Pub Date : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10054863
Hasibul Hasan Sabuj, Nigar Sultana Nuha, Paul Richie Gomes, Aiman Lameesa, Md. Ashraful Alam
Bangladesh’s garment industry is widely recognized and plays a significant role in the current global market. The nation’s per capita income and citizens’ living standards have risen significantly with the noteworthy hard work performed by the employees in this industry. The garment sector is more efficient once the target production can be achieved without any difficulties. But a frequent issue that comprises within this industry is, often the actual garment producing productivity of the people working there do not reach the previously determined target-productivity. The business suffers a significant loss when the productivity gap appears in this process. This approach seeks to address this issue by prediction of the actual productivity of the workers. To attain this goal, a machine learning approach is suggested for the productivity prediction of the employees, after experimentation with five machine learning models. The proposed approach displays a reassuring level of prediction accuracy, with a minimalist MAE (Mean Absolute Error) of 0.072, which is less than the existing Deep Learning model with a MAE of 0.086. This indicates that, application of this process can play a vital role in setting an accurate target production which might lead to more profit and production in the sector. Also, this work contains an explainable AI technique named SHAP for interpreting the model in order to see further information within it.
{"title":"Interpretable Garment Workers’ Productivity Prediction in Bangladesh Using Machine Learning Algorithms and Explainable AI","authors":"Hasibul Hasan Sabuj, Nigar Sultana Nuha, Paul Richie Gomes, Aiman Lameesa, Md. Ashraful Alam","doi":"10.1109/ICCIT57492.2022.10054863","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10054863","url":null,"abstract":"Bangladesh’s garment industry is widely recognized and plays a significant role in the current global market. The nation’s per capita income and citizens’ living standards have risen significantly with the noteworthy hard work performed by the employees in this industry. The garment sector is more efficient once the target production can be achieved without any difficulties. But a frequent issue that comprises within this industry is, often the actual garment producing productivity of the people working there do not reach the previously determined target-productivity. The business suffers a significant loss when the productivity gap appears in this process. This approach seeks to address this issue by prediction of the actual productivity of the workers. To attain this goal, a machine learning approach is suggested for the productivity prediction of the employees, after experimentation with five machine learning models. The proposed approach displays a reassuring level of prediction accuracy, with a minimalist MAE (Mean Absolute Error) of 0.072, which is less than the existing Deep Learning model with a MAE of 0.086. This indicates that, application of this process can play a vital role in setting an accurate target production which might lead to more profit and production in the sector. Also, this work contains an explainable AI technique named SHAP for interpreting the model in order to see further information within it.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"90 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116648934","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10055089
A. Hasib Uddin, Joygun Khatun, Mehera Afroz Meghna, Prince Mahmud
The automatic recognition of handwritten English material has seen a lot of progress. However, research on automatic Bangla handwriting numerals recognition is far behind. Even the most effective recognizers now in use do not produce an adequate performance for real-world applications. This paper suggested a strategy based on deep neural networks. In this paper, we have used the BanglaLekha-Isolated handwriting dataset along with ResNet50 and DensNet201 models for benchmarking process. Then we proposed two new models one is a Gated Recurrent Unit (GRU) based and another one is a Hybrid of Convolutional Neural Network (CNN) and Convolutional Long Short-term Memory (ConvLSTM). As for our proposed GRU model it performs closely to the DensNet201 and REsNet50 models while requiring very few parameters compared to these two models. On the other hand, our proposed Hybrid ConvLSTM model outperforms both of the aforementioned benchmarking models. Finally, we have developed a new Bangla Handwriting Numerical dataset containing a total of seven thousand training, one thousand validation, and two thousand test images. Our proposed best-performing model (Hybrid ConvLSTM) achieves 98.84% accuracy in the test data of our dataset while the GRU model gained 91.33% test accuracy without any help of image preprocessing steps.
{"title":"Bangla Handwritten Digit Recognition using RNN-CNN Hybrid Approach","authors":"A. Hasib Uddin, Joygun Khatun, Mehera Afroz Meghna, Prince Mahmud","doi":"10.1109/ICCIT57492.2022.10055089","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055089","url":null,"abstract":"The automatic recognition of handwritten English material has seen a lot of progress. However, research on automatic Bangla handwriting numerals recognition is far behind. Even the most effective recognizers now in use do not produce an adequate performance for real-world applications. This paper suggested a strategy based on deep neural networks. In this paper, we have used the BanglaLekha-Isolated handwriting dataset along with ResNet50 and DensNet201 models for benchmarking process. Then we proposed two new models one is a Gated Recurrent Unit (GRU) based and another one is a Hybrid of Convolutional Neural Network (CNN) and Convolutional Long Short-term Memory (ConvLSTM). As for our proposed GRU model it performs closely to the DensNet201 and REsNet50 models while requiring very few parameters compared to these two models. On the other hand, our proposed Hybrid ConvLSTM model outperforms both of the aforementioned benchmarking models. Finally, we have developed a new Bangla Handwriting Numerical dataset containing a total of seven thousand training, one thousand validation, and two thousand test images. Our proposed best-performing model (Hybrid ConvLSTM) achieves 98.84% accuracy in the test data of our dataset while the GRU model gained 91.33% test accuracy without any help of image preprocessing steps.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125383159","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10054721
Afridi Ibn Rahman, Zebel-E.-Noor Akhand, Tasin Al Nahian Khan, Anirudh Sarda, Subhi Bhuiyan, Mma Rakib, Zubayer Ahmed Fahim, Indronil Kundu
The COVID-19 pandemic has obligated people to adopt the virtual lifestyle. Currently, the use of videoconferencing to conduct business meetings is prevalent owing to the numerous benefits it presents. However, a large number of people with speech impediment find themselves handicapped to the new normal as they cannot communicate their ideas effectively, especially in fast paced meetings. Therefore, this paper aims to introduce an enriched dataset using an action recognition method with the most common phrases translated into American Sign Language (ASL) that are routinely used in professional meetings. It further proposes a sign language detecting and classifying model employing deep learning architectures, namely, CNN and LSTM. The performances of these models are analysed by employing different performance metrics like accuracy, recall, F1- Score and Precision. CNN and LSTM models yield an accuracy of 93.75% and 96.54% respectively, after being trained with the dataset introduced in this study. Therefore, the incorporation of the LSTM model into different cloud services, virtual private networks and softwares will allow people with speech impairment to use sign language, which will automatically be translated into captions using moving camera circumstances in real time. This will in turn equip other people with the tool to understand and grasp the message that is being conveyed and easily discuss and effectuate the ideas.
{"title":"Continuous Sign Language Interpretation to Text Using Deep Learning Models","authors":"Afridi Ibn Rahman, Zebel-E.-Noor Akhand, Tasin Al Nahian Khan, Anirudh Sarda, Subhi Bhuiyan, Mma Rakib, Zubayer Ahmed Fahim, Indronil Kundu","doi":"10.1109/ICCIT57492.2022.10054721","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10054721","url":null,"abstract":"The COVID-19 pandemic has obligated people to adopt the virtual lifestyle. Currently, the use of videoconferencing to conduct business meetings is prevalent owing to the numerous benefits it presents. However, a large number of people with speech impediment find themselves handicapped to the new normal as they cannot communicate their ideas effectively, especially in fast paced meetings. Therefore, this paper aims to introduce an enriched dataset using an action recognition method with the most common phrases translated into American Sign Language (ASL) that are routinely used in professional meetings. It further proposes a sign language detecting and classifying model employing deep learning architectures, namely, CNN and LSTM. The performances of these models are analysed by employing different performance metrics like accuracy, recall, F1- Score and Precision. CNN and LSTM models yield an accuracy of 93.75% and 96.54% respectively, after being trained with the dataset introduced in this study. Therefore, the incorporation of the LSTM model into different cloud services, virtual private networks and softwares will allow people with speech impairment to use sign language, which will automatically be translated into captions using moving camera circumstances in real time. This will in turn equip other people with the tool to understand and grasp the message that is being conveyed and easily discuss and effectuate the ideas.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126585367","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}
Nowadays, fraud correlated with credit cards became very prevalent since a lot of people use credit cards for buying goods and services. Because of e-commerce and technological advancement, most transactions are happening online, which is increasing the risk of fraudulent transactions and resulting in huge losses financially. Therefore, an effective detection technique, as the quickest prediction option, should be developed to deter fraud from propagating. This paper targeted to develop a deep learning (DL)-based model on SMOTE oversampling technique to predict the fraudulent transactions of credit cards. The system used three popular DL algorithms: Artificial Neural Network (ANN), Convolutional Neural Network (CNN), and Long Short-Term Memory Recurrent Neural Network (LSTM RNN), and measured the best performer in terms of evaluation metrics. However, the results confirm that the CNN algorithm outperformed both ANN and LSTM RNN. Additionally, compared to previous studies, our CNN fraud detection program recorded high rates of accuracy in identifying fraudulent activity. The system achieved an accuracy of 99.97%, precision of 99.94%, recall of 99.99%, and F1-Score of 99.96%. This proposed scheme can help to reduce financial loss by detecting credit card scams or frauds globally.
{"title":"SMOTE Based Credit Card Fraud Detection Using Convolutional Neural Network","authors":"Md. Nawab Yousuf Ali, Taniya Kabir, Noushin Laila Raka, Sanzida Siddikha Toma, Md. Lizur Rahman, J. Ferdaus","doi":"10.1109/ICCIT57492.2022.10054727","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10054727","url":null,"abstract":"Nowadays, fraud correlated with credit cards became very prevalent since a lot of people use credit cards for buying goods and services. Because of e-commerce and technological advancement, most transactions are happening online, which is increasing the risk of fraudulent transactions and resulting in huge losses financially. Therefore, an effective detection technique, as the quickest prediction option, should be developed to deter fraud from propagating. This paper targeted to develop a deep learning (DL)-based model on SMOTE oversampling technique to predict the fraudulent transactions of credit cards. The system used three popular DL algorithms: Artificial Neural Network (ANN), Convolutional Neural Network (CNN), and Long Short-Term Memory Recurrent Neural Network (LSTM RNN), and measured the best performer in terms of evaluation metrics. However, the results confirm that the CNN algorithm outperformed both ANN and LSTM RNN. Additionally, compared to previous studies, our CNN fraud detection program recorded high rates of accuracy in identifying fraudulent activity. The system achieved an accuracy of 99.97%, precision of 99.94%, recall of 99.99%, and F1-Score of 99.96%. This proposed scheme can help to reduce financial loss by detecting credit card scams or frauds globally.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125172527","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10055123
Abdullah Mohammad Sakib, Bilkis Jamal Ferdosi, S. Jahan, Kashfia Jashim
The right to health is one of the fundamental human rights. Every state is obliged to provide healthcare facilities to its population. In Bangladesh, the government is working hard to provide a better healthcare system, though the country needs to go a long way to have a unified healthcare system. There is a lack of a proper referral system in the country, and proper diagnosis is hindered due to a patient’s lack of medical history. In this paper, we propose a system that helps the patient to create a medical history from images of the prescriptions. Our system extracts and classifies data from an unstructured Bangladeshi medical prescription that can be used to create a repository of medical history. The proposed method works in four phases: phase I text localization and extraction from the images of prescriptions, phase II - classification of the extracted images, phase III - image to text conversion using OCR, and phase IV - classification of the text in four categories symptoms, medicines, diagnostic tests, and others. For image classification, we use a very deep convolutional network, VGG-16 and for text classification, we use the Bidirectional Encoder Representations from Transformers (BERT) model. Performance evaluation of the proposed system is very promising and the system can be used in any country like Bangladesh to facilitate better treatment.
{"title":"Medical Text Extraction and Classification from Prescription Images","authors":"Abdullah Mohammad Sakib, Bilkis Jamal Ferdosi, S. Jahan, Kashfia Jashim","doi":"10.1109/ICCIT57492.2022.10055123","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055123","url":null,"abstract":"The right to health is one of the fundamental human rights. Every state is obliged to provide healthcare facilities to its population. In Bangladesh, the government is working hard to provide a better healthcare system, though the country needs to go a long way to have a unified healthcare system. There is a lack of a proper referral system in the country, and proper diagnosis is hindered due to a patient’s lack of medical history. In this paper, we propose a system that helps the patient to create a medical history from images of the prescriptions. Our system extracts and classifies data from an unstructured Bangladeshi medical prescription that can be used to create a repository of medical history. The proposed method works in four phases: phase I text localization and extraction from the images of prescriptions, phase II - classification of the extracted images, phase III - image to text conversion using OCR, and phase IV - classification of the text in four categories symptoms, medicines, diagnostic tests, and others. For image classification, we use a very deep convolutional network, VGG-16 and for text classification, we use the Bidirectional Encoder Representations from Transformers (BERT) model. Performance evaluation of the proposed system is very promising and the system can be used in any country like Bangladesh to facilitate better treatment.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123584448","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10055391
N. Khan, Md Shamiul Islam, Fuad Chowdhury, Abdur Samad Siham, Nazmus Sakib
In our daily lives, newspapers and online news portals have become ubiquitous. These provide us with information on global events. Of all the news available in newspapers, crime news is the most significant. People read this kind of news with sincerity and considerable curiosity. We read a lot of Bangla newspapers and news sources, but we didn’t find any news on crime that was categorized. Perhaps categorizing the Bangla crime news would be helpful for the readers. Therefore, we decided to work on Bengali crime news classification, which will have a big influence in the Bengali community. However, categorizing crime news from daily newspaper headlines is not an easy task for a human. In this paper, we introduced a practical model to automatically annotate crime news from Bengali newspaper headlines in 6 predetermined crimes. In order to accomplish this goal, we have used TF-IDF for extracting features with 8 different machine learning and language classifier models (SVM, Decision Tree,Random Forest, LSTM, Bi-LSTM, BERT etc) and got best result by Sagor Sarkar’s Bangla-Bert-Base. The experimental result with 6293 training and 1574 testing samples shows 90.15% accuracy. This research output and dataset can be utilized by enthusiasts for further research purposes like subsetting crimes, crime status or judgment analysis etc. Our dataset will be available upon request @https://tinyurl.com/5n7wwaek.
{"title":"Bengali Crime News Classification Based on Newspaper Headlines using NLP","authors":"N. Khan, Md Shamiul Islam, Fuad Chowdhury, Abdur Samad Siham, Nazmus Sakib","doi":"10.1109/ICCIT57492.2022.10055391","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055391","url":null,"abstract":"In our daily lives, newspapers and online news portals have become ubiquitous. These provide us with information on global events. Of all the news available in newspapers, crime news is the most significant. People read this kind of news with sincerity and considerable curiosity. We read a lot of Bangla newspapers and news sources, but we didn’t find any news on crime that was categorized. Perhaps categorizing the Bangla crime news would be helpful for the readers. Therefore, we decided to work on Bengali crime news classification, which will have a big influence in the Bengali community. However, categorizing crime news from daily newspaper headlines is not an easy task for a human. In this paper, we introduced a practical model to automatically annotate crime news from Bengali newspaper headlines in 6 predetermined crimes. In order to accomplish this goal, we have used TF-IDF for extracting features with 8 different machine learning and language classifier models (SVM, Decision Tree,Random Forest, LSTM, Bi-LSTM, BERT etc) and got best result by Sagor Sarkar’s Bangla-Bert-Base. The experimental result with 6293 training and 1574 testing samples shows 90.15% accuracy. This research output and dataset can be utilized by enthusiasts for further research purposes like subsetting crimes, crime status or judgment analysis etc. Our dataset will be available upon request @https://tinyurl.com/5n7wwaek.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125533576","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10055555
M. Ahmed, S. Afrose, Ashik Adnan, Nazifa Khanom, Md Sabbir Hossain, Md Humaion Kabir Mehedi, Annajiat Alim Rasel
An excellent substitute for red meat, mushrooms are a rich, calorie-efficient source of protein, fiber, and antioxidants. Mushrooms may also be rich sources of potent medications. Therefore, it’s important to classify edible and poisonous mushrooms. An interpretable system for the identification of mushrooms is being developed using machine learning methods and Explainable Artificial Intelligence (XAI) models. The Mushroom dataset from the UC Irvine Machine Learning Repository was the one utilized in this study. Among the six ML models, Decision Tree, Random Forest, and KNN performed flawlessly in this dataset, achieving 100% accuracy. Whereas, SVM had a 98% accuracy rate, compared to 95% for Logistic Regression and 93% for Naive Bayes. The two XAI models SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model Agnostic Explanation) were used to interpret the top three ML models.
蘑菇是红肉的极好替代品，富含蛋白质、纤维和抗氧化剂。蘑菇也可能是有效药物的丰富来源。因此，对食用蘑菇和有毒蘑菇进行分类是很重要的。利用机器学习方法和可解释的人工智能(XAI)模型，正在开发一种用于识别蘑菇的可解释系统。来自加州大学欧文分校机器学习库的蘑菇数据集是本研究中使用的数据集。在六个ML模型中，决策树、随机森林和KNN在该数据集中表现完美，准确率达到100%。然而，SVM的准确率为98%，而逻辑回归的准确率为95%，朴素贝叶斯的准确率为93%。两个XAI模型SHAP (SHapley Additive explanatory)和LIME (Local Interpretable Model Agnostic Explanation)被用来解释前三个ML模型。
{"title":"Comparative Analysis of Interpretable Mushroom Classification using Several Machine Learning Models","authors":"M. Ahmed, S. Afrose, Ashik Adnan, Nazifa Khanom, Md Sabbir Hossain, Md Humaion Kabir Mehedi, Annajiat Alim Rasel","doi":"10.1109/ICCIT57492.2022.10055555","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055555","url":null,"abstract":"An excellent substitute for red meat, mushrooms are a rich, calorie-efficient source of protein, fiber, and antioxidants. Mushrooms may also be rich sources of potent medications. Therefore, it’s important to classify edible and poisonous mushrooms. An interpretable system for the identification of mushrooms is being developed using machine learning methods and Explainable Artificial Intelligence (XAI) models. The Mushroom dataset from the UC Irvine Machine Learning Repository was the one utilized in this study. Among the six ML models, Decision Tree, Random Forest, and KNN performed flawlessly in this dataset, achieving 100% accuracy. Whereas, SVM had a 98% accuracy rate, compared to 95% for Logistic Regression and 93% for Naive Bayes. The two XAI models SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model Agnostic Explanation) were used to interpret the top three ML models.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125665015","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10055136
Farjana Parvin, Md. Al Mamun
Early identification of brain tumors greatly influences the clinical diagnosis process of a brain tumor patient. Therefore, this study suggests a brain tumor detection approach that merges deep and shallow features extracted from the brain MRI images in order to distinguish between non-tumor and tumor classes. We combine some pre-trained deep CNN architectures and the concept of transfer learning in our proposed framework to obtain high-level features from magnetic resonance images. Following the extraction, a Support Vector Machine classifier with radial basis function was used to evaluate the deep features. A deep feature vector is then created by combining the best three deep features that perform well on the SVM classifier. Even though deep features are crucial for classification, as the network becomes deeper, some low-level features might be lost. Therefore, a shallow network was intended to learn low-level information from the brain MRI. Deep and shallow features are then merged to compensate for the information loss. The fused feature vector is then employed, in order to train a support vector machine classifier. The experimental results were obtained on a publicly available dataset. Our proposed framework has achieved a high accuracy of 92.48% (with a precision of 93.64%, recall of 94.55%, and f1-score of 93.97%). The results also showed that utilizing this feature fusion enhances the performance of the classification framework and these results ensure the hypothesis that features fusion enables the compensation of low-level information lost. Moreover, our classification approach outperformed others when compared to state-of-the-art studies.
{"title":"Feature Fusion Based Effective Brain Tumor Detection Approach Using MRI","authors":"Farjana Parvin, Md. Al Mamun","doi":"10.1109/ICCIT57492.2022.10055136","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055136","url":null,"abstract":"Early identification of brain tumors greatly influences the clinical diagnosis process of a brain tumor patient. Therefore, this study suggests a brain tumor detection approach that merges deep and shallow features extracted from the brain MRI images in order to distinguish between non-tumor and tumor classes. We combine some pre-trained deep CNN architectures and the concept of transfer learning in our proposed framework to obtain high-level features from magnetic resonance images. Following the extraction, a Support Vector Machine classifier with radial basis function was used to evaluate the deep features. A deep feature vector is then created by combining the best three deep features that perform well on the SVM classifier. Even though deep features are crucial for classification, as the network becomes deeper, some low-level features might be lost. Therefore, a shallow network was intended to learn low-level information from the brain MRI. Deep and shallow features are then merged to compensate for the information loss. The fused feature vector is then employed, in order to train a support vector machine classifier. The experimental results were obtained on a publicly available dataset. Our proposed framework has achieved a high accuracy of 92.48% (with a precision of 93.64%, recall of 94.55%, and f1-score of 93.97%). The results also showed that utilizing this feature fusion enhances the performance of the classification framework and these results ensure the hypothesis that features fusion enables the compensation of low-level information lost. Moreover, our classification approach outperformed others when compared to state-of-the-art studies.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120947690","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10054811
Rakhi Rani Paul, S. Paul, Md. Ekramul Hamid
recognizing emotions from speech signals is one of the active research fields in the area of human information processing as well as man-machine interaction. Different persons have different emotions and altogether different ways of expressing them. In this paper, a 2D Convolutional Neural Network (CNN) based method is presented for human emotion classification. We consider RAVDESS and SAVEE datasets to evaluate the performance of the model. Initially, Mel-frequency cepstral coefficients MFCC features are extracted from the speech signals which are used for the training purpose. Here, we consider only forty (40) cepstrum coefficients per frame. The proposed 2D CNN model is trained to classify seven different emotional states (neutral, calm, happy, sad, angry, scared, disgust, surprised). We achieve 89.86% overall accuracy from our proposed model for the RAVDESS dataset and 83.57% for the SAVEE dataset respectively. It is found that happy class is classified with an accuracy of 96% for the RAVDESS dataset and 92% for the SAVEE dataset. Lastly, the result of our proposed model is compared with the other recent existing works. The performance of our proposed model is good enough because it achieves better accuracy than other models. This work has many real-life applications such as man-machine interaction, auto supervision, auxiliary lie detection, the discovery of dissatisfaction with the client’s mode, detecting neurological disordered patients and so on.
{"title":"A 2D Convolution Neural Network Based Method for Human Emotion Classification from Speech Signal","authors":"Rakhi Rani Paul, S. Paul, Md. Ekramul Hamid","doi":"10.1109/ICCIT57492.2022.10054811","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10054811","url":null,"abstract":"recognizing emotions from speech signals is one of the active research fields in the area of human information processing as well as man-machine interaction. Different persons have different emotions and altogether different ways of expressing them. In this paper, a 2D Convolutional Neural Network (CNN) based method is presented for human emotion classification. We consider RAVDESS and SAVEE datasets to evaluate the performance of the model. Initially, Mel-frequency cepstral coefficients MFCC features are extracted from the speech signals which are used for the training purpose. Here, we consider only forty (40) cepstrum coefficients per frame. The proposed 2D CNN model is trained to classify seven different emotional states (neutral, calm, happy, sad, angry, scared, disgust, surprised). We achieve 89.86% overall accuracy from our proposed model for the RAVDESS dataset and 83.57% for the SAVEE dataset respectively. It is found that happy class is classified with an accuracy of 96% for the RAVDESS dataset and 92% for the SAVEE dataset. Lastly, the result of our proposed model is compared with the other recent existing works. The performance of our proposed model is good enough because it achieves better accuracy than other models. This work has many real-life applications such as man-machine interaction, auto supervision, auxiliary lie detection, the discovery of dissatisfaction with the client’s mode, detecting neurological disordered patients and so on.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"117 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133788134","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 : 2022-12-17DOI: 10.1109/ICCIT57492.2022.10055232
Chokder Alamgir Bartend Russell, Shahriar Khan
With the recent blackouts in Texas, Manhattan, and NE US, the importance of analyzing outage is greater than ever. Contingency Analysis is helpful to increase the resiliency of power system by analyzing impact of different contingencies. The effect of single transmission line outage in a transmission network has been studied with IEEE 39 bus network. Each of the branches has been disconnected one at a time to find effect on generator constraints, voltage constraints of buses, transmission line loading and possibility of islanding. PSS®E Xplore 34 was used for the simulation. The results show that single transmission line outage may impact generator power factors, increase demand of reactive power from the generators, and overload other transmission lines. Transmission line outage may lead to several violations of system constraints leading to islanding. This study may help research on different impacts of single transmission line outage and improve power system resilience.
{"title":"Single Line Outage Analysis on IEEE 39 Bus Network","authors":"Chokder Alamgir Bartend Russell, Shahriar Khan","doi":"10.1109/ICCIT57492.2022.10055232","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055232","url":null,"abstract":"With the recent blackouts in Texas, Manhattan, and NE US, the importance of analyzing outage is greater than ever. Contingency Analysis is helpful to increase the resiliency of power system by analyzing impact of different contingencies. The effect of single transmission line outage in a transmission network has been studied with IEEE 39 bus network. Each of the branches has been disconnected one at a time to find effect on generator constraints, voltage constraints of buses, transmission line loading and possibility of islanding. PSS®E Xplore 34 was used for the simulation. The results show that single transmission line outage may impact generator power factors, increase demand of reactive power from the generators, and overload other transmission lines. Transmission line outage may lead to several violations of system constraints leading to islanding. This study may help research on different impacts of single transmission line outage and improve power system resilience.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131067596","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：