Pub Date : 2024-01-10DOI: 10.1088/2632-2153/ad1d05
Jizhen Cai, Clotilde Boust, Alamin Mansouri
Recently, the Convolutional Neural Network (CNN) has gained increasing importance in hyperspectral image classification thanks to its superior performance. However, most of the previous research has mainly focused on 2D-CNN, and the limited applications of 3D-CNN have been attributed to its complexity, despite its potential to enhance information extraction between adjacent channels of the image. Moreover, 1D-CNN is typically restricted to the field of signal processing as it ignores the spatial information of hyperspectral images. In this paper, we propose a novel CNN model named ATSFCNN (Attention-based Triple-Stream Fused Convolutional Neural Network) that fuses the features of 1D-CNN, 2D-CNN, and 3D-CNN to consider all the relevant information of the hyperspectral dataset. Our contributions are twofold: First, we propose a strategy to extract and homogenize features from 1D, 2D, and 3D CNN. Secondly, we propose a way to efficiently fuse these features. This attention-based methodology adeptly integrates features from the triple streams, thereby transcending the former limitations of singular stream utilization. Consequently, it becomes capable of attaining elevated outcomes in the context of hyperspectral classification, marked by increased levels of both accuracy and stability. We compared the results of ATSFCNN with those of other deep learning models, including 1D-CNN, 2D-CNN, 2D-CNN+PCA, 3D-CNN, and 3D-CNN+PCA, and demonstrated its superior performance and robustness. Quantitative assessments, predicated on the metrics of Overall Accuracy (OA), Average Accuracy (AA), and Kappa Coefficient (κ) emphatically corroborate the preeminence of ATSFCNN. Notably, spanning three remote sensing datasets, ATSFCNN consistently achieves peak levels of Overall Accuracy, quantified at 98.38%, 97.09%, and 96.93% respectively. This prowess is further accentuated by concomitant Average Accuracy scores of 98.47%, 95.80%, and 95.80%, as well as Kappa Coefficient values amounting to 97.41%, 96.14%, and 95.21%.
{"title":"ATSFCNN: A Novel Attention-based Triple-Stream Fused CNN Model for Hyperspectral Image Classification","authors":"Jizhen Cai, Clotilde Boust, Alamin Mansouri","doi":"10.1088/2632-2153/ad1d05","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1d05","url":null,"abstract":"\u0000 Recently, the Convolutional Neural Network (CNN) has gained increasing importance in hyperspectral image classification thanks to its superior performance. However, most of the previous research has mainly focused on 2D-CNN, and the limited applications of 3D-CNN have been attributed to its complexity, despite its potential to enhance information extraction between adjacent channels of the image. Moreover, 1D-CNN is typically restricted to the field of signal processing as it ignores the spatial information of hyperspectral images. In this paper, we propose a novel CNN model named ATSFCNN (Attention-based Triple-Stream Fused Convolutional Neural Network) that fuses the features of 1D-CNN, 2D-CNN, and 3D-CNN to consider all the relevant information of the hyperspectral dataset. Our contributions are twofold: First, we propose a strategy to extract and homogenize features from 1D, 2D, and 3D CNN. Secondly, we propose a way to efficiently fuse these features. This attention-based methodology adeptly integrates features from the triple streams, thereby transcending the former limitations of singular stream utilization. Consequently, it becomes capable of attaining elevated outcomes in the context of hyperspectral classification, marked by increased levels of both accuracy and stability. We compared the results of ATSFCNN with those of other deep learning models, including 1D-CNN, 2D-CNN, 2D-CNN+PCA, 3D-CNN, and 3D-CNN+PCA, and demonstrated its superior performance and robustness. Quantitative assessments, predicated on the metrics of Overall Accuracy (OA), Average Accuracy (AA), and Kappa Coefficient (κ) emphatically corroborate the preeminence of ATSFCNN. Notably, spanning three remote sensing datasets, ATSFCNN consistently achieves peak levels of Overall Accuracy, quantified at 98.38%, 97.09%, and 96.93% respectively. This prowess is further accentuated by concomitant Average Accuracy scores of 98.47%, 95.80%, and 95.80%, as well as Kappa Coefficient values amounting to 97.41%, 96.14%, and 95.21%.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"83 14","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139440810","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 : 2024-01-10DOI: 10.1088/2632-2153/ad1d06
Yu Zhao, Xiaoyan Shen, Jiadong Chen, Wei Qian, He Ma, Liang Sang
Purpose Medical image segmentation is essential to image-based disease analysis and has proven to be significantly helpful for doctors to make decisions. Due to the low-contrast of some medical images, the accurate segmentation of medical images has always been a challenging problem. The experiment found that UNet with current loss functions cannot capture subtle information in target contours or regions in low-contrast medical images, which are crucial for subsequent disease diagnosis. Methods We propose a robust loss by incorporating the difference in average radial derivative (ARD), length and region area to further help the network to achieve more accurate segmentation results. We evaluated the proposed loss function using UNet as the base segmentation network compared to five conventional loss functions on one private and four public medical image datasets. Results Experimental results illustrate that UNet with the proposed loss function can achieve the best segmentation performance, even better than the outstanding deep learning models with original loss functions. Furthermore, three representative datasets were chosen to validate the effectiveness of the proposed δARD loss function with seven different models. Conclusion The experiments revealed δARD loss's plug-and-play feature and its robustness over multiple models and datasets.
{"title":"δARD loss for low-contrast medical image segmentation","authors":"Yu Zhao, Xiaoyan Shen, Jiadong Chen, Wei Qian, He Ma, Liang Sang","doi":"10.1088/2632-2153/ad1d06","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1d06","url":null,"abstract":"\u0000 Purpose Medical image segmentation is essential to image-based disease analysis and has proven to be significantly helpful for doctors to make decisions. Due to the low-contrast of some medical images, the accurate segmentation of medical images has always been a challenging problem. The experiment found that UNet with current loss functions cannot capture subtle information in target contours or regions in low-contrast medical images, which are crucial for subsequent disease diagnosis. Methods We propose a robust loss by incorporating the difference in average radial derivative (ARD), length and region area to further help the network to achieve more accurate segmentation results. We evaluated the proposed loss function using UNet as the base segmentation network compared to five conventional loss functions on one private and four public medical image datasets. Results Experimental results illustrate that UNet with the proposed loss function can achieve the best segmentation performance, even better than the outstanding deep learning models with original loss functions. Furthermore, three representative datasets were chosen to validate the effectiveness of the proposed δARD loss function with seven different models. Conclusion The experiments revealed δARD loss's plug-and-play feature and its robustness over multiple models and datasets.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"4 12","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139439893","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 : 2024-01-08DOI: 10.1088/2632-2153/ad1c32
Chibuike Chiedozie Ibebuchi
Understanding the spatial patterns of the North Atlantic Oscillation (NAO) is vital for climate science. For this reason, empirical orthogonal function (EOF) analysis is commonly applied to sea-level pressure (SLP) anomaly data in the North Atlantic region. This study evaluated the traditional EOF-based definition of the NAO index against the Autoencoder (AE) neural network-based definition, using the Hurrell NAO Index (Station-Based) as a reference. Specifically, EOF and AE were applied to monthly SLP anomaly data from ERA5 (1950-2022) to derive spatial modes of variability in the North Atlantic region. Both methods produced spatial patterns consistent with the traditional NAO definition, with dipole centers of action between the Icelandic Low and the Azores High. During boreal winter (December to March), when the NAO is most active, the AE-based method achieved a correlation of 0.96 with the reference NAO index, outperforming the EOF-based method's correlation of 0.90. The all-season Adjusted R-squared values were 50% for the AE-based index and 34% for the EOF-based index. Notably, the AE-based index revealed several other non-linear patterns of the NAO, with more than one encoded pattern correlating at least 0.90 with the reference NAO index during boreal winter. These results not only demonstrate the AE's superiority over traditional EOF in representing the station-based index but also uncover previously unexplored complexities in the NAO that are close to the reference temporal pattern. This suggests that AE offers a promising approach for defining climate modes of variability, potentially capturing intricacies that traditional linear methods like EOF might miss.
{"title":"Redefining the North Atlantic Oscillation Index Generation using Autoencoder Neural Network","authors":"Chibuike Chiedozie Ibebuchi","doi":"10.1088/2632-2153/ad1c32","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1c32","url":null,"abstract":"\u0000 Understanding the spatial patterns of the North Atlantic Oscillation (NAO) is vital for climate science. For this reason, empirical orthogonal function (EOF) analysis is commonly applied to sea-level pressure (SLP) anomaly data in the North Atlantic region. This study evaluated the traditional EOF-based definition of the NAO index against the Autoencoder (AE) neural network-based definition, using the Hurrell NAO Index (Station-Based) as a reference. Specifically, EOF and AE were applied to monthly SLP anomaly data from ERA5 (1950-2022) to derive spatial modes of variability in the North Atlantic region. Both methods produced spatial patterns consistent with the traditional NAO definition, with dipole centers of action between the Icelandic Low and the Azores High. During boreal winter (December to March), when the NAO is most active, the AE-based method achieved a correlation of 0.96 with the reference NAO index, outperforming the EOF-based method's correlation of 0.90. The all-season Adjusted R-squared values were 50% for the AE-based index and 34% for the EOF-based index. Notably, the AE-based index revealed several other non-linear patterns of the NAO, with more than one encoded pattern correlating at least 0.90 with the reference NAO index during boreal winter. These results not only demonstrate the AE's superiority over traditional EOF in representing the station-based index but also uncover previously unexplored complexities in the NAO that are close to the reference temporal pattern. This suggests that AE offers a promising approach for defining climate modes of variability, potentially capturing intricacies that traditional linear methods like EOF might miss.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"48 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139448174","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 : 2024-01-08DOI: 10.1088/2632-2153/ad1c30
Pragyan Banerjee, Shivam Milind Akarte, Prakhar Kumar, Muhammad Shamsuzzaman, Ankit Butola, Krishna Agarwal, dilip kumar prasad, F. Melandsø, A. Habib
Acoustic microscopy is a cutting-edge label-free imaging technology that allows us to see the surface and interior structure of industrial and biological materials. The acoustic image is created by focusing high-frequency acoustic waves on the object and then detecting reflected signals. On the other hand, the quality of the acoustic image's resolution is influenced by the signal-to-noise ratio, the scanning step size, and the frequency of the transducer. Deep learning-based high-resolution imaging in acoustic microscopy is proposed in this paper. To illustrate 4 times resolution improvement in acoustic images, five distinct models are used: SRGAN, ESRGAN, IMDN, DBPN-RES-MR64-3, and SwinIR. The trained model's performance is assessed by calculating the PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index) between the network-predicted and ground truth images. To avoid the model from over-fitting, transfer learning was incorporated during the procedure. SwinIR had average SSIM and PSNR values of 0.95 and 35, respectively. The model was also evaluated using a biological sample from Reindeer Antler, yielding an SSIM score of 0.88 and a PSNR score of 32.93. Our framework is relevant to a wide range of industrial applications, including electronic production, material micro-structure analysis, and other biological applications in general.
{"title":"High-resolution imaging in acoustic microscopy using deep learning","authors":"Pragyan Banerjee, Shivam Milind Akarte, Prakhar Kumar, Muhammad Shamsuzzaman, Ankit Butola, Krishna Agarwal, dilip kumar prasad, F. Melandsø, A. Habib","doi":"10.1088/2632-2153/ad1c30","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1c30","url":null,"abstract":"\u0000 Acoustic microscopy is a cutting-edge label-free imaging technology that allows us to see the surface and interior structure of industrial and biological materials. The acoustic image is created by focusing high-frequency acoustic waves on the object and then detecting reflected signals. On the other hand, the quality of the acoustic image's resolution is influenced by the signal-to-noise ratio, the scanning step size, and the frequency of the transducer. Deep learning-based high-resolution imaging in acoustic microscopy is proposed in this paper. To illustrate 4 times resolution improvement in acoustic images, five distinct models are used: SRGAN, ESRGAN, IMDN, DBPN-RES-MR64-3, and SwinIR. The trained model's performance is assessed by calculating the PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index) between the network-predicted and ground truth images. To avoid the model from over-fitting, transfer learning was incorporated during the procedure. SwinIR had average SSIM and PSNR values of 0.95 and 35, respectively. The model was also evaluated using a biological sample from Reindeer Antler, yielding an SSIM score of 0.88 and a PSNR score of 32.93. Our framework is relevant to a wide range of industrial applications, including electronic production, material micro-structure analysis, and other biological applications in general.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"59 20","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139447702","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 : 2024-01-08DOI: 10.1088/2632-2153/ad1c34
Chibuike Chiedozie Ibebuchi, O. Obarein, Itohan-Osa Abu
Spatial regionalization is instrumental in simplifying the spatial complexity of the climate system. To identify regions of significant climate variability, pattern extraction is often required prior to spatial regionalization with a clustering algorithm. In this study, the autoencoder artificial neural network (AE) was applied to extract the inherent patterns of global temperature data (from 1901 to 2021). Subsequently, Fuzzy C-means clustering was applied to the extracted patterns to classify the global temperature regions. Our analysis involved comparing AE-based and principal component analysis (PCA)-based clustering results to assess consistency. We determined the number of clusters by examining the average percentage decrease in Fuzzy Partition Coefficient and its 95% confidence interval, seeking a balance between obtaining a high Fuzzy Partition Coefficient and avoiding over-segmentation. This approach suggested that for a more general model, four clusters is reasonable. The Adjusted Rand Index between the AE-based and PCA-based clusters is 0.75, indicating that the AE-based and PCA-based clusters have considerable overlap. The observed difference between the AE-based clusters and PCA-based clusters is suggested to be associated with AE’s capability to learn and extract complex non-linear patterns, and this attribute, for example, enabled the clustering algorithm to accurately detect the Himalayas region as the “third pole” with similar temperature characteristics as the polar regions. Finally, when the analysis period is divided into two (1901-1960 and 1961-2021), the Adjusted Rand Index between the two clusters is 0.96 which suggests that historical climate change has not significantly affected the defined temperature regions over the two periods. In essence, this study indicates both AE's potential to enhance our understanding of climate variability and reveals the stability of the historical temperature regions.
{"title":"Application of Autoencoders Artificial Neural Network and Principal Component Analysis for Pattern Extraction and Spatial Regionalization of Global Temperature Data","authors":"Chibuike Chiedozie Ibebuchi, O. Obarein, Itohan-Osa Abu","doi":"10.1088/2632-2153/ad1c34","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1c34","url":null,"abstract":"\u0000 Spatial regionalization is instrumental in simplifying the spatial complexity of the climate system. To identify regions of significant climate variability, pattern extraction is often required prior to spatial regionalization with a clustering algorithm. In this study, the autoencoder artificial neural network (AE) was applied to extract the inherent patterns of global temperature data (from 1901 to 2021). Subsequently, Fuzzy C-means clustering was applied to the extracted patterns to classify the global temperature regions. Our analysis involved comparing AE-based and principal component analysis (PCA)-based clustering results to assess consistency. We determined the number of clusters by examining the average percentage decrease in Fuzzy Partition Coefficient and its 95% confidence interval, seeking a balance between obtaining a high Fuzzy Partition Coefficient and avoiding over-segmentation. This approach suggested that for a more general model, four clusters is reasonable. The Adjusted Rand Index between the AE-based and PCA-based clusters is 0.75, indicating that the AE-based and PCA-based clusters have considerable overlap. The observed difference between the AE-based clusters and PCA-based clusters is suggested to be associated with AE’s capability to learn and extract complex non-linear patterns, and this attribute, for example, enabled the clustering algorithm to accurately detect the Himalayas region as the “third pole” with similar temperature characteristics as the polar regions. Finally, when the analysis period is divided into two (1901-1960 and 1961-2021), the Adjusted Rand Index between the two clusters is 0.96 which suggests that historical climate change has not significantly affected the defined temperature regions over the two periods. In essence, this study indicates both AE's potential to enhance our understanding of climate variability and reveals the stability of the historical temperature regions.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"19 26","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139445266","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 : 2024-01-08DOI: 10.1088/2632-2153/ad1c31
M. Pathan, Suresh M Pradhan, T. P. Selvam, Balvinder Kaur Sapra
In the present age, marked by data-driven advancements in various fields, the importance of machine learning holds a prominent position. The ability of machine learning algorithms to resolve complex patterns and extract insights from large datasets has solidified its transformative potential in various scientific domains. This paper introduces an innovative application of machine learning techniques in the domain of radiation dosimetry. Specifically, it shows the applicability of machine learning in estimating the radiation dose received by occupational workers. This estimation is expressed in terms of personal dose equivalent, and it involves the utilization of thermoluminescence signals emitted by CaSO4:Dy–based personnel monitoring badges. To estimate personal dose equivalent, three-stage algorithm driven by machine learning models is proposed. This algorithm systematically identifies the photon energy ranges, calculates the average photon energy, and determines personal dose equivalent. By implementing this approach to the conventional three-element dosimeter, the study overcomes existing limitations and enhances accuracy in dose estimation. The algorithm demonstrates 97.8% classification accuracy in discerning photon energy ranges and achieves a coefficient of determination of 0.988 for estimating average photon energy. Importantly, it also reduces the coefficient of variation of relative deviations by up to 6% for estimated personal dose equivalent, compared to existing algorithms. The study improves accuracy and establishes a new methodology for evaluating radiation exposure to occupational workers using conventional thermoluminescent dosimeter badge.
{"title":"A Multi-Stage Machine Learning Algorithm for Estimating Personal Dose Equivalent using Thermoluminescent Dosimeter","authors":"M. Pathan, Suresh M Pradhan, T. P. Selvam, Balvinder Kaur Sapra","doi":"10.1088/2632-2153/ad1c31","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1c31","url":null,"abstract":"\u0000 In the present age, marked by data-driven advancements in various fields, the importance of machine learning holds a prominent position. The ability of machine learning algorithms to resolve complex patterns and extract insights from large datasets has solidified its transformative potential in various scientific domains. This paper introduces an innovative application of machine learning techniques in the domain of radiation dosimetry. Specifically, it shows the applicability of machine learning in estimating the radiation dose received by occupational workers. This estimation is expressed in terms of personal dose equivalent, and it involves the utilization of thermoluminescence signals emitted by CaSO4:Dy–based personnel monitoring badges. To estimate personal dose equivalent, three-stage algorithm driven by machine learning models is proposed. This algorithm systematically identifies the photon energy ranges, calculates the average photon energy, and determines personal dose equivalent. By implementing this approach to the conventional three-element dosimeter, the study overcomes existing limitations and enhances accuracy in dose estimation. The algorithm demonstrates 97.8% classification accuracy in discerning photon energy ranges and achieves a coefficient of determination of 0.988 for estimating average photon energy. Importantly, it also reduces the coefficient of variation of relative deviations by up to 6% for estimated personal dose equivalent, compared to existing algorithms. The study improves accuracy and establishes a new methodology for evaluating radiation exposure to occupational workers using conventional thermoluminescent dosimeter badge.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"41 15","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139447575","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 : 2024-01-08DOI: 10.1088/2632-2153/ad1c33
Jens Grauer, F. J. Schwarzendahl, Hartmut Löwen, B. Liebchen
Bacteria and other self-propelling microorganisms produce and respond to signaling molecules to communicate with each other (quorum sensing) and to direct their collective behavior. Here, we explore agents (active particles) which communicate with each other to coordinate their collective dynamics to optimize nutrient consumption. Using reinforcement learning and neural networks, we identify three different strategies: a "clustering strategy", where the agents accumulate in regions of high nutrient concentration; a "spreading strategy", where particles stay away from each other to avoid competing for sparse resources; and an "adaptive strategy", where the agents adaptively decide to either follow or stay away from others. Our work exemplifies the idea that machine learning can be used to determine parameters that are evolutionarily optimized in biological systems but often occur as unknown parameters in mathematical models describing their dynamics.
{"title":"Optimizing collective behavior of communicating active particles with machine learning","authors":"Jens Grauer, F. J. Schwarzendahl, Hartmut Löwen, B. Liebchen","doi":"10.1088/2632-2153/ad1c33","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1c33","url":null,"abstract":"\u0000 Bacteria and other self-propelling microorganisms produce and respond to signaling molecules to communicate with each other (quorum sensing) and to direct their collective behavior. Here, we explore agents (active particles) which communicate with each other to coordinate their collective dynamics to optimize nutrient consumption. Using reinforcement learning and neural networks, we identify three different strategies: a \"clustering strategy\", where the agents accumulate in regions of high nutrient concentration; a \"spreading strategy\", where particles stay away from each other to avoid competing for sparse resources; and an \"adaptive strategy\", where the agents adaptively decide to either follow or stay away from others. Our work exemplifies the idea that machine learning can be used to determine parameters that are evolutionarily optimized in biological systems but often occur as unknown parameters in mathematical models describing their dynamics.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"49 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139446474","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}
Sparse-view computed tomography (SVCT) is regarded as a promising technique to accelerate data acquisition and reduce radiation dose. However, in the presence of metallic implants, SVCT inevitably makes the reconstructed CT images suffer from severe metal artifacts and streaking artifacts due to the lack of sufficient projection data. Previous stand-alone SVCT and metal artifact reduction (MAR) methods to solve the problem of simultaneously sparse-view and metal artifact reduction (SVMAR) are plagued by insufficient correction accuracy. To overcome this limitation, we propose a multi-domain deep unrolling network, called Mud-Net, for SVMAR. Specifically, we establish a joint sinogram, image, artifact, and coding domains deep unrolling reconstruction model to recover high-quality CT images from the under-sampled sinograms corrupted by metallic implants. To train this multi-domain network effectively, we embed multi-domain knowledge into the network training process. Comprehensive experiments demonstrate that our method is superior to both existing MAR methods in the full-view MAR task and previous SVCT methods in the SVMAR task.
稀疏视图计算机断层扫描（SVCT）被认为是一种很有前途的技术，可加快数据采集速度并减少辐射剂量。然而，在存在金属植入物的情况下，由于缺乏足够的投影数据，SVCT 不可避免地会使重建的 CT 图像出现严重的金属伪影和条纹伪影。以往独立的 SVCT 和金属伪影还原（MAR）方法在解决同时还原稀疏视图和金属伪影（SVMAR）的问题时，受到校正精度不足的困扰。为了克服这一局限性，我们提出了一种用于 SVMAR 的多域深度展开网络，称为 Mud-Net。具体来说，我们建立了一个联合窦状图、图像、伪影和编码域的深度展开重建模型，以从被金属植入物破坏的低采样窦状图中恢复高质量的 CT 图像。为了有效地训练这个多域网络，我们在网络训练过程中嵌入了多域知识。综合实验证明，我们的方法在全视图 MAR 任务中优于现有的 MAR 方法，在 SVMAR 任务中优于之前的 SVCT 方法。
{"title":"Mud-Net: Multi-domain deep unrolling network for simultaneous sparse-view and metal artifact reduction in computed tomography","authors":"Baoshun Shi, Ke Jiang, Shaolei Zhang, Qiusheng Lian, Yanwei Qin, Yunsong Zhao","doi":"10.1088/2632-2153/ad1b8e","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1b8e","url":null,"abstract":"\u0000 Sparse-view computed tomography (SVCT) is regarded as a promising technique to accelerate data acquisition and reduce radiation dose. However, in the presence of metallic implants, SVCT inevitably makes the reconstructed CT images suffer from severe metal artifacts and streaking artifacts due to the lack of sufficient projection data. Previous stand-alone SVCT and metal artifact reduction (MAR) methods to solve the problem of simultaneously sparse-view and metal artifact reduction (SVMAR) are plagued by insufficient correction accuracy. To overcome this limitation, we propose a multi-domain deep unrolling network, called Mud-Net, for SVMAR. Specifically, we establish a joint sinogram, image, artifact, and coding domains deep unrolling reconstruction model to recover high-quality CT images from the under-sampled sinograms corrupted by metallic implants. To train this multi-domain network effectively, we embed multi-domain knowledge into the network training process. Comprehensive experiments demonstrate that our method is superior to both existing MAR methods in the full-view MAR task and previous SVCT methods in the SVMAR task.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"14 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139382823","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 : 2024-01-03DOI: 10.1088/2632-2153/ad1a50
Yifan Wang, Yang Li, Jiaqi Wang, Haofeng Lv, Jinshi Guo
During the tracking of moving targets in dynamic scenes, efficiently handling outliers in the optical flow and maintaining robustness across various motion amplitudes represents a critical challenge. So far, studies have used thresholding and local consistency based approaches to deal with optical outliers. However, there is subjectivity through expert-defined thresholds or delineated regions, and therefore these methods do not perform consistently enough under different target motion amplitudes. Other studies have focused on complex statistical-mathematical modelling which, although theoretically valid, requires significant computational resources. Aiming at the above problems this paper proposes a new method to calculate the optical outliers by using stochastic neighborhood graph combined with the Borda counting method, which reduces the computation amount on the basis of objectively eliminating the outliers. Sparse optical flow values are used as the overall population and the outlier and inlier sparse optical flow values are used as samples. Analyze the dissimilarity between sparse optical flow data points, obtaining the dissimilarity matrix, introducing the Gaussian function to smooth and reduce the dimensionality of the dissimilarity matrix, and then normalizing the smoothing matrix to generate the binding matrix, where the probability sum of each node to other nodes in the matrix is equal to 1. Stochastic neighborhood graphs are then generated based on a binding matrix to obtain the outlier probabilities of data points in different neighborhood graphs, and outlier samples are obtained based on the probability. To avoid the subjectivity of the expert thresholds, the outlier probabilities are weighted and ranked to calculate the data point Borda scores to obtain accurate optical outliers. The experimental results show that the method in this paper is robust to different amplitude motions and real scenarios, and the accuracy, precision and recall of outliers elimination are better than the current mainstream algorithms.
{"title":"Sparse Optical Flow Outliers Elimination Method Based on Borda Stochastic Neighborhood Graph","authors":"Yifan Wang, Yang Li, Jiaqi Wang, Haofeng Lv, Jinshi Guo","doi":"10.1088/2632-2153/ad1a50","DOIUrl":"https://doi.org/10.1088/2632-2153/ad1a50","url":null,"abstract":"\u0000 During the tracking of moving targets in dynamic scenes, efficiently handling outliers in the optical flow and maintaining robustness across various motion amplitudes represents a critical challenge. So far, studies have used thresholding and local consistency based approaches to deal with optical outliers. However, there is subjectivity through expert-defined thresholds or delineated regions, and therefore these methods do not perform consistently enough under different target motion amplitudes. Other studies have focused on complex statistical-mathematical modelling which, although theoretically valid, requires significant computational resources. Aiming at the above problems this paper proposes a new method to calculate the optical outliers by using stochastic neighborhood graph combined with the Borda counting method, which reduces the computation amount on the basis of objectively eliminating the outliers. Sparse optical flow values are used as the overall population and the outlier and inlier sparse optical flow values are used as samples. Analyze the dissimilarity between sparse optical flow data points, obtaining the dissimilarity matrix, introducing the Gaussian function to smooth and reduce the dimensionality of the dissimilarity matrix, and then normalizing the smoothing matrix to generate the binding matrix, where the probability sum of each node to other nodes in the matrix is equal to 1. Stochastic neighborhood graphs are then generated based on a binding matrix to obtain the outlier probabilities of data points in different neighborhood graphs, and outlier samples are obtained based on the probability. To avoid the subjectivity of the expert thresholds, the outlier probabilities are weighted and ranked to calculate the data point Borda scores to obtain accurate optical outliers. The experimental results show that the method in this paper is robust to different amplitude motions and real scenarios, and the accuracy, precision and recall of outliers elimination are better than the current mainstream algorithms.","PeriodicalId":503691,"journal":{"name":"Machine Learning: Science and Technology","volume":"18 7","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139388556","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：