Pub Date : 2023-07-02DOI: 10.1109/SSP53291.2023.10208005
P. Sherman
In this work we first review and elaborate on the eigenstructure of the covariance matrix for an autoregressive process of order 1. We then address the statistical elements related to its estimator in relation to the maximum eigenvalue. Bias, uncertainty, and distributions are provided in relation to the estimators of the various parameters associated with both the eigenvalue and eigenvector.
{"title":"On the Eigenstructure of the AR(1) Covariance","authors":"P. Sherman","doi":"10.1109/SSP53291.2023.10208005","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208005","url":null,"abstract":"In this work we first review and elaborate on the eigenstructure of the covariance matrix for an autoregressive process of order 1. We then address the statistical elements related to its estimator in relation to the maximum eigenvalue. Bias, uncertainty, and distributions are provided in relation to the estimators of the various parameters associated with both the eigenvalue and eigenvector.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125856779","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208029
O. Y. Chén, Duy Thanh Vu, Gilbert Greub, H. Cao, Xingru He, Yannick Muller, Constantinos Petrovas, H. Shou, Viet-Dung Nguyen, Bangdong Zhi, Laurent Perez, J. Raisaro, G. Nagels, M. Vos, Wei He, R. Gottardo, Palie Smart, M. Munafo, Giuseppe Pantaleo
We present here a systematical approach to studying the varying brain. We first distinguish different types of brain variability and provide examples for them. Next, we show classical analysis of covariance (ANCOVA) as well as advanced residual analysis via statistical- and deep-learning aim to decompose the total variance of the brain or behaviour data into explainable variance components. Additionally, we discuss innate and acquired brain variability. For varying big brain data, we define the neural law of large numbers and discuss methods for extracting representations from large-scale, potentially high-dimensional brain data. Finally, we examine the gut-brain axis, an often lurking, yet important, source of brain variability.
{"title":"The Statistical Analysis of the Varying Brain","authors":"O. Y. Chén, Duy Thanh Vu, Gilbert Greub, H. Cao, Xingru He, Yannick Muller, Constantinos Petrovas, H. Shou, Viet-Dung Nguyen, Bangdong Zhi, Laurent Perez, J. Raisaro, G. Nagels, M. Vos, Wei He, R. Gottardo, Palie Smart, M. Munafo, Giuseppe Pantaleo","doi":"10.1109/SSP53291.2023.10208029","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208029","url":null,"abstract":"We present here a systematical approach to studying the varying brain. We first distinguish different types of brain variability and provide examples for them. Next, we show classical analysis of covariance (ANCOVA) as well as advanced residual analysis via statistical- and deep-learning aim to decompose the total variance of the brain or behaviour data into explainable variance components. Additionally, we discuss innate and acquired brain variability. For varying big brain data, we define the neural law of large numbers and discuss methods for extracting representations from large-scale, potentially high-dimensional brain data. Finally, we examine the gut-brain axis, an often lurking, yet important, source of brain variability.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123527568","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208001
Fraser Williams, D. Jayalath, Anju Jose Tom, Terrence Martin, C. Fookes
Radio emitter localization methods have traditionally incorporated many sources of information such as time difference of arrival (TDOA), angle of arrival (AOA), and received signal strength (RSS) into a two-step position estimation model. Modern direct position determination (DPD) methods have since superseded the performance of two-step methods in low signal-to-noise ratio (SNR) environments. However, the current DPD literature has neglected the use of RSS information to enhance localization accuracy, despite its prevalence in predecessor two-step methods. As signal strength information is always present at receiver nodes, regardless of operating hardware, this information could be used to better estimate emitter position. We propose an RSS method as applied to spatially distributed receiver arrays incorporating beamforming. Monte Carlo simulations show improved accuracy at medium to high SNR as compared to methods exploiting only time and angle information, while having reduced performance at very low SNR.
{"title":"Enhancing Emitter Localization Accuracy Through Integration of Received Signal Strength in Direct Position Determination","authors":"Fraser Williams, D. Jayalath, Anju Jose Tom, Terrence Martin, C. Fookes","doi":"10.1109/SSP53291.2023.10208001","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208001","url":null,"abstract":"Radio emitter localization methods have traditionally incorporated many sources of information such as time difference of arrival (TDOA), angle of arrival (AOA), and received signal strength (RSS) into a two-step position estimation model. Modern direct position determination (DPD) methods have since superseded the performance of two-step methods in low signal-to-noise ratio (SNR) environments. However, the current DPD literature has neglected the use of RSS information to enhance localization accuracy, despite its prevalence in predecessor two-step methods. As signal strength information is always present at receiver nodes, regardless of operating hardware, this information could be used to better estimate emitter position. We propose an RSS method as applied to spatially distributed receiver arrays incorporating beamforming. Monte Carlo simulations show improved accuracy at medium to high SNR as compared to methods exploiting only time and angle information, while having reduced performance at very low SNR.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123760902","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208065
Chenyu Gao, Ziping Zhao, D. Palomar
Generalized autoregressive conditional heteroskedasticity (GARCH) is a popular model to describe the time-varying conditional volatility of a time series, which is widely used in signal processing and machine learning. In this paper, we focus on the model parameter estimation of GARCH based on the Gaussian maximum likelihood estimation method. Due to the recursively coupling nature of parameters in GARCH, the optimization problem is highly non-convex. In this paper, we propose a novel algorithm based on the block majorization-minimization algorithmic framework, which can take care of the per-block variable structures for efficient problem solving. Numerical experiments demonstrate that the proposed algorithm can achieve comparable and even better performance in terms of parameter estimation errors. More importantly, estimated parameters from our algorithm always guarantee a stationary model, which is a desirable property in time series volatility modeling.
{"title":"A Novel Algorithm for GARCH Model Estimation","authors":"Chenyu Gao, Ziping Zhao, D. Palomar","doi":"10.1109/SSP53291.2023.10208065","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208065","url":null,"abstract":"Generalized autoregressive conditional heteroskedasticity (GARCH) is a popular model to describe the time-varying conditional volatility of a time series, which is widely used in signal processing and machine learning. In this paper, we focus on the model parameter estimation of GARCH based on the Gaussian maximum likelihood estimation method. Due to the recursively coupling nature of parameters in GARCH, the optimization problem is highly non-convex. In this paper, we propose a novel algorithm based on the block majorization-minimization algorithmic framework, which can take care of the per-block variable structures for efficient problem solving. Numerical experiments demonstrate that the proposed algorithm can achieve comparable and even better performance in terms of parameter estimation errors. More importantly, estimated parameters from our algorithm always guarantee a stationary model, which is a desirable property in time series volatility modeling.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129593548","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208022
Meesam Jafri, Sana Anwer, Suraj Srivastava, A. Jagannatham
This paper considers a joint radar and communication (JRC) system towards radar cross-section (RCS) parameter and channel estimation in millimeter wave (mmWave) multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems. The proposed algorithms are based on the hybrid mmWave MIMO architecture. First, the orthogonal matching pursuit (OMP)-based framework is conceived for radar target parameter estimation. Next, a novel multiple measurement vector (MMV)-based Bayesian learning (MBL) algorithm is proposed for mmWave MIMO channel estimation in JRC systems. Subsequently, these quantities are employed at the dual-functional radar-communication (DFRC) base station (BS) and at the user equipment (UE) toward successful data transmission and detection, respectively. The proposed techniques exploit the sparsity inherent in the radar scattering environment and the simultaneous sparsity of the wireless channel across all the subcarriers for improved performance. Numerical results demonstrate the efficacy of the proposed techniques and the improved performance in comparison to existing sparse recovery techniques as well as the conventional non-sparse parameter estimation algorithms.
{"title":"Sparse Estimation in mmWave MIMO-OFDM Joint Radar and Communication (JRC) Systems","authors":"Meesam Jafri, Sana Anwer, Suraj Srivastava, A. Jagannatham","doi":"10.1109/SSP53291.2023.10208022","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208022","url":null,"abstract":"This paper considers a joint radar and communication (JRC) system towards radar cross-section (RCS) parameter and channel estimation in millimeter wave (mmWave) multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems. The proposed algorithms are based on the hybrid mmWave MIMO architecture. First, the orthogonal matching pursuit (OMP)-based framework is conceived for radar target parameter estimation. Next, a novel multiple measurement vector (MMV)-based Bayesian learning (MBL) algorithm is proposed for mmWave MIMO channel estimation in JRC systems. Subsequently, these quantities are employed at the dual-functional radar-communication (DFRC) base station (BS) and at the user equipment (UE) toward successful data transmission and detection, respectively. The proposed techniques exploit the sparsity inherent in the radar scattering environment and the simultaneous sparsity of the wireless channel across all the subcarriers for improved performance. Numerical results demonstrate the efficacy of the proposed techniques and the improved performance in comparison to existing sparse recovery techniques as well as the conventional non-sparse parameter estimation algorithms.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133045191","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10207992
H. M. Tran, K. T. Pham, Thanh M. Vo, L. T. That, T. T. M. Huynh, S. Dao
The physical characteristics of agricultural products are crucial for developing grading, sizing, and packaging systems. So that, accurately measuring irregularly shaped products like starfruit is a challenging task. This paper proposes a technique that two cameras are used to estimate the dimensions, volume, and mass of starfruit with high accuracy. Firstly, top-view and body-view images of the starfruit are captured, and image processing techniques, conical frustum method are employed to find the volume based on the area ratio of star shape area over its bounding box and volume of multiple pieces along the longitudinal axis. Then, the density of the starfruit is used to estimate its mass. The proposed method has been validated with a highest average accuracy of 99.16% for the volume and 98.59% mass using 255 training samples. This technology is simple to adopt in starfruit and other fruit manufacturing lines
{"title":"Physical Characteristics Estimation for Irregularly Shaped Fruit Using Two Cameras","authors":"H. M. Tran, K. T. Pham, Thanh M. Vo, L. T. That, T. T. M. Huynh, S. Dao","doi":"10.1109/SSP53291.2023.10207992","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10207992","url":null,"abstract":"The physical characteristics of agricultural products are crucial for developing grading, sizing, and packaging systems. So that, accurately measuring irregularly shaped products like starfruit is a challenging task. This paper proposes a technique that two cameras are used to estimate the dimensions, volume, and mass of starfruit with high accuracy. Firstly, top-view and body-view images of the starfruit are captured, and image processing techniques, conical frustum method are employed to find the volume based on the area ratio of star shape area over its bounding box and volume of multiple pieces along the longitudinal axis. Then, the density of the starfruit is used to estimate its mass. The proposed method has been validated with a highest average accuracy of 99.16% for the volume and 98.59% mass using 255 training samples. This technology is simple to adopt in starfruit and other fruit manufacturing lines","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114701486","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10207983
Jad Abou Chaaya, Batoul Zaraket, Hassan Harb, A. Mansour
After the invasion of the Covid-19 virus, governments started containing the spread of the virus by forcing people to wear face masks in public places. Therefore, automatic face mask detection has become very important to limit the virus spread. Unfortunately, existing methods present limited performance in accurately detecting masks in crowded areas due to the significant number of faces per scene. In order to tackle this challenge, we propose a two-stage neural network-based architecture that can accurately detect face masks in crowded environments. Several simulations have been conducted to investigate the efficiency of the proposed architecture and the results show a high accuracy of detection that can reach up to 96.5%.
{"title":"Convolutional Neural Network-based Architecture for Detecting Face Mask in Crowded Areas","authors":"Jad Abou Chaaya, Batoul Zaraket, Hassan Harb, A. Mansour","doi":"10.1109/SSP53291.2023.10207983","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10207983","url":null,"abstract":"After the invasion of the Covid-19 virus, governments started containing the spread of the virus by forcing people to wear face masks in public places. Therefore, automatic face mask detection has become very important to limit the virus spread. Unfortunately, existing methods present limited performance in accurately detecting masks in crowded areas due to the significant number of faces per scene. In order to tackle this challenge, we propose a two-stage neural network-based architecture that can accurately detect face masks in crowded environments. Several simulations have been conducted to investigate the efficiency of the proposed architecture and the results show a high accuracy of detection that can reach up to 96.5%.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"52 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131980930","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10207954
Kommineni Jenni, M. Srinivas, Roshni Sannapu, Murukessan Perumal
Convolutional networks are a key component of many computer vision applications. However, convolutions have a serious flaw. It only works in a small area, hence it lacks global information. The Attention method, on the other hand, is a new improvement in capturing long range interactions that has mostly been used to sequence modeling and generative modeling tasks. As an alternative to convolutions, we investigate the use of convolutions with an attention mechanism in a video question answering task. We present a unique self-attention mechanism based on convolutions that outperforms convolutions in the video question answering task. We discovered that combining convolutions with self-attention produces the greatest outcomes in experiments. As a result, we propose a hybrid idea, which combines convolutional operators with the self-attention mechanism. We combine convolutional feature maps with self-attention feature maps. Experiments show that convolution with self-attention improves video question answering tasks on the MSRVTT-QA dataset.
{"title":"CSA-BERT: Video Question Answering","authors":"Kommineni Jenni, M. Srinivas, Roshni Sannapu, Murukessan Perumal","doi":"10.1109/SSP53291.2023.10207954","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10207954","url":null,"abstract":"Convolutional networks are a key component of many computer vision applications. However, convolutions have a serious flaw. It only works in a small area, hence it lacks global information. The Attention method, on the other hand, is a new improvement in capturing long range interactions that has mostly been used to sequence modeling and generative modeling tasks. As an alternative to convolutions, we investigate the use of convolutions with an attention mechanism in a video question answering task. We present a unique self-attention mechanism based on convolutions that outperforms convolutions in the video question answering task. We discovered that combining convolutions with self-attention produces the greatest outcomes in experiments. As a result, we propose a hybrid idea, which combines convolutional operators with the self-attention mechanism. We combine convolutional feature maps with self-attention feature maps. Experiments show that convolution with self-attention improves video question answering tasks on the MSRVTT-QA dataset.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"123 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134377193","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208078
Minh Le Nguyen, Tinh T. Bui, L. Nguyen, E. Garcia-Palacios, H. Zepernick, T. Duong
In this paper, we consider an Internet-of-Things network supported by several satellites and multiple cache-assisted unmanned aerial vehicles (UAVs). We propose an optimisation problem with the aim of minimising the total network latency. To reduce the complexity of the original problem, it is divided into three sub-problems, namely, clustering ground users associated with UAVs, cache placement in UAVs (to support the network in avoiding backhaul congestion), and power allocation for satellites and UAVs. A non-cooperative game is designed to obtain the solution to the clustering problem; a genetic algorithm, which is powerful in the scenario of many variables, is employed to obtain the optimal solution to the high-complexity caching problem; and a quick estimation technique is used for power allocation. The total network latency is then minimised by using alternating optimisation technique. Numerical results prove the efficiency of our methods compared to other traditional ones.
{"title":"Real-Time Large-Scale 6G Satellite-UAV Networks","authors":"Minh Le Nguyen, Tinh T. Bui, L. Nguyen, E. Garcia-Palacios, H. Zepernick, T. Duong","doi":"10.1109/SSP53291.2023.10208078","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208078","url":null,"abstract":"In this paper, we consider an Internet-of-Things network supported by several satellites and multiple cache-assisted unmanned aerial vehicles (UAVs). We propose an optimisation problem with the aim of minimising the total network latency. To reduce the complexity of the original problem, it is divided into three sub-problems, namely, clustering ground users associated with UAVs, cache placement in UAVs (to support the network in avoiding backhaul congestion), and power allocation for satellites and UAVs. A non-cooperative game is designed to obtain the solution to the clustering problem; a genetic algorithm, which is powerful in the scenario of many variables, is employed to obtain the optimal solution to the high-complexity caching problem; and a quick estimation technique is used for power allocation. The total network latency is then minimised by using alternating optimisation technique. Numerical results prove the efficiency of our methods compared to other traditional ones.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"144 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116078598","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208061
Q. A. Nguyen, Nhung T. C. Nguyen, Son Nguyen, Phuong T. K. Doan, N. H. Thinh, Tung H. Tran, A. L. T. Luong, Ha V. Le, H. M. Luu
Chromosomal image analysis is an important method to diagnose chromosomal disorders. However, the image can be curved after cultivation, resulting in difficulty in chromosome recognition and analyzing the bands. While manual work of straightening the chromosomes requires an intensive labor, the computer-aided method can increase the performance as well as preserve the image details. In this paper, we investigate a method of straightening the curved chromosomes using Spatial Transformer Network (SPN) and to what extend the method affects the chromosome classification using a CNN-based method. The experiments were carried on a dataset of 28,106 chromosome images. The results show that SPN achieved compatible performance to manual method on the curved chromosomes with straight ratio of higher than 90%, yielding improvements of the classification accuracy to that of the original curved images from 3% to 5% on average. The source code and processed data are shared to support further studies.
{"title":"Improving Classification of Curved Chromosomes in Karyotyping using CNN-based Deformation","authors":"Q. A. Nguyen, Nhung T. C. Nguyen, Son Nguyen, Phuong T. K. Doan, N. H. Thinh, Tung H. Tran, A. L. T. Luong, Ha V. Le, H. M. Luu","doi":"10.1109/SSP53291.2023.10208061","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208061","url":null,"abstract":"Chromosomal image analysis is an important method to diagnose chromosomal disorders. However, the image can be curved after cultivation, resulting in difficulty in chromosome recognition and analyzing the bands. While manual work of straightening the chromosomes requires an intensive labor, the computer-aided method can increase the performance as well as preserve the image details. In this paper, we investigate a method of straightening the curved chromosomes using Spatial Transformer Network (SPN) and to what extend the method affects the chromosome classification using a CNN-based method. The experiments were carried on a dataset of 28,106 chromosome images. The results show that SPN achieved compatible performance to manual method on the curved chromosomes with straight ratio of higher than 90%, yielding improvements of the classification accuracy to that of the original curved images from 3% to 5% on average. The source code and processed data are shared to support further studies.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123461380","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：