Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10164881
A. Mary, A. Edison
Deep learning is a sophisticated and adaptable technique that has found widespread use in fields such as natural language processing, machine learning, and computer vision. It is one of the most recent deep learning-powered applications to emerge. Deep fakes are altered, high-quality, realistic videos/images that have lately gained popularity. Many incredible uses of this technology are being investigated. Malicious uses of fake videos, such as fake news, celebrity pornographic videos, financial scams, and revenge porn are currently on the rise in the digital world. As a result, celebrities, politicians, and other well-known persons are particularly vulnerable to the Deep fake detection challenge. Numerous research has been undertaken in recent years to understand how deep fakes function and many deep learning-based algorithms to detect deep fake videos or pictures have been presented.This study comprehensively evaluates deep fake production and detection technologies based on several deep learning algorithms. In addition, the limits of current approaches and the availability of databases in society will be discussed. A deep fake detection system that is both precise and automatic. Given the ease with which deep fake videos/images may be generated and shared, the lack of an effective deep fake detection system creates a serious problem for the world. However, there have been various attempts to address this issue, and deep learning-related solutions outperform traditional approaches.
{"title":"Deep fake Detection using deep learning techniques: A Literature Review","authors":"A. Mary, A. Edison","doi":"10.1109/ICCC57789.2023.10164881","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10164881","url":null,"abstract":"Deep learning is a sophisticated and adaptable technique that has found widespread use in fields such as natural language processing, machine learning, and computer vision. It is one of the most recent deep learning-powered applications to emerge. Deep fakes are altered, high-quality, realistic videos/images that have lately gained popularity. Many incredible uses of this technology are being investigated. Malicious uses of fake videos, such as fake news, celebrity pornographic videos, financial scams, and revenge porn are currently on the rise in the digital world. As a result, celebrities, politicians, and other well-known persons are particularly vulnerable to the Deep fake detection challenge. Numerous research has been undertaken in recent years to understand how deep fakes function and many deep learning-based algorithms to detect deep fake videos or pictures have been presented.This study comprehensively evaluates deep fake production and detection technologies based on several deep learning algorithms. In addition, the limits of current approaches and the availability of databases in society will be discussed. A deep fake detection system that is both precise and automatic. Given the ease with which deep fake videos/images may be generated and shared, the lack of an effective deep fake detection system creates a serious problem for the world. However, there have been various attempts to address this issue, and deep learning-related solutions outperform traditional approaches.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131911422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165205
A. Np, Pournami P.N., J. P. B.
Medical image diagnostics now benefit greatly from the use of deep convolutional neural networks. The CNN-based deep neural networks are extensively used in the medical classification tasks. Although deep learning algorithms have gained performance comparable to pathologists in interpreting whole slide images for the detection of tumours, very few researchers have explored the possibility of detecting urothelial carcinoma from microscopic images. In this study, we propose a novel deep learning model for urine cytology smear-based detection of urothelial cell cancer (UCC). The network is based on Inception architecture that can efficiently learn the features of varied size cells in the image and produced relatively high accuracy when compared to state-of-the-art techniques. The proposed technique is evaluated using a dataset that includes the cytology samples of 115 individuals, 59 of whom had UCC instances that were histologically confirmed and the remaining 56 benign cases were identified through routine cytology samples. The suggested approach offers 98.63% accuracy with fewer parameters. The method’s performance in terms of accuracy and parameter count is highly encouraging.
{"title":"An Inception based Urothelial Cell Classification Network for the detection of Bladder Carcinoma from Urine Cytology Microscopic Images","authors":"A. Np, Pournami P.N., J. P. B.","doi":"10.1109/ICCC57789.2023.10165205","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165205","url":null,"abstract":"Medical image diagnostics now benefit greatly from the use of deep convolutional neural networks. The CNN-based deep neural networks are extensively used in the medical classification tasks. Although deep learning algorithms have gained performance comparable to pathologists in interpreting whole slide images for the detection of tumours, very few researchers have explored the possibility of detecting urothelial carcinoma from microscopic images. In this study, we propose a novel deep learning model for urine cytology smear-based detection of urothelial cell cancer (UCC). The network is based on Inception architecture that can efficiently learn the features of varied size cells in the image and produced relatively high accuracy when compared to state-of-the-art techniques. The proposed technique is evaluated using a dataset that includes the cytology samples of 115 individuals, 59 of whom had UCC instances that were histologically confirmed and the remaining 56 benign cases were identified through routine cytology samples. The suggested approach offers 98.63% accuracy with fewer parameters. The method’s performance in terms of accuracy and parameter count is highly encouraging.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133402987","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165131
J. Vijaya, Muskan Jain, Nandita Yadav
Machine learning is developing swiftly as an everexpanding field. The development of the same is occurring rapidly and has made many theoretical breakthroughs in recent times. Due to its importance as a part of machine learning, intelligent optimization algorithms are expected to become increasingly. The exponential growth of data volume and the increase in model complexity present increasing challenges for machine learning optimization strategies. Numerous initiatives have been launched to improve machine learning optimization approaches or address optimization-related problems. Future optimization and machine-learning research can be guided by a detailed evaluation and analysis of optimization strategies from a machine-learning perspective. Machine learning uses a variety of optimization strategies, which makes it easier to compare and analyze how well they function in various situations. In this study, we analyze and contrast seven well-known bio-inspired data engineering techniques and their effectiveness. We apply these techniques to the Radar Returns from the Ionosphere data-set and assess the results with a range of assessment metrics.
{"title":"Comparison Analysis of Various Optimization Algorithms for Classification of Radar Returns from the Ionosphere","authors":"J. Vijaya, Muskan Jain, Nandita Yadav","doi":"10.1109/ICCC57789.2023.10165131","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165131","url":null,"abstract":"Machine learning is developing swiftly as an everexpanding field. The development of the same is occurring rapidly and has made many theoretical breakthroughs in recent times. Due to its importance as a part of machine learning, intelligent optimization algorithms are expected to become increasingly. The exponential growth of data volume and the increase in model complexity present increasing challenges for machine learning optimization strategies. Numerous initiatives have been launched to improve machine learning optimization approaches or address optimization-related problems. Future optimization and machine-learning research can be guided by a detailed evaluation and analysis of optimization strategies from a machine-learning perspective. Machine learning uses a variety of optimization strategies, which makes it easier to compare and analyze how well they function in various situations. In this study, we analyze and contrast seven well-known bio-inspired data engineering techniques and their effectiveness. We apply these techniques to the Radar Returns from the Ionosphere data-set and assess the results with a range of assessment metrics.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121756303","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165496
Suhailam P, Raju Yerolla, C. Besta
Beer has existed for centuries. Fermentation is necessary. Over time, fermentation techniques have evolved. Despite having identical ingredients, each beer is unique. Beer production requires a source of carbs and yeast. Microbes aid brewers from the manufacture of raw materials to packaging stability. Some people overlook beer because it is made through the fermentation of food. Esters, acids, and higher alcohols are all susceptible to the effects of temperature. Fermentation temperatures can boost acidity and fruitiness. The influence of temperature on them is modeled using MATLAB/Simulink, and the temperature profile of an industrial environment is incorporated into the resulting model. The dynamic model is used to determine appropriate controller settings for optimal control, and the PID and MPC controllers are employed to obtain recognizable temperature profiles for flavor development.
{"title":"Modeling and advanced control strategies for the beer fermentation process","authors":"Suhailam P, Raju Yerolla, C. Besta","doi":"10.1109/ICCC57789.2023.10165496","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165496","url":null,"abstract":"Beer has existed for centuries. Fermentation is necessary. Over time, fermentation techniques have evolved. Despite having identical ingredients, each beer is unique. Beer production requires a source of carbs and yeast. Microbes aid brewers from the manufacture of raw materials to packaging stability. Some people overlook beer because it is made through the fermentation of food. Esters, acids, and higher alcohols are all susceptible to the effects of temperature. Fermentation temperatures can boost acidity and fruitiness. The influence of temperature on them is modeled using MATLAB/Simulink, and the temperature profile of an industrial environment is incorporated into the resulting model. The dynamic model is used to determine appropriate controller settings for optimal control, and the PID and MPC controllers are employed to obtain recognizable temperature profiles for flavor development.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128839711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165147
Reshmi V, M. Ebenezer
Renewable energy sources seem more feasible and cost-effective than power plants using fossil fuels for power generation. Maintaining an appropriate power balance between the source and the demand is almost impossible since renewable energy sources like wind and solar are unpredictable and have fluctuating loads. Microgrids are being elevated as an effective technique for combining small-scale distributed energy sources to produce electricity at the distribution voltage level. Because of the nature of the energy formed, some of the energy produced from these sources cannot be used immediately. Power electronic interfaces are required for the flexible, secure and reliable functioning of distribution systems between micro sources. Dealing with power quality concerns is among the main technical difficulties in the operation and control of microgrid systems that are either grid-connected or stand-alone. These significant issues are mostly caused by the design, mode of operation, nature and performance of distributed energy sources in the microgrid system. Current and voltage harmonics, voltage sag/swell, fluctuation and unbalance are the main power quality problems brought on by the large penetration of distributed generators, nonlinear and unbalanced loads. This study intends to analyze major power quality challenges and the control mechanisms that have been developed for enhancing the power quality in microgrids.
{"title":"Power Quality Enhancement Techniques in Microgrid-A Review","authors":"Reshmi V, M. Ebenezer","doi":"10.1109/ICCC57789.2023.10165147","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165147","url":null,"abstract":"Renewable energy sources seem more feasible and cost-effective than power plants using fossil fuels for power generation. Maintaining an appropriate power balance between the source and the demand is almost impossible since renewable energy sources like wind and solar are unpredictable and have fluctuating loads. Microgrids are being elevated as an effective technique for combining small-scale distributed energy sources to produce electricity at the distribution voltage level. Because of the nature of the energy formed, some of the energy produced from these sources cannot be used immediately. Power electronic interfaces are required for the flexible, secure and reliable functioning of distribution systems between micro sources. Dealing with power quality concerns is among the main technical difficulties in the operation and control of microgrid systems that are either grid-connected or stand-alone. These significant issues are mostly caused by the design, mode of operation, nature and performance of distributed energy sources in the microgrid system. Current and voltage harmonics, voltage sag/swell, fluctuation and unbalance are the main power quality problems brought on by the large penetration of distributed generators, nonlinear and unbalanced loads. This study intends to analyze major power quality challenges and the control mechanisms that have been developed for enhancing the power quality in microgrids.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129586358","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165168
S. S, D. S
Underwater image processing has been an active research topic over the past few years as interest in marine observation and the use of ocean resources has increased. Different from conventional images, marine ecosystems are frequently subjected to challenging conditions such as underwater turbulence, low contrast, and high colour distortion as a result of the light's non-uniform attenuation as it passes through the water. To overcome these challenges, a good amount of work in conventional and deep learning based underwater image processing has been published over a period of time. Deep learning has demonstrated excellent performance improvement than the conventional approaches on the challenging vision tasks. In this survey, important underwater image processing methods using deep learning have been discussed. The major underwater metrics, common datasets, and challenges are also presented.
{"title":"A Comprehensive Analysis of Underwater Image Processing based on Deep Learning Techniques","authors":"S. S, D. S","doi":"10.1109/ICCC57789.2023.10165168","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165168","url":null,"abstract":"Underwater image processing has been an active research topic over the past few years as interest in marine observation and the use of ocean resources has increased. Different from conventional images, marine ecosystems are frequently subjected to challenging conditions such as underwater turbulence, low contrast, and high colour distortion as a result of the light's non-uniform attenuation as it passes through the water. To overcome these challenges, a good amount of work in conventional and deep learning based underwater image processing has been published over a period of time. Deep learning has demonstrated excellent performance improvement than the conventional approaches on the challenging vision tasks. In this survey, important underwater image processing methods using deep learning have been discussed. The major underwater metrics, common datasets, and challenges are also presented.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117144917","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165227
G. P, A. P, Gayathri Vinayan, Gokuldath G, Ponmalar M, Aswini S H
Optical flow is a powerful application of image processing that is used in a variety of applications, primarily in object tracking and motion estimation. In this paper, we implement a system for vehicle motion tracking and velocity estimation using Lucas-Kanade (L-K) algorithm based optical flow method. The work includes two applications of optical flow: tracking the movement of the vehicle in the case of a fixed camera and velocity estimation of a vehicle with a camera mounted on it. Pre-processing steps include gaussian smoothing, and computing spatial and temporal gradients. This is followed by the further formulation of Lucas kanade equation in the form of matrices. The system of equations is then solved using the least square error criteria, and the flow vectors are obtained. Processes such as segmentation, blob analysis, camera calibration, and thresholding are further done which are used for velocity estimation as well as motion tracking. The functionality was tested and verified on video sequences obtained from the lab testing scenarios and real-world camera visuals taken from various sources.
{"title":"Lucas Kanade based Optical Flow for Vehicle Motion Tracking and Velocity Estimation","authors":"G. P, A. P, Gayathri Vinayan, Gokuldath G, Ponmalar M, Aswini S H","doi":"10.1109/ICCC57789.2023.10165227","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165227","url":null,"abstract":"Optical flow is a powerful application of image processing that is used in a variety of applications, primarily in object tracking and motion estimation. In this paper, we implement a system for vehicle motion tracking and velocity estimation using Lucas-Kanade (L-K) algorithm based optical flow method. The work includes two applications of optical flow: tracking the movement of the vehicle in the case of a fixed camera and velocity estimation of a vehicle with a camera mounted on it. Pre-processing steps include gaussian smoothing, and computing spatial and temporal gradients. This is followed by the further formulation of Lucas kanade equation in the form of matrices. The system of equations is then solved using the least square error criteria, and the flow vectors are obtained. Processes such as segmentation, blob analysis, camera calibration, and thresholding are further done which are used for velocity estimation as well as motion tracking. The functionality was tested and verified on video sequences obtained from the lab testing scenarios and real-world camera visuals taken from various sources.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131534901","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10164867
Anagha Anil, V. R. Jisha
Over the years, there has been a substantial increase in the number of vehicular traffic, which has led to vital problems like car crashes and congestion. More than 90 percent of collisions are the result of human error. Technology that allows for autonomous driving has the potential to enhance traffic efficiency and safety. Based on knowledge about the nearby traffic, an autonomous vehicle can create a trajectory and follow it using control algorithms. A significant technology in the study and implementation of autonomous vehicles is trajectory tracking control. Paths are a series of instructions that provide directional directives to get to a specific location, whereas a trajectory includes the schedule of velocity and higher order words, such as acceleration in terms of the body’s longitudinal and lateral motion, that are necessary to reach there. In this study, PID controllers and model predictive controllers (MPC) are used to govern the trajectory of an autonomous vehicle. The performance of the autonomous vehicle using both the controllers are then compared. The work is validated using simulations on MATLAB simulink.
{"title":"Trajectory Tracking Control of an Autonomous Vehicle using Model Predictive Control and PID Controller","authors":"Anagha Anil, V. R. Jisha","doi":"10.1109/ICCC57789.2023.10164867","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10164867","url":null,"abstract":"Over the years, there has been a substantial increase in the number of vehicular traffic, which has led to vital problems like car crashes and congestion. More than 90 percent of collisions are the result of human error. Technology that allows for autonomous driving has the potential to enhance traffic efficiency and safety. Based on knowledge about the nearby traffic, an autonomous vehicle can create a trajectory and follow it using control algorithms. A significant technology in the study and implementation of autonomous vehicles is trajectory tracking control. Paths are a series of instructions that provide directional directives to get to a specific location, whereas a trajectory includes the schedule of velocity and higher order words, such as acceleration in terms of the body’s longitudinal and lateral motion, that are necessary to reach there. In this study, PID controllers and model predictive controllers (MPC) are used to govern the trajectory of an autonomous vehicle. The performance of the autonomous vehicle using both the controllers are then compared. The work is validated using simulations on MATLAB simulink.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131607120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165171
Sabooj Ray, S. S.
Accurate measurement of pressure using piezoresistive MEMS sensors requires proper quantification and correction of its errors. Algorithmic temperature compensation of such a pressure sensor gives an accuracy that is one order better than resistive-compensation, but it is a time-consuming process. This paper discusses a MATLAB-based method to extract the correction coefficients for a MEMS pressure sensor and evaluates the extent of their effectiveness before they are fed into the system and experimented in hardware. The contribution of non-linearity error is analyzed and reduced. A method to improve the match between sensors as required in some aerospace applications is also discussed.
{"title":"Accuracy improvement for MEMS Piezoresistive Pressure Sensors","authors":"Sabooj Ray, S. S.","doi":"10.1109/ICCC57789.2023.10165171","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165171","url":null,"abstract":"Accurate measurement of pressure using piezoresistive MEMS sensors requires proper quantification and correction of its errors. Algorithmic temperature compensation of such a pressure sensor gives an accuracy that is one order better than resistive-compensation, but it is a time-consuming process. This paper discusses a MATLAB-based method to extract the correction coefficients for a MEMS pressure sensor and evaluates the extent of their effectiveness before they are fed into the system and experimented in hardware. The contribution of non-linearity error is analyzed and reduced. A method to improve the match between sensors as required in some aerospace applications is also discussed.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"2010 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127356360","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-05-19DOI: 10.1109/ICCC57789.2023.10165135
Shilna E., Athira Vinod, Jeena R. S., Anurenjan P. R., S. G.
A brain tumour is an abnormality in brain tissue that may cause harm to the nervous system and in severe cases can lead to death. Being a disease with a high mortality rate, the detection and accurate segmentation of brain tumour sub-regions is a crucial task in the disease diagnosis and treatment procedure. The manual segmentation process requires anatomical knowledge, is expensive, time-consuming, and inaccurate due to human errors. So, automatic and reliable segmentation methods are required; however, the large spatial and structural variability among brain tumours make automatic segmentation a challenging problem. This work proposes Reswave-Net, a deep learning network using an encoder-decoder (U-Net) architecture with residual connections to automate and standardize the task of tumour segmentation, which also incorporates wavelet decomposition of the input images. The network is trained and evaluated on Brain Tumour Segmentation (BraTS) Challenge-2020 dataset and achieves a mean Dice Score of 87.36%, 70.45%, and 72.55% and the Hausdorff distance of 6.87, 34.16 and 23.42 for the whole tumour, enhancing tumour and tumour core, respectively. For overall survival prediction, a random forest model is used where the radiomic features extracted from the image and age of the subject are used for training. The model achieves an accuracy of 58.4%.
{"title":"Reswave-Net: A wavelet based Residual U-Net for Brain Tumour Segmentation and Overall Survival Prediction","authors":"Shilna E., Athira Vinod, Jeena R. S., Anurenjan P. R., S. G.","doi":"10.1109/ICCC57789.2023.10165135","DOIUrl":"https://doi.org/10.1109/ICCC57789.2023.10165135","url":null,"abstract":"A brain tumour is an abnormality in brain tissue that may cause harm to the nervous system and in severe cases can lead to death. Being a disease with a high mortality rate, the detection and accurate segmentation of brain tumour sub-regions is a crucial task in the disease diagnosis and treatment procedure. The manual segmentation process requires anatomical knowledge, is expensive, time-consuming, and inaccurate due to human errors. So, automatic and reliable segmentation methods are required; however, the large spatial and structural variability among brain tumours make automatic segmentation a challenging problem. This work proposes Reswave-Net, a deep learning network using an encoder-decoder (U-Net) architecture with residual connections to automate and standardize the task of tumour segmentation, which also incorporates wavelet decomposition of the input images. The network is trained and evaluated on Brain Tumour Segmentation (BraTS) Challenge-2020 dataset and achieves a mean Dice Score of 87.36%, 70.45%, and 72.55% and the Hausdorff distance of 6.87, 34.16 and 23.42 for the whole tumour, enhancing tumour and tumour core, respectively. For overall survival prediction, a random forest model is used where the radiomic features extracted from the image and age of the subject are used for training. The model achieves an accuracy of 58.4%.","PeriodicalId":192909,"journal":{"name":"2023 International Conference on Control, Communication and Computing (ICCC)","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115895282","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：