The remote sensing industry is actively discussing the classification of hyperspectral images (HSIs). For the first time, the idea of subgrouping dimensionality is presented using a modified deep learning model, and this research presents a novel framework for dimensionality reduction in HSI classification as a result. In particular, our system uses the subgrouping model to extract many characteristics from a dataset and then apply a selection criterion. First, we performed data reduction and subgrouping by extracting the correlation matrix. After that, we resample the data and use it as input for a hyperspectral picture classification. In the proposed framework, we combine NMF on spectral dimensions with information-based feature selection and a wavelet-based 2D CNN on spatial dimensions to classify spectral-spatial data. Based on the experimental findings, it is clear that this framework delivers the most excellent classification accuracy compared to other approaches, including traditional classifiers like PCA and MNF-based deep learning methods.
{"title":"Subgrouping-Based NMF with Imbalanced Class Handling for Hyperspectral Image Classification","authors":"Md. Touhid Islam, Mohadeb Kumar, Md. Rashedul Islam, Md. Sohrawordi","doi":"10.1109/ICCIT57492.2022.10055177","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055177","url":null,"abstract":"The remote sensing industry is actively discussing the classification of hyperspectral images (HSIs). For the first time, the idea of subgrouping dimensionality is presented using a modified deep learning model, and this research presents a novel framework for dimensionality reduction in HSI classification as a result. In particular, our system uses the subgrouping model to extract many characteristics from a dataset and then apply a selection criterion. First, we performed data reduction and subgrouping by extracting the correlation matrix. After that, we resample the data and use it as input for a hyperspectral picture classification. In the proposed framework, we combine NMF on spectral dimensions with information-based feature selection and a wavelet-based 2D CNN on spatial dimensions to classify spectral-spatial data. Based on the experimental findings, it is clear that this framework delivers the most excellent classification accuracy compared to other approaches, including traditional classifiers like PCA and MNF-based deep learning methods.","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":"124460173","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/ICCIT57492.2022.10054673
Ishraq Hasan, Muhammad Munswarim Khan, Kazi Tasnim Rahman, Anika Siddiqui Mayesha, Zinia Sultana, M. Islam
Virtual Reality (VR) is one of the pioneering technologies in the current decade. An increasing number of users are migrating to the virtual world as time progresses. This has led to users desiring more intuitive and natural control for their inputs in the digital world, surpassing the need for traditional input systems such as mouse and keyboard. VR based natural inputs systems are scarce, and the available ones are expensive. Again, very few of them translates the normal hand movements into the virtual control inputs. Therefore, the objective is to design and develop a wearable input system for controlling VR with enhanced usability and high accuracy. To attain this objective, user requirements were firstly elicited through semi-structured interviews. Then, a cost-effective and usable wearable system (VR Glove) was developed based on the revealed requirements for controlling VR. Finally, the system was evaluated with 20 test-participants (novice and expert); and it was found that the VR Glove was usable both to the novice and expert users, though the system was more usable to experts than the novice users. Participants were comfortable with the working mechanism of the proposed VR Glove system and also found the system very responsive.
{"title":"VR Glove: A Virtual Input System for Controlling VR with Enhanced Usability and High Accuracy","authors":"Ishraq Hasan, Muhammad Munswarim Khan, Kazi Tasnim Rahman, Anika Siddiqui Mayesha, Zinia Sultana, M. Islam","doi":"10.1109/ICCIT57492.2022.10054673","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10054673","url":null,"abstract":"Virtual Reality (VR) is one of the pioneering technologies in the current decade. An increasing number of users are migrating to the virtual world as time progresses. This has led to users desiring more intuitive and natural control for their inputs in the digital world, surpassing the need for traditional input systems such as mouse and keyboard. VR based natural inputs systems are scarce, and the available ones are expensive. Again, very few of them translates the normal hand movements into the virtual control inputs. Therefore, the objective is to design and develop a wearable input system for controlling VR with enhanced usability and high accuracy. To attain this objective, user requirements were firstly elicited through semi-structured interviews. Then, a cost-effective and usable wearable system (VR Glove) was developed based on the revealed requirements for controlling VR. Finally, the system was evaluated with 20 test-participants (novice and expert); and it was found that the VR Glove was usable both to the novice and expert users, though the system was more usable to experts than the novice users. Participants were comfortable with the working mechanism of the proposed VR Glove system and also found the system very responsive.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"93 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":"115828472","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/ICCIT57492.2022.10055557
M. H. Maruf, Syed Iftekhar Ali
Memristor-based ternary content addressable memory (MTCAM) is a form of special memory where the memristor controls the primary operation instead of transistors. In addition, a memristor is a kind of particular passive element with two terminals that keeps the data as memory when the power goes down. This paper proposes a novel 5T4M MTCAM that is compact in size, efficient in energy consumption, and capable of restoring data. The proposed design uses BSIM 32nm CMOS PTM as a transistor model and a modified Biolek model as a memristor model for simulation. 16x16 array of MTCAM has been used with pre-charge low matchline (ML) sensing. This novel MTCAM offers 633ps of search time and 1.65fJ/digit/search of search energy which are lower than the other existing designs. In addition, this design can restore the data in successive search cycles though it performs its write and search operations using the same nodes.
{"title":"Novel Memristor-based Energy Efficient Compact 5T4M Ternary Content Addressable Memory","authors":"M. H. Maruf, Syed Iftekhar Ali","doi":"10.1109/ICCIT57492.2022.10055557","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055557","url":null,"abstract":"Memristor-based ternary content addressable memory (MTCAM) is a form of special memory where the memristor controls the primary operation instead of transistors. In addition, a memristor is a kind of particular passive element with two terminals that keeps the data as memory when the power goes down. This paper proposes a novel 5T4M MTCAM that is compact in size, efficient in energy consumption, and capable of restoring data. The proposed design uses BSIM 32nm CMOS PTM as a transistor model and a modified Biolek model as a memristor model for simulation. 16x16 array of MTCAM has been used with pre-charge low matchline (ML) sensing. This novel MTCAM offers 633ps of search time and 1.65fJ/digit/search of search energy which are lower than the other existing designs. In addition, this design can restore the data in successive search cycles though it performs its write and search operations using the same nodes.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"21 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":"127449885","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/ICCIT57492.2022.10055054
Partho Ghose, Md. Ashraf Uddin, Mohammad Manzurul Islam, Manowarul Islam, U. Acharjee
Breast cancer has become a common disease that affects women all over the world. Early detection and diagnosis of the breast cancer is crucial for an effective medication and treatment. But, detection of breast cancer at the primary stage is challenging due to the ambiguity of the mammograms. Many researchers have explored Machine learning (ML) based model to detect breast cancer. Most of the developed models have not been clinically effective. To address this, in this paper, we propose an optimized SVM based model for the prediction of breast cancer where Bayesian search method is applied to discover the best hyper-parameters of the SVM classifier. Performance of the model with default hyper-parameter for the SVM is compared to the performance with tuned hyper-parameter. The comparison shows that performance is significantly improved when the tuned hyper-parameter is used for training SVM classifier. Our findings show that SVM’s performance with default parameters is 96% whereas the maximum accuracy level 98% is obtained using tuned hyper-parameter.
{"title":"A Breast Cancer Detection Model using a Tuned SVM Classifier","authors":"Partho Ghose, Md. Ashraf Uddin, Mohammad Manzurul Islam, Manowarul Islam, U. Acharjee","doi":"10.1109/ICCIT57492.2022.10055054","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055054","url":null,"abstract":"Breast cancer has become a common disease that affects women all over the world. Early detection and diagnosis of the breast cancer is crucial for an effective medication and treatment. But, detection of breast cancer at the primary stage is challenging due to the ambiguity of the mammograms. Many researchers have explored Machine learning (ML) based model to detect breast cancer. Most of the developed models have not been clinically effective. To address this, in this paper, we propose an optimized SVM based model for the prediction of breast cancer where Bayesian search method is applied to discover the best hyper-parameters of the SVM classifier. Performance of the model with default hyper-parameter for the SVM is compared to the performance with tuned hyper-parameter. The comparison shows that performance is significantly improved when the tuned hyper-parameter is used for training SVM classifier. Our findings show that SVM’s performance with default parameters is 96% whereas the maximum accuracy level 98% is obtained using tuned hyper-parameter.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"260 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":"123473699","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/ICCIT57492.2022.10054785
Syed Taha Yeasin Ramadan, T. Sakib, Md. Ahsan Rahat, Md. Mushfique Hossain, Raiyan Rahman, Md. Mahbubur Rahman
Intelligible speech, while it provides an excellent means of communication for humans and sets us apart from other lifeforms, our abuse of speech creates deep and lasting issues in our society. The use of derogatory language has a significant impact not only on children’s mental health but also on adults, for instance, in an abusive work environment. Accountability for such actions is one of the key steps toward maintaining a healthy atmosphere or at least making it less frequent. In this paper, we describe our work on detecting abusive or hate speech in Bangla in real time. Our system converts the speech to text and then uses NLP and deep learning to detect such occurrences in real-time. Also, if the voice is registered on our system, it identifies the person engaging in abusive words, opening ways to greater workplace accountability. We also describe our mobile application and the microcontroller-based standalone embedded system that can be deployed in target places (for instance, daycare centers, schools, workplaces, etc.) to record audio and detect the abusive speech and the speaker in real-time. Several datasets have been deployed on the LSTM, Bi-LSTM, GRU, and BERT models to assess the system’s efficacy. Identification of the individual speaking the words is done using the audio signal extraction feature MFCC. The experimental results show that the BERT model provides the highest accuracy compared to other algorithms.
{"title":"An Integrated Embedded System Towards Abusive Bengali Speech and Speaker Detection Using NLP and Deep Learning","authors":"Syed Taha Yeasin Ramadan, T. Sakib, Md. Ahsan Rahat, Md. Mushfique Hossain, Raiyan Rahman, Md. Mahbubur Rahman","doi":"10.1109/ICCIT57492.2022.10054785","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10054785","url":null,"abstract":"Intelligible speech, while it provides an excellent means of communication for humans and sets us apart from other lifeforms, our abuse of speech creates deep and lasting issues in our society. The use of derogatory language has a significant impact not only on children’s mental health but also on adults, for instance, in an abusive work environment. Accountability for such actions is one of the key steps toward maintaining a healthy atmosphere or at least making it less frequent. In this paper, we describe our work on detecting abusive or hate speech in Bangla in real time. Our system converts the speech to text and then uses NLP and deep learning to detect such occurrences in real-time. Also, if the voice is registered on our system, it identifies the person engaging in abusive words, opening ways to greater workplace accountability. We also describe our mobile application and the microcontroller-based standalone embedded system that can be deployed in target places (for instance, daycare centers, schools, workplaces, etc.) to record audio and detect the abusive speech and the speaker in real-time. Several datasets have been deployed on the LSTM, Bi-LSTM, GRU, and BERT models to assess the system’s efficacy. Identification of the individual speaking the words is done using the audio signal extraction feature MFCC. The experimental results show that the BERT model provides the highest accuracy compared to other algorithms.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"1 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":"129323622","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/ICCIT57492.2022.10055066
S. M. Rahman, Md. Abdullah Al Mamun, Md. Asraf Ali
Acceleration, angular velocity and electromyographic (EMG) signal at the lower limb muscles, specially over both leg's Tibialis Anterior muscles are highly non-stationary, even if no perturbing influences can be identified during walking at any speed. This study analyzed the fractal dynamics (i.e., complexity of gait time series) in the walking gait time series of four types of signals obtained from wearable sensors such as IMUs (inertial measurement units), i.e., accelerometer signals which represents the acceleration experienced by the body, gyroscope signals which is the angular velocity, and magnetometer signals which is magnetic field vector, and Electromyographic (EMG) signal from both leg’s Tibialis Anterior muscles. Gait time series from twenty-two healthy participants were analyzed while they performed walking at their comfortable speed. The scaling exponents (i.e., α-values) of the gait dynamics were accomplished by evaluating their fluctuation through detrended fluctuation analysis (DFA), which is most common and widely used non-linear technique for any non-stationary time series. DFA (the scaling exponents α) results established an anti-persistent in EMG and acceleration signal, less persistent pattern in angular velocity and persistent (i.e., long-range or fractal-like correlations) in magnetometer signal. This fractal complexity or noise patterns obtained from the EMG and inertial signals might provide new approaches for assessing and forecasting sudden injury risk during walking.
{"title":"Fractal Pattern Identification from Wearable Inertial and Electromyographic Signals Data during Walking","authors":"S. M. Rahman, Md. Abdullah Al Mamun, Md. Asraf Ali","doi":"10.1109/ICCIT57492.2022.10055066","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055066","url":null,"abstract":"Acceleration, angular velocity and electromyographic (EMG) signal at the lower limb muscles, specially over both leg's Tibialis Anterior muscles are highly non-stationary, even if no perturbing influences can be identified during walking at any speed. This study analyzed the fractal dynamics (i.e., complexity of gait time series) in the walking gait time series of four types of signals obtained from wearable sensors such as IMUs (inertial measurement units), i.e., accelerometer signals which represents the acceleration experienced by the body, gyroscope signals which is the angular velocity, and magnetometer signals which is magnetic field vector, and Electromyographic (EMG) signal from both leg’s Tibialis Anterior muscles. Gait time series from twenty-two healthy participants were analyzed while they performed walking at their comfortable speed. The scaling exponents (i.e., α-values) of the gait dynamics were accomplished by evaluating their fluctuation through detrended fluctuation analysis (DFA), which is most common and widely used non-linear technique for any non-stationary time series. DFA (the scaling exponents α) results established an anti-persistent in EMG and acceleration signal, less persistent pattern in angular velocity and persistent (i.e., long-range or fractal-like correlations) in magnetometer signal. This fractal complexity or noise patterns obtained from the EMG and inertial signals might provide new approaches for assessing and forecasting sudden injury risk during walking.","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":"121492892","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/ICCIT57492.2022.10054941
Md. Ibne Joha, Md. Shafiul Islam, S. Ahamed
Internet connectivity has become an essential aspect of the 21st century. Internet of Things (IoT) has embedded devices, including various sensors, software, and appliances. IoT-based intelligent home automation system offers the automated monitoring, operation, and control of home appliances via the internet. This paper presents an IoT-based, intelligent control and protection system for home appliances using NodeMCU and Blynk app. System control and operation via the app are independent of the Wi-Fi network’s nature. It eliminates additional coding to connect the system to any Wi-Fi network rather than home Wi-Fi. Moreover, it automatically detects and protects home appliances from damage against overheating, overloading, gas leakage, and fire hazards. Furthermore, the dimmer circuit offers fan speed and light intensity control. The system also provides a time-scheduling option and voice command control via Google Assistant. Above all, it saves energy and enhances human comfort.
{"title":"IoT-Based Smart Control and Protection System for Home Appliances","authors":"Md. Ibne Joha, Md. Shafiul Islam, S. Ahamed","doi":"10.1109/ICCIT57492.2022.10054941","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10054941","url":null,"abstract":"Internet connectivity has become an essential aspect of the 21st century. Internet of Things (IoT) has embedded devices, including various sensors, software, and appliances. IoT-based intelligent home automation system offers the automated monitoring, operation, and control of home appliances via the internet. This paper presents an IoT-based, intelligent control and protection system for home appliances using NodeMCU and Blynk app. System control and operation via the app are independent of the Wi-Fi network’s nature. It eliminates additional coding to connect the system to any Wi-Fi network rather than home Wi-Fi. Moreover, it automatically detects and protects home appliances from damage against overheating, overloading, gas leakage, and fire hazards. Furthermore, the dimmer circuit offers fan speed and light intensity control. The system also provides a time-scheduling option and voice command control via Google Assistant. Above all, it saves energy and enhances human comfort.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"31 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":"121533356","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/ICCIT57492.2022.10055553
Sk. Azmaeen Bin Amir, Abid Hossain Khan
A thermal power plant's efficiency and output power are very sensitive to its surrounding weather conditions. Since a nuclear power plant (NPP) usually runs at lower thermodynamic efficiency compared to other thermal power plants, an additional decrease in output power may challenge the economic viability of the project. Thus, it is very important to establish a sufficiently accurate model than can depict the correlation between NPP output power and condenser pressure. This work attempts to calibrate a simplified thermodynamic model using two evolutionary algorithms, Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). For GA, the initial population is varied in the range of 10-1000, while the mutation and crossover rates are taken as 0.01 and 0.50, respectively. For PSO, the swarm size is varied within the range of 100-1000. Results reveal that the calibrated model has more accurate predictions compared to the original model. The model calibrated with GA is found to be slightly better performing than the one calibrated with PSO. Additionally, the calibration process is observed to be insensitive to the reference condenser pressure. Finally, it is estimated that the efficiency of the plant can go down to 33.56% at 15kPa condenser pressure compared to 37.30% at 4kPa.
{"title":"Calibration of a simplified thermodynamic model for VVER-1200-based nuclear power plants using evolutionary algorithms","authors":"Sk. Azmaeen Bin Amir, Abid Hossain Khan","doi":"10.1109/ICCIT57492.2022.10055553","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055553","url":null,"abstract":"A thermal power plant's efficiency and output power are very sensitive to its surrounding weather conditions. Since a nuclear power plant (NPP) usually runs at lower thermodynamic efficiency compared to other thermal power plants, an additional decrease in output power may challenge the economic viability of the project. Thus, it is very important to establish a sufficiently accurate model than can depict the correlation between NPP output power and condenser pressure. This work attempts to calibrate a simplified thermodynamic model using two evolutionary algorithms, Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). For GA, the initial population is varied in the range of 10-1000, while the mutation and crossover rates are taken as 0.01 and 0.50, respectively. For PSO, the swarm size is varied within the range of 100-1000. Results reveal that the calibrated model has more accurate predictions compared to the original model. The model calibrated with GA is found to be slightly better performing than the one calibrated with PSO. Additionally, the calibration process is observed to be insensitive to the reference condenser pressure. Finally, it is estimated that the efficiency of the plant can go down to 33.56% at 15kPa condenser pressure compared to 37.30% at 4kPa.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"10 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":"126366519","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/ICCIT57492.2022.10055421
Afrah Rashid, Syeda Sohana Binta Farhad, Afsana Bhuyian, N. Yeasmin, Mohammad Abdul Azim, Z. Alom
Breast cancer is one of the most common malignancies affecting women worldwide, with many fatalities yearly. The risk of death suffered by breast cancer is increasing exponentially. Due to a surge of development of research in the medical field, providing more timely and possible early detection of disease has become a time-demanding option. By far, radiologists have manually checked cancer images and diagnosed them. Research has shown that a considerable number of ultrasound images are created every individual day. However, the number of radiologists is limited, so they cannot provide service on time. However, they often misclassify breast lesions, resulting in a high false-positive rate. An automatic system for detecting disease assists radiologists in disease diagnosis and provides reliable, productive, and reduces the risk of death. In this paper, we compare six machine learning models, namely (i) Support Vector Machine (SVM), (ii) Naive Bayes (NB), (iii) Logistic Regression (LR), (iv) Decision Tree (DT), (v) Random Forest (RF), and (vi) k-Nearest Neighbors (k-NN) on two different datasets (i) the Wisconsin Breast Cancer Dataset (WBCD) and (ii) the Breast Cancer Coimbra Dataset (BCCD). This study aims to create different classification models to analyze the obtained results and compare them to predict breast cancer. We use several performance metrics to select the best classification model among them. Our comparative analysis shows that SVM models can achieve better performance metrics, and thus the model of this research possesses relevant to use in clinical applications.
{"title":"A Comparative Analysis of Machine Learning techniques on Breast Cancer diagnosis using WEKA","authors":"Afrah Rashid, Syeda Sohana Binta Farhad, Afsana Bhuyian, N. Yeasmin, Mohammad Abdul Azim, Z. Alom","doi":"10.1109/ICCIT57492.2022.10055421","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055421","url":null,"abstract":"Breast cancer is one of the most common malignancies affecting women worldwide, with many fatalities yearly. The risk of death suffered by breast cancer is increasing exponentially. Due to a surge of development of research in the medical field, providing more timely and possible early detection of disease has become a time-demanding option. By far, radiologists have manually checked cancer images and diagnosed them. Research has shown that a considerable number of ultrasound images are created every individual day. However, the number of radiologists is limited, so they cannot provide service on time. However, they often misclassify breast lesions, resulting in a high false-positive rate. An automatic system for detecting disease assists radiologists in disease diagnosis and provides reliable, productive, and reduces the risk of death. In this paper, we compare six machine learning models, namely (i) Support Vector Machine (SVM), (ii) Naive Bayes (NB), (iii) Logistic Regression (LR), (iv) Decision Tree (DT), (v) Random Forest (RF), and (vi) k-Nearest Neighbors (k-NN) on two different datasets (i) the Wisconsin Breast Cancer Dataset (WBCD) and (ii) the Breast Cancer Coimbra Dataset (BCCD). This study aims to create different classification models to analyze the obtained results and compare them to predict breast cancer. We use several performance metrics to select the best classification model among them. Our comparative analysis shows that SVM models can achieve better performance metrics, and thus the model of this research possesses relevant to use in clinical applications.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"3 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":"121968608","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/ICCIT57492.2022.10055470
Shabbir Ahmed, Md Abu Marjan, M. Rahman, Md. Shahriar Haque Shemul, Md. Palash Uddin, M. I. Afjal
Remote sensing through neighboring constrained spectral wavelength bands, the hyperspectral image (HSI) contains significant information about the land objects. Using all of the original HSI features (bands), it appears that the classification performance is inadequate. To attenuate this, band (dimensionality) reduction schemes using feature extraction and feature selection techniques are frequently used in order to enhance classification performance. Despite being often employed for HSI feature reduction, Principal Component Analysis (PCA) usually struggles to retrieve the local desired HSI features since it only evaluates the HSI’s global statistics. Therefore, Spectrally-Segmented-PCA (SSPCA) and Incremental-PCA (IPCA) are presented to supplant the classical PCA. In this paper, we propose the Spectrally-Segmented-Incremental-PCA (SSIPCA) feature extraction approach to make use of the utility of both the SSPCA and the IPCA. Specifically, SSIPCA divides the whole HSI into a number of spectrally separated bands’ subgroups before applying the standard IPCA to each subgroup independently. We experiment with the Indian Pines mixed agricultural HSI classification to assess the proposed SSIPCA employing a perpixel Support Vector Machine (SVM) as the classifier. Based on the classification accuracy, we evince that the proposed SSIPCA approach (90.78% & 88.702%) outperforms the entire original bands of HSI (87.610% & 86.361%), PCA (88.78% & 86.985%), IPCA (89.171% & 86.576%) and SSPCA (90.634% & 88.468%) feature extraction methods.
{"title":"Spectrally-Segmented-Incremental-PCA for Hyperspectral Image Classification","authors":"Shabbir Ahmed, Md Abu Marjan, M. Rahman, Md. Shahriar Haque Shemul, Md. Palash Uddin, M. I. Afjal","doi":"10.1109/ICCIT57492.2022.10055470","DOIUrl":"https://doi.org/10.1109/ICCIT57492.2022.10055470","url":null,"abstract":"Remote sensing through neighboring constrained spectral wavelength bands, the hyperspectral image (HSI) contains significant information about the land objects. Using all of the original HSI features (bands), it appears that the classification performance is inadequate. To attenuate this, band (dimensionality) reduction schemes using feature extraction and feature selection techniques are frequently used in order to enhance classification performance. Despite being often employed for HSI feature reduction, Principal Component Analysis (PCA) usually struggles to retrieve the local desired HSI features since it only evaluates the HSI’s global statistics. Therefore, Spectrally-Segmented-PCA (SSPCA) and Incremental-PCA (IPCA) are presented to supplant the classical PCA. In this paper, we propose the Spectrally-Segmented-Incremental-PCA (SSIPCA) feature extraction approach to make use of the utility of both the SSPCA and the IPCA. Specifically, SSIPCA divides the whole HSI into a number of spectrally separated bands’ subgroups before applying the standard IPCA to each subgroup independently. We experiment with the Indian Pines mixed agricultural HSI classification to assess the proposed SSIPCA employing a perpixel Support Vector Machine (SVM) as the classifier. Based on the classification accuracy, we evince that the proposed SSIPCA approach (90.78% & 88.702%) outperforms the entire original bands of HSI (87.610% & 86.361%), PCA (88.78% & 86.985%), IPCA (89.171% & 86.576%) and SSPCA (90.634% & 88.468%) feature extraction methods.","PeriodicalId":255498,"journal":{"name":"2022 25th International Conference on Computer and Information Technology (ICCIT)","volume":"149 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":"115959692","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：