Pub Date : 2022-12-01DOI: 10.1109/OCIT56763.2022.00030
Md. Abu Sayeed, Fatahi Nasrin, S. Mohanty, E. Kougianos
Epilepsy is a neurological disorder marked by recurrent seizures. At least 3 million Americans and 1% of the global population have epilepsy, requiring a low-latency seizure detection system necessary for effective epilepsy treatment. In this paper, a pulse exclusion mechanism (PEM) based novel seizure detection system has been presented in the internet of medical things (IoMT), which uses a PEM to eliminate unnecessary features or channels and allocate desired pulses in a time frame. An optimized deep neural network (DNN) algorithm is used for feature classification. The proposed approach has been evaluated using CHB-MIT Scalp database. The results of the experiments indicate that the proposed eSeiz 2.0 offers a high specificity of 100% and a low latency of 1.05 sec, which can be useful for wearable biomedical applications as well as real-world epilepsy treatment.
{"title":"eSeiz 2.0: An IoMT Framework for Accurate Low-Latency Seizure Detection using Pulse Exclusion Mechanism","authors":"Md. Abu Sayeed, Fatahi Nasrin, S. Mohanty, E. Kougianos","doi":"10.1109/OCIT56763.2022.00030","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00030","url":null,"abstract":"Epilepsy is a neurological disorder marked by recurrent seizures. At least 3 million Americans and 1% of the global population have epilepsy, requiring a low-latency seizure detection system necessary for effective epilepsy treatment. In this paper, a pulse exclusion mechanism (PEM) based novel seizure detection system has been presented in the internet of medical things (IoMT), which uses a PEM to eliminate unnecessary features or channels and allocate desired pulses in a time frame. An optimized deep neural network (DNN) algorithm is used for feature classification. The proposed approach has been evaluated using CHB-MIT Scalp database. The results of the experiments indicate that the proposed eSeiz 2.0 offers a high specificity of 100% and a low latency of 1.05 sec, which can be useful for wearable biomedical applications as well as real-world epilepsy treatment.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123683563","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00036
Payel Patra, Tripty Singh
Diabetic Retinopathy (DR) could be a mortal eye ailment that happens in people who have the disease named diabetics which hurts mainly on retina and after a long duration, it may lead to visual lacking. Diabetic Retinopathy Detection (DRD) through the integration of state of the art Profound Proficiency styles. This research used dataset, which was obtained from Eye Foundation Hospital Bangalore and Narayana Netralaya Bangalore, In this paper authors designed the frameworks within the field of profound Convolutional Neural Networks (CNNs), which have demonstrated progressive changes in numerous areas of computer vision counting therapeutic imaging, and researchers bring their control to the conclusion of eye fundus images. This proposed outline is combination of three stages. To begin with, the fundus picture is pre-processed utilizing an intensity of normalised procedure and augmented method. 2nd, the pre-processed picture is input to distinctive foundations of the CNN architecture in arrange to extricate a point vector for the evaluating process. 3rd, a classification is utilized for DRD and decides its review (e.g., no DR, mild, severe, moderate, or Proliferative Diabetic Retinopa-thy). A trained model with Resnet50, Inception V3, VGG-19, DenseNet-121 and MobileNetV2 architectures will extricate the Indus images of the eye. The outcome is coming with amazing exactness of 93.79 percentile, which is better by 7% than earlier work, by utilizing several activation functions in the new DiabRetNet architecture.
{"title":"Diabetic Retinopathy Detection using an Improved ResNet 50-InceptionV3 and hybrid DiabRetNet Structures","authors":"Payel Patra, Tripty Singh","doi":"10.1109/OCIT56763.2022.00036","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00036","url":null,"abstract":"Diabetic Retinopathy (DR) could be a mortal eye ailment that happens in people who have the disease named diabetics which hurts mainly on retina and after a long duration, it may lead to visual lacking. Diabetic Retinopathy Detection (DRD) through the integration of state of the art Profound Proficiency styles. This research used dataset, which was obtained from Eye Foundation Hospital Bangalore and Narayana Netralaya Bangalore, In this paper authors designed the frameworks within the field of profound Convolutional Neural Networks (CNNs), which have demonstrated progressive changes in numerous areas of computer vision counting therapeutic imaging, and researchers bring their control to the conclusion of eye fundus images. This proposed outline is combination of three stages. To begin with, the fundus picture is pre-processed utilizing an intensity of normalised procedure and augmented method. 2nd, the pre-processed picture is input to distinctive foundations of the CNN architecture in arrange to extricate a point vector for the evaluating process. 3rd, a classification is utilized for DRD and decides its review (e.g., no DR, mild, severe, moderate, or Proliferative Diabetic Retinopa-thy). A trained model with Resnet50, Inception V3, VGG-19, DenseNet-121 and MobileNetV2 architectures will extricate the Indus images of the eye. The outcome is coming with amazing exactness of 93.79 percentile, which is better by 7% than earlier work, by utilizing several activation functions in the new DiabRetNet architecture.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123922393","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00057
M. Sayeed, Deepa Gupta
Global universities are establishing institutional setups that offer a hybrid format of education. The next step of education is to maintain quality and flexibility, such as providing the option to convert online courses such as Massive Open Online Courses (MOOCS) to course credits. However, several universities are reluctant to completely transition to online-based education due to poor digital experience in educational tools. The available evaluation tools such as Multiple-choice answers (MCQ) aren't able to evaluate students holistically. In this study, research work aims for an improvised reference-based approach (utilizing student and reference answers) that evaluates descriptive answers with the Siamese architecture- Roberta bi-encoder based transformer models for Automated Short Answer Grading (ASAG). The architecture was designed considering ASAG tasks constrained to feasible compute resources. The research work presents the competitive performance of the models, further improvised with finetuning and hyperparameter optimization process on the benchmark SemEval-2013 2way task dataset.
{"title":"Automate Descriptive Answer Grading using Reference based Models","authors":"M. Sayeed, Deepa Gupta","doi":"10.1109/OCIT56763.2022.00057","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00057","url":null,"abstract":"Global universities are establishing institutional setups that offer a hybrid format of education. The next step of education is to maintain quality and flexibility, such as providing the option to convert online courses such as Massive Open Online Courses (MOOCS) to course credits. However, several universities are reluctant to completely transition to online-based education due to poor digital experience in educational tools. The available evaluation tools such as Multiple-choice answers (MCQ) aren't able to evaluate students holistically. In this study, research work aims for an improvised reference-based approach (utilizing student and reference answers) that evaluates descriptive answers with the Siamese architecture- Roberta bi-encoder based transformer models for Automated Short Answer Grading (ASAG). The architecture was designed considering ASAG tasks constrained to feasible compute resources. The research work presents the competitive performance of the models, further improvised with finetuning and hyperparameter optimization process on the benchmark SemEval-2013 2way task dataset.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127175499","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00112
Isha Padhy, P. Kanungo, S. Sahoo
A shadow in an image can disturb the actual outcome in computer vision and pattern recognition applications. The reason is that the shadow will act as an individual object resulting in the false interpretation and performance degradation of subsequent computer vision tasks. Here we propose a process to detect and remove shadows from an image using the YCbCr colour model. A small portion of the image is identified as a shadow area. The features at the pixel level and along the boundaries in the shadow area are learned. A method based on the locations of the border of the shadow is applied to remove the shadow. Experiments have been conducted on the benchmark camouflaged image dataset and the non-camouflaged image dataset to evaluate the approach. The methodology achieves promising performance in detecting and removing shadows from an image.
{"title":"A YCbCr Model Based Shadow Detection and Removal Approach On Camouflaged Images","authors":"Isha Padhy, P. Kanungo, S. Sahoo","doi":"10.1109/OCIT56763.2022.00112","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00112","url":null,"abstract":"A shadow in an image can disturb the actual outcome in computer vision and pattern recognition applications. The reason is that the shadow will act as an individual object resulting in the false interpretation and performance degradation of subsequent computer vision tasks. Here we propose a process to detect and remove shadows from an image using the YCbCr colour model. A small portion of the image is identified as a shadow area. The features at the pixel level and along the boundaries in the shadow area are learned. A method based on the locations of the border of the shadow is applied to remove the shadow. Experiments have been conducted on the benchmark camouflaged image dataset and the non-camouflaged image dataset to evaluate the approach. The methodology achieves promising performance in detecting and removing shadows from an image.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"88 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130670133","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00025
Binon Teji, Swarup Roy
Graph representation learning recently has proven their excellent competency in understanding large graphs and their inner engineering for various downstream tasks. Link completion is an important computational task to guess missing edges in a network. The traditional methods extract local, pairwise information based on specific proximity statistics that are always ineffective in inferring missing links from a global topological perspective. Graph Convolutional Network (GCN) based em-bedding methods may be an effective alternative. In this work, we try to experimentally assess the power of GCN-based graph embedding techniques, namely Graph Auto Encoder (GAE) and its variants GraphSAGE, and Graph Attention Network (GAT) for link prediction tasks. Experimental results show that the GAE-based encoding methods are able to achieve superior results for predicting missing links in various real large-scale networks in comparison to traditional link prediction methods. Interestingly, our results reveal that the above techniques successfully recreate the original network with high true positive and negative rates. However, it has been observed that they produce many extra edges with an overall very high false positive rate.
{"title":"Missing Link Identification from Node Embeddings using Graph Auto Encoders and its Variants","authors":"Binon Teji, Swarup Roy","doi":"10.1109/OCIT56763.2022.00025","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00025","url":null,"abstract":"Graph representation learning recently has proven their excellent competency in understanding large graphs and their inner engineering for various downstream tasks. Link completion is an important computational task to guess missing edges in a network. The traditional methods extract local, pairwise information based on specific proximity statistics that are always ineffective in inferring missing links from a global topological perspective. Graph Convolutional Network (GCN) based em-bedding methods may be an effective alternative. In this work, we try to experimentally assess the power of GCN-based graph embedding techniques, namely Graph Auto Encoder (GAE) and its variants GraphSAGE, and Graph Attention Network (GAT) for link prediction tasks. Experimental results show that the GAE-based encoding methods are able to achieve superior results for predicting missing links in various real large-scale networks in comparison to traditional link prediction methods. Interestingly, our results reveal that the above techniques successfully recreate the original network with high true positive and negative rates. However, it has been observed that they produce many extra edges with an overall very high false positive rate.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126877838","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}
Consumption of alcohol among students, mainly college or university students, has risen immensely over the past couple of years. It has been determined that students experiment with alcohol during their college years and around 80% of students consume alcohol in some manner or degree and 50% are involved in binge drinking. This is mainly due to students wanting to explore their newfound independence and freedom which they didn't have during their school years. In this paper, we have analyzed students belonging to two courses of a Secondary School-Maths and Portuguese Language Course. We have applied Feature Scaling along with various machine learning classification models to determine higher alcohol consumption where the Random Forest Model outperformed all other models that have been applied such as Linear, Ridge, and Lasso Regression, Decision Tree, k-NN, XG Boost, Support Vector Machine, ADA Boosting Regressor and Gradient Boosting Regressor for analysis of alcohol consumption among secondary school students.
{"title":"Alcohol Consumption Rate Prediction using Machine Learning Algorithms","authors":"Advait Singh, Vinay Singh, Mahendra Kumar Gourisaria, Ashish Sharma","doi":"10.1109/OCIT56763.2022.00026","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00026","url":null,"abstract":"Consumption of alcohol among students, mainly college or university students, has risen immensely over the past couple of years. It has been determined that students experiment with alcohol during their college years and around 80% of students consume alcohol in some manner or degree and 50% are involved in binge drinking. This is mainly due to students wanting to explore their newfound independence and freedom which they didn't have during their school years. In this paper, we have analyzed students belonging to two courses of a Secondary School-Maths and Portuguese Language Course. We have applied Feature Scaling along with various machine learning classification models to determine higher alcohol consumption where the Random Forest Model outperformed all other models that have been applied such as Linear, Ridge, and Lasso Regression, Decision Tree, k-NN, XG Boost, Support Vector Machine, ADA Boosting Regressor and Gradient Boosting Regressor for analysis of alcohol consumption among secondary school students.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115785050","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00080
Anusha Vaishnav, Amulya Ratna Swain, M. R. Lenka
As the world is moving forward to the Fifth Generation (5G) of wireless technology, the demand for efficient communication techniques has also increased. 5G provides a far higher level of performance than previous generations of wireless communication in terms of low latency, increased throughput, and increased spectral efficiency. In 5G, some companion technologies have been added to strengthen the communication efficiency among the users. Device-to-Device(D2D) communication is one of these technologies to be used for modern cellular networks like 5G. D2D technology allows devices to communicate with each other without the assistance of a base station. The primary benefits of D2D communication include increased spectrum, energy efficiency, reduced transmission delay, and improved system throughput. Along with these benefits, several technical challenges include device discovery, resource allocation, mode selection, interference management, privacy, and security. In this paper, we discuss one of the challenges and the primary aspect of D2D communication, i.e., Device Discovery. The device discovery process starts when the devices transmit a discovery signal to an intermediate device to enhance the communication process by connecting with that device. Finding a potential intermediate device that will not disrupt the communication channel can sometimes become challenging. The device discovery process cannot be overlooked as it is an important step that is required before the establishment of D2D communication as well as during the communication process. In other words, device discovery is one of the key building blocks of D2D-based networks. This paper thoroughly reviews most of the important device discovery techniques for D2D communication.
{"title":"Device Discovery Approaches in D2D Communication: A Survey","authors":"Anusha Vaishnav, Amulya Ratna Swain, M. R. Lenka","doi":"10.1109/OCIT56763.2022.00080","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00080","url":null,"abstract":"As the world is moving forward to the Fifth Generation (5G) of wireless technology, the demand for efficient communication techniques has also increased. 5G provides a far higher level of performance than previous generations of wireless communication in terms of low latency, increased throughput, and increased spectral efficiency. In 5G, some companion technologies have been added to strengthen the communication efficiency among the users. Device-to-Device(D2D) communication is one of these technologies to be used for modern cellular networks like 5G. D2D technology allows devices to communicate with each other without the assistance of a base station. The primary benefits of D2D communication include increased spectrum, energy efficiency, reduced transmission delay, and improved system throughput. Along with these benefits, several technical challenges include device discovery, resource allocation, mode selection, interference management, privacy, and security. In this paper, we discuss one of the challenges and the primary aspect of D2D communication, i.e., Device Discovery. The device discovery process starts when the devices transmit a discovery signal to an intermediate device to enhance the communication process by connecting with that device. Finding a potential intermediate device that will not disrupt the communication channel can sometimes become challenging. The device discovery process cannot be overlooked as it is an important step that is required before the establishment of D2D communication as well as during the communication process. In other words, device discovery is one of the key building blocks of D2D-based networks. This paper thoroughly reviews most of the important device discovery techniques for D2D communication.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131753912","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00034
S. Nayak, Satchidananda Dehuri, Sung-Bae Cho
Time series data are correlated in a nonlinear fashion which makes the future data prediction challenging. Particularly, the correlation among data at the fluctuation points is insignificant and it is hard to capture the underlaying nonlinearity at those points by conventional prediction systems. The accuracy of time series forecasting (TSF) is vastly influenced by the current and immediate past data rather by far away data points. This article proposes an extreme learning-based method for exploration of virtual data positions (ELMVDP) from the training data and incorporates them to the original time series to intensify the TSF accuracy of a single hidden layer neural network. Specifically, this method is useful for the time series having less volume of data which may not suffice to train a TSF model. The effectiveness of ELMVDP method is evaluated on time series available in the literature, compared with few similar deterministic and stochastic approaches, and observations from simulation studies show that ELMVDP method yields better predictions than others.
{"title":"ELMVDP: extreme learning based virtual data position exploration and incorporation method for escalation of time series forecasting accuracy","authors":"S. Nayak, Satchidananda Dehuri, Sung-Bae Cho","doi":"10.1109/OCIT56763.2022.00034","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00034","url":null,"abstract":"Time series data are correlated in a nonlinear fashion which makes the future data prediction challenging. Particularly, the correlation among data at the fluctuation points is insignificant and it is hard to capture the underlaying nonlinearity at those points by conventional prediction systems. The accuracy of time series forecasting (TSF) is vastly influenced by the current and immediate past data rather by far away data points. This article proposes an extreme learning-based method for exploration of virtual data positions (ELMVDP) from the training data and incorporates them to the original time series to intensify the TSF accuracy of a single hidden layer neural network. Specifically, this method is useful for the time series having less volume of data which may not suffice to train a TSF model. The effectiveness of ELMVDP method is evaluated on time series available in the literature, compared with few similar deterministic and stochastic approaches, and observations from simulation studies show that ELMVDP method yields better predictions than others.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"256 3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116876435","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00042
Ch. Sanjeev Kumar Dash, A. K. Behera, S. Nayak, Satchidananda Dehuri, J. P. Mohanty
Economic activities have deteriorated the quality of air, which is a vital natural resource. There has been a lot of research on predicting when terrible air quality will occur, but much of it is limited by a lack of data collected, making it unable to account for periodic and other factors. This article develops and analyses the performances of two higher order neural networks-based forecasts such as pi-sigma neural network (PSNN) and functional link artificial neural network (FLANN) on estimating the air quality index (AQI) of Brarajanagar and Talcher industrial region of Odisha State, India. AQIs at the daily level of two cities are collected from the Kaggle source, preprocessed, and used for modeling and forecasting by the two higher-order neural networks. Simulation outcomes and comparative studies are in favor of PSNN and FLANN-based forecasting
{"title":"Estimation of Air Quality Index of Brajarajnagar and Talcher Industrial Region of Odisha State: A Higher Order Neural Network Approach","authors":"Ch. Sanjeev Kumar Dash, A. K. Behera, S. Nayak, Satchidananda Dehuri, J. P. Mohanty","doi":"10.1109/OCIT56763.2022.00042","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00042","url":null,"abstract":"Economic activities have deteriorated the quality of air, which is a vital natural resource. There has been a lot of research on predicting when terrible air quality will occur, but much of it is limited by a lack of data collected, making it unable to account for periodic and other factors. This article develops and analyses the performances of two higher order neural networks-based forecasts such as pi-sigma neural network (PSNN) and functional link artificial neural network (FLANN) on estimating the air quality index (AQI) of Brarajanagar and Talcher industrial region of Odisha State, India. AQIs at the daily level of two cities are collected from the Kaggle source, preprocessed, and used for modeling and forecasting by the two higher-order neural networks. Simulation outcomes and comparative studies are in favor of PSNN and FLANN-based forecasting","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"141 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116046344","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 : 2022-12-01DOI: 10.1109/OCIT56763.2022.00060
Prachi Vijayeeta, Parthasarathi Pattnayak
Deep Learning mechanisms can be leveraged for playing the type of music based on the emotions of an individual entity. This can be done by detecting the human facial expressions, color, posture, orientation, lightning, etc. An interface is designed which makes the system to analyze the possible variability of faces. The basic pre-requisite for emotion recognition is appropriate selection of facial features that helps in identifying the mood of a person. Traditionally, grouping songs into various playlist was manual interpreted that consumed lot of time and it was indeed a tedious task. However, the advent of Facial Expression Based Music System emphasizes an automatic creation of music playlist based on real time mental state of an individual. In this work we have employed Haar Cascade-CNN classifier and SVM classifier to detect the emotions in an image. Haar Cascade is a machine learning-based approach where a lot of positive and negative images are used to train the classifier. The learning algorithm keeps on training the input feature vector based on the image captured. The gray scale image of the face is used by the system to classify five basic emotions such as surprise, disgust, neutral, anger and happiness. The emotion classification is achieved by observing the parts of the face, like eyes, lips movement, etc. A comparative study of these two classifiers are conducted based on the trained datasets. This electronic document is a “live” template and already defines the components of your paper [title, text, heads, etc.] in its style sheet.
{"title":"A Deep Learning approach for Emotion Based Music Player","authors":"Prachi Vijayeeta, Parthasarathi Pattnayak","doi":"10.1109/OCIT56763.2022.00060","DOIUrl":"https://doi.org/10.1109/OCIT56763.2022.00060","url":null,"abstract":"Deep Learning mechanisms can be leveraged for playing the type of music based on the emotions of an individual entity. This can be done by detecting the human facial expressions, color, posture, orientation, lightning, etc. An interface is designed which makes the system to analyze the possible variability of faces. The basic pre-requisite for emotion recognition is appropriate selection of facial features that helps in identifying the mood of a person. Traditionally, grouping songs into various playlist was manual interpreted that consumed lot of time and it was indeed a tedious task. However, the advent of Facial Expression Based Music System emphasizes an automatic creation of music playlist based on real time mental state of an individual. In this work we have employed Haar Cascade-CNN classifier and SVM classifier to detect the emotions in an image. Haar Cascade is a machine learning-based approach where a lot of positive and negative images are used to train the classifier. The learning algorithm keeps on training the input feature vector based on the image captured. The gray scale image of the face is used by the system to classify five basic emotions such as surprise, disgust, neutral, anger and happiness. The emotion classification is achieved by observing the parts of the face, like eyes, lips movement, etc. A comparative study of these two classifiers are conducted based on the trained datasets. This electronic document is a “live” template and already defines the components of your paper [title, text, heads, etc.] in its style sheet.","PeriodicalId":425541,"journal":{"name":"2022 OITS International Conference on Information Technology (OCIT)","volume":"373 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124673252","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：