the other uses a longer sequence of frames for predicting the behavior as a time domain parameter.
{"title":"Classifying Human Activities using CNN and ConvLSTM in Video Sequences","authors":"Reema Gera, Kalyan Ram Ambati, Pallavi G. Chakole, Naveen Cheggoju, V. Kamble, V. Satpute","doi":"10.1109/PCEMS58491.2023.10136043","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136043","url":null,"abstract":"Video surveillance plays an important role to analyze any anomaly activity in the given premises. However, cameras can only capture the video information but cannot determine the type of activity on its own. Therefore, such systems require regular human intervention and monitoring. This requires a lot of time and manual efforts. This calls for the need of automatic human activity recognition (HAR) system. This is possible using latest technologies like computer vision and deep learning based systems. Recognizing human activities in videos is a challenging task in computer vision. The main function of intelligent video systems is to automatically identify and tag the actions performed by people in video sequences accurately. The objective of this research is to develop a model that can accurately recognize and classify human activities from video footage. The information captured by the cameras i.e., videos can be used to determine the type of activity using deep learning based networks. Such a network should be capable of classifying the videos using the available spatial and temporal information. In this paper, a framework is proposed where the data is pre-processed initially to reject redundant information. This data is fed then into deep network to predict the event. In this paper, for HAR, two different network models are presented based on the size of the sequence of frames. One network takes in just the most significant frame and the other uses a longer sequence of frames for predicting the behavior as a time domain parameter.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"25 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":"132628439","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}
This paper presents a non-intrusive video-based identification system with a single camera view placed at 90° angle to the subject. The aim is to efficiently extract biometric features from a low-resolution video (acquired from a common CCTV camera) for recognizing individuals. The proposed method uses gait cues and face profile features for identification of individuals. A fusion rule is applied to these feature sets to obtain a new set of attributes. Thus, three different recognition models are developed using the gait, face, and fused feature sets. An ensemble technique is defined over the three classification models based on these sets of cues to identify an individual. This approach is experimentally validated on the CASIA-B dataset that achieves 99.33% identification accuracy.
{"title":"Gait-Face Based Human Recognition From Distant Video","authors":"Gayatri Mudliar, Koushiki Nath, Yogesh Saini, Praveen Kumar","doi":"10.1109/PCEMS58491.2023.10136028","DOIUrl":"https://doi.org/10.1109/PCEMS58491.2023.10136028","url":null,"abstract":"This paper presents a non-intrusive video-based identification system with a single camera view placed at 90° angle to the subject. The aim is to efficiently extract biometric features from a low-resolution video (acquired from a common CCTV camera) for recognizing individuals. The proposed method uses gait cues and face profile features for identification of individuals. A fusion rule is applied to these feature sets to obtain a new set of attributes. Thus, three different recognition models are developed using the gait, face, and fused feature sets. An ensemble technique is defined over the three classification models based on these sets of cues to identify an individual. This approach is experimentally validated on the CASIA-B dataset that achieves 99.33% identification accuracy.","PeriodicalId":330870,"journal":{"name":"2023 2nd International Conference on Paradigm Shifts in Communications Embedded Systems, Machine Learning and Signal Processing (PCEMS)","volume":"26 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":"114301237","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：