Predicting grapevine phenological stages(GPHS)is critical for precisely managing vineyard operations,including plant disease treatments,pruning,and harvest.Solutions commonly used to address viticulture challenges rel...Predicting grapevine phenological stages(GPHS)is critical for precisely managing vineyard operations,including plant disease treatments,pruning,and harvest.Solutions commonly used to address viticulture challenges rely on image processing techniques,which have achieved significant results.However,they require the installation of dedicated hardware in the vineyard,making it invasive and difficult to maintain.Moreover,accurate prediction is influenced by the interplay of climatic factors,especially temperature,and the impact of global warming,which are difficult to model using images.Another problem frequently found in GPHS prediction is the persistent issue of missing values in viticultural datasets,particularly in phenological stages.This paper proposes a semi-supervised approach that begins with a small set of labeled phenological stage examples and automatically generates new annotations for large volumes of unlabeled climatic data.This approach aims to address key challenges in phenological analysis.This novel climatic data-based approach offers advantages over common image processing methods,as it is non-intrusive,cost-effective,and adaptable for vineyards of various sizes and technological levels.To ensure the robustness of the proposed Pseudo-labelling strategy,we integrated it into eight machine-learning algorithms.We evaluated its performance across seven diverse datasets,each exhibiting varying percentages of missing values.Performance metrics,including the coefficient of determination(R2)and root-mean-square error(RMSE),are employed to assess the effectiveness of the models.The study demonstrates that integrating the proposed Pseudo-labeling strategy with supervised learning approaches significantly improves predictive accuracy.Moreover,the study shows that the proposed methodology can also be integrated with explainable artificial intelligence techniques to determine the importance of the input features.In particular,the investigation highlights that growing degree days are crucial for improved GPHS prediction.展开更多
Few-shot learning attempts to identify novel categories by exploiting limited labeled training data,while the performances of existing methods still have much room for improvement.Thanks to a very low cost,many recent...Few-shot learning attempts to identify novel categories by exploiting limited labeled training data,while the performances of existing methods still have much room for improvement.Thanks to a very low cost,many recent methods resort to additional unlabeled training data to boost performance,known as semi-supervised few-shot learning(SSFSL).The general idea of SSFSL methods is to first generate pseudo labels for all unlabeled data and then augment the labeled training set with selected pseudo-labeled data.However,almost all previous SSFSL methods only take supervision signal from pseudo-labeling,ignoring that the distribution of training data can also be utilized as an effective unsupervised regularization.In this paper,we propose a simple yet effective SSFSL method named feature reconstruction based regression method(TENET),which takes low-rank feature reconstruction as the unsupervised objective function and pseudo labels as the supervised constraint.We provide several theoretical insights on why TENET can mitigate overfitting on low-quality training data,and why it can enhance the robustness against inaccurate pseudo labels.Extensive experiments on four popular datasets validate the effectiveness of TENET.展开更多
Age estimation using forensics odontology is an important process in identifying victims in criminal or mass disaster cases.Traditionally,this process is done manually by human expert.However,the speed and accuracy ma...Age estimation using forensics odontology is an important process in identifying victims in criminal or mass disaster cases.Traditionally,this process is done manually by human expert.However,the speed and accuracy may vary depending on the expertise level of the human expert and other human factors such as level of fatigue and attentiveness.To improve the recognition speed and consistency,researchers have proposed automated age estimation using deep learning techniques such as Convolutional Neural Network(CNN).CNN requires many training images to obtain high percentage of recognition accuracy.Unfortunately,it is very difficult to get large number of samples of dental images for training the CNN due to the need to comply to privacy acts.A promising solution to this problem is a technique called Generative Adversarial Network(GAN).GAN is a technique that can generate synthetic images that has similar statistics as the training set.A variation of GAN called Conditional GAN(CGAN)enables the generation of the synthetic images to be controlled more precisely such that only the specified type of images will be generated.This paper proposes a CGAN for generating new dental images to increase the number of images available for training a CNN model to perform age estimation.We also propose a pseudolabelling technique to label the generated images with proper age and gender.We used the combination of real and generated images to trainDentalAge and Sex Net(DASNET),which is a CNN model for dental age estimation.Based on the experiment conducted,the accuracy,coefficient of determination(R2)and Absolute Error(AE)of DASNET have improved to 87%,0.85 and 1.18 years respectively as opposed to 74%,0.72 and 3.45 years when DASNET is trained using real,but smaller number of images.展开更多
By leveraging data from a fully labeled source domain,unsupervised domain adaptation(UDA)im-proves classification performance on an unlabeled target domain through explicit discrepancy minimization of data distributio...By leveraging data from a fully labeled source domain,unsupervised domain adaptation(UDA)im-proves classification performance on an unlabeled target domain through explicit discrepancy minimization of data distribution or adversarial learning.As an enhancement,category alignment is involved during adaptation to reinforce target feature discrimination by utilizing model prediction.However,there remain unexplored prob-lems about pseudo-label inaccuracy incurred by wrong category predictions on target domain,and distribution deviation caused by overfitting on source domain.In this paper,we propose a model-agnostic two-stage learning framework,which greatly reduces flawed model predictions using soft pseudo-label strategy and avoids overfitting on source domain with a curriculum learning strategy.Theoretically,it successfully decreases the combined risk in the upper bound of expected error on the target domain.In the first stage,we train a model with distribution alignment-based UDA method to obtain soft semantic label on target domain with rather high confidence.To avoid overfitting on source domain,in the second stage,we propose a curriculum learning strategy to adaptively control the weighting between losses from the two domains so that the focus of the training stage is gradually shifted from source distribution to target distribution with prediction confidence boosted on the target domain.Extensive experiments on two well-known benchmark datasets validate the universal effectiveness of our proposed framework on promoting the performance of the top-ranked UDA algorithms and demonstrate its consistent su-perior performance.展开更多
The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fau...The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fault characteristics under different loads is markedly inconsistent,and data is hard to label,which makes it difficult for the traditional diagnosis method based on single-condition training to generalize to different conditions.To address these issues,the paper proposes a novel transfer discriminant neural network(TDNN)for gear fault diagnosis.Specifically,an optimized joint distribution adaptive mechanism(OJDA)is designed to solve the distribution alignment problem between two domains.To improve the classification effect within the domain and the feature recognition capability for a few labeled data,metric learning is introduced to distinguish features from different fault categories.In addition,TDNN adopts a new pseudo-label training strategy to achieve label replacement by comparing the maximum probability of the pseudo-label with the test result.The proposed TDNN is verified in the experimental data set of the artillery manipulator device,and the diagnosis can achieve 99.5%,significantly outperforming other traditional adaptation methods.展开更多
Deep learning networks are increasingly exploited in the field of neuronal soma segmentation.However,annotating dataset is also an expensive and time-consuming task.Unsupervised domain adaptation is an effective metho...Deep learning networks are increasingly exploited in the field of neuronal soma segmentation.However,annotating dataset is also an expensive and time-consuming task.Unsupervised domain adaptation is an effective method to mitigate the problem,which is able to learn an adaptive segmentation model by transferring knowledge from a rich-labeled source domain.In this paper,we propose a multi-level distribution alignment-based unsupervised domain adaptation network(MDA-Net)for segmentation of 3D neuronal soma images.Distribution alignment is performed in both feature space and output space.In the feature space,features from different scales are adaptively fused to enhance the feature extraction capability for small target somata and con-strained to be domain invariant by adversarial adaptation strategy.In the output space,local discrepancy maps that can reveal the spatial structures of somata are constructed on the predicted segmentation results.Then thedistribution alignment is performed on the local discrepancies maps across domains to obtain a superior discrepancy map in the target domain,achieving refined segmentation performance of neuronal somata.Additionally,after a period of distribution align-ment procedure,a portion of target samples with high confident pseudo-labels are selected as training data,which assist in learning a more adaptive segmentation network.We verified the superiority of the proposed algorithm by comparing several domain adaptation networks on two 3D mouse brain neuronal somata datasets and one macaque brain neuronal soma dataset.展开更多
A common necessity for prior unsupervised domain adaptation methods that can improve the domain adaptation in unlabeled target domain dataset is access to source domain data-set and target domain dataset simultaneousl...A common necessity for prior unsupervised domain adaptation methods that can improve the domain adaptation in unlabeled target domain dataset is access to source domain data-set and target domain dataset simultaneously.However,data privacy makes it not always possible to access source domain dataset and target domain dataset in actual industrial equipment simulta-neously,especially for aviation component like Electro-Mechanical Actuator(EMA)whose dataset are often not shareable due to the data copyright and confidentiality.To address this problem,this paper proposes a source free unsupervised domain adaptation framework for EMA fault diagnosis.The proposed framework is a combination of feature network and classifier.Firstly,source domain datasets are only applied to train a source model.Secondly,the well-trained source model is trans-ferred to target domain and classifier is frozen based on source domain hypothesis.Thirdly,nearest centroid filtering is introduced to filter the reliable pseudo labels for unlabeled target domain data-set,and finally,supervised learning and pseudo label clustering are applied to fine-tune the trans-ferred model.In comparison with several traditional unsupervised domain adaptation methods,case studies based on low-and high-frequency monitoring signals on EMA indicate the effectiveness of the proposed method.展开更多
AIM:To address the challenges of data labeling difficulties,data privacy,and necessary large amount of labeled data for deep learning methods in diabetic retinopathy(DR)identification,the aim of this study is to devel...AIM:To address the challenges of data labeling difficulties,data privacy,and necessary large amount of labeled data for deep learning methods in diabetic retinopathy(DR)identification,the aim of this study is to develop a source-free domain adaptation(SFDA)method for efficient and effective DR identification from unlabeled data.METHODS:A multi-SFDA method was proposed for DR identification.This method integrates multiple source models,which are trained from the same source domain,to generate synthetic pseudo labels for the unlabeled target domain.Besides,a softmax-consistence minimization term is utilized to minimize the intra-class distances between the source and target domains and maximize the inter-class distances.Validation is performed using three color fundus photograph datasets(APTOS2019,DDR,and EyePACS).RESULTS:The proposed model was evaluated and provided promising results with respectively 0.8917 and 0.9795 F1-scores on referable and normal/abnormal DR identification tasks.It demonstrated effective DR identification through minimizing intra-class distances and maximizing inter-class distances between source and target domains.CONCLUSION:The multi-SFDA method provides an effective approach to overcome the challenges in DR identification.The method not only addresses difficulties in data labeling and privacy issues,but also reduces the need for large amounts of labeled data required by deep learning methods,making it a practical tool for early detection and preservation of vision in diabetic patients.展开更多
Unsupervised domain adaptation(UDA),which aims to use knowledge from a label-rich source domain to help learn unlabeled target domain,has recently attracted much attention.UDA methods mainly concentrate on source clas...Unsupervised domain adaptation(UDA),which aims to use knowledge from a label-rich source domain to help learn unlabeled target domain,has recently attracted much attention.UDA methods mainly concentrate on source classification and distribution alignment between domains to expect the correct target prediction.While in this paper,we attempt to learn the target prediction end to end directly,and develop a Self-corrected unsupervised domain adaptation(SCUDA)method with probabilistic label correction.SCUDA adopts a probabilistic label corrector to learn and correct the target labels directly.Specifically,besides model parameters,those target pseudo-labels are also updated in learning and corrected by the anchor-variable,which preserves the class candidates for samples.Experiments on real datasets show the competitiveness of SCUDA.展开更多
基金supported by the Departmental Strategic Plan(PSD)of the University of Udine-Interdepartmental Project on Artificial Intelligence(2020-25)this study was carried out within the Agritech National Research Center and received funding from the European Union's Next-Generation EU(National Recovery and Resilience Plan(PNRR)-Mission 4,Component 2,Investment 1.4-Decree No.1032 of 17/06/2022,CN00000022).
文摘Predicting grapevine phenological stages(GPHS)is critical for precisely managing vineyard operations,including plant disease treatments,pruning,and harvest.Solutions commonly used to address viticulture challenges rely on image processing techniques,which have achieved significant results.However,they require the installation of dedicated hardware in the vineyard,making it invasive and difficult to maintain.Moreover,accurate prediction is influenced by the interplay of climatic factors,especially temperature,and the impact of global warming,which are difficult to model using images.Another problem frequently found in GPHS prediction is the persistent issue of missing values in viticultural datasets,particularly in phenological stages.This paper proposes a semi-supervised approach that begins with a small set of labeled phenological stage examples and automatically generates new annotations for large volumes of unlabeled climatic data.This approach aims to address key challenges in phenological analysis.This novel climatic data-based approach offers advantages over common image processing methods,as it is non-intrusive,cost-effective,and adaptable for vineyards of various sizes and technological levels.To ensure the robustness of the proposed Pseudo-labelling strategy,we integrated it into eight machine-learning algorithms.We evaluated its performance across seven diverse datasets,each exhibiting varying percentages of missing values.Performance metrics,including the coefficient of determination(R2)and root-mean-square error(RMSE),are employed to assess the effectiveness of the models.The study demonstrates that integrating the proposed Pseudo-labeling strategy with supervised learning approaches significantly improves predictive accuracy.Moreover,the study shows that the proposed methodology can also be integrated with explainable artificial intelligence techniques to determine the importance of the input features.In particular,the investigation highlights that growing degree days are crucial for improved GPHS prediction.
基金supported in part by the Beijing Natural Science Foundation,China(No.L221013)in part by the National Science Foundation of China(Nos.U20B2070 and 61832016).
文摘Few-shot learning attempts to identify novel categories by exploiting limited labeled training data,while the performances of existing methods still have much room for improvement.Thanks to a very low cost,many recent methods resort to additional unlabeled training data to boost performance,known as semi-supervised few-shot learning(SSFSL).The general idea of SSFSL methods is to first generate pseudo labels for all unlabeled data and then augment the labeled training set with selected pseudo-labeled data.However,almost all previous SSFSL methods only take supervision signal from pseudo-labeling,ignoring that the distribution of training data can also be utilized as an effective unsupervised regularization.In this paper,we propose a simple yet effective SSFSL method named feature reconstruction based regression method(TENET),which takes low-rank feature reconstruction as the unsupervised objective function and pseudo labels as the supervised constraint.We provide several theoretical insights on why TENET can mitigate overfitting on low-quality training data,and why it can enhance the robustness against inaccurate pseudo labels.Extensive experiments on four popular datasets validate the effectiveness of TENET.
文摘Age estimation using forensics odontology is an important process in identifying victims in criminal or mass disaster cases.Traditionally,this process is done manually by human expert.However,the speed and accuracy may vary depending on the expertise level of the human expert and other human factors such as level of fatigue and attentiveness.To improve the recognition speed and consistency,researchers have proposed automated age estimation using deep learning techniques such as Convolutional Neural Network(CNN).CNN requires many training images to obtain high percentage of recognition accuracy.Unfortunately,it is very difficult to get large number of samples of dental images for training the CNN due to the need to comply to privacy acts.A promising solution to this problem is a technique called Generative Adversarial Network(GAN).GAN is a technique that can generate synthetic images that has similar statistics as the training set.A variation of GAN called Conditional GAN(CGAN)enables the generation of the synthetic images to be controlled more precisely such that only the specified type of images will be generated.This paper proposes a CGAN for generating new dental images to increase the number of images available for training a CNN model to perform age estimation.We also propose a pseudolabelling technique to label the generated images with proper age and gender.We used the combination of real and generated images to trainDentalAge and Sex Net(DASNET),which is a CNN model for dental age estimation.Based on the experiment conducted,the accuracy,coefficient of determination(R2)and Absolute Error(AE)of DASNET have improved to 87%,0.85 and 1.18 years respectively as opposed to 74%,0.72 and 3.45 years when DASNET is trained using real,but smaller number of images.
基金the 111 Project(No.BP0719010)the Project of the Science and Technology Commission of Shanghai Municipality(No.18DZ2270700)。
文摘By leveraging data from a fully labeled source domain,unsupervised domain adaptation(UDA)im-proves classification performance on an unlabeled target domain through explicit discrepancy minimization of data distribution or adversarial learning.As an enhancement,category alignment is involved during adaptation to reinforce target feature discrimination by utilizing model prediction.However,there remain unexplored prob-lems about pseudo-label inaccuracy incurred by wrong category predictions on target domain,and distribution deviation caused by overfitting on source domain.In this paper,we propose a model-agnostic two-stage learning framework,which greatly reduces flawed model predictions using soft pseudo-label strategy and avoids overfitting on source domain with a curriculum learning strategy.Theoretically,it successfully decreases the combined risk in the upper bound of expected error on the target domain.In the first stage,we train a model with distribution alignment-based UDA method to obtain soft semantic label on target domain with rather high confidence.To avoid overfitting on source domain,in the second stage,we propose a curriculum learning strategy to adaptively control the weighting between losses from the two domains so that the focus of the training stage is gradually shifted from source distribution to target distribution with prediction confidence boosted on the target domain.Extensive experiments on two well-known benchmark datasets validate the universal effectiveness of our proposed framework on promoting the performance of the top-ranked UDA algorithms and demonstrate its consistent su-perior performance.
文摘The ammunition loading system manipulator is susceptible to gear failure due to high-frequency,heavyload reciprocating motions and the absence of protective gear components.After a fault occurs,the distribution of fault characteristics under different loads is markedly inconsistent,and data is hard to label,which makes it difficult for the traditional diagnosis method based on single-condition training to generalize to different conditions.To address these issues,the paper proposes a novel transfer discriminant neural network(TDNN)for gear fault diagnosis.Specifically,an optimized joint distribution adaptive mechanism(OJDA)is designed to solve the distribution alignment problem between two domains.To improve the classification effect within the domain and the feature recognition capability for a few labeled data,metric learning is introduced to distinguish features from different fault categories.In addition,TDNN adopts a new pseudo-label training strategy to achieve label replacement by comparing the maximum probability of the pseudo-label with the test result.The proposed TDNN is verified in the experimental data set of the artillery manipulator device,and the diagnosis can achieve 99.5%,significantly outperforming other traditional adaptation methods.
基金supported by the Fund of Key Laboratory of Biomedical Engineering of Hainan Province(No.BME20240001)the STI2030-Major Projects(No.2021ZD0200104)the National Natural Science Foundations of China under Grant 61771437.
文摘Deep learning networks are increasingly exploited in the field of neuronal soma segmentation.However,annotating dataset is also an expensive and time-consuming task.Unsupervised domain adaptation is an effective method to mitigate the problem,which is able to learn an adaptive segmentation model by transferring knowledge from a rich-labeled source domain.In this paper,we propose a multi-level distribution alignment-based unsupervised domain adaptation network(MDA-Net)for segmentation of 3D neuronal soma images.Distribution alignment is performed in both feature space and output space.In the feature space,features from different scales are adaptively fused to enhance the feature extraction capability for small target somata and con-strained to be domain invariant by adversarial adaptation strategy.In the output space,local discrepancy maps that can reveal the spatial structures of somata are constructed on the predicted segmentation results.Then thedistribution alignment is performed on the local discrepancies maps across domains to obtain a superior discrepancy map in the target domain,achieving refined segmentation performance of neuronal somata.Additionally,after a period of distribution align-ment procedure,a portion of target samples with high confident pseudo-labels are selected as training data,which assist in learning a more adaptive segmentation network.We verified the superiority of the proposed algorithm by comparing several domain adaptation networks on two 3D mouse brain neuronal somata datasets and one macaque brain neuronal soma dataset.
基金supported by the National Natural Science Foundation of China(No.52075349)the Aeronautical Science Foundation of China(No.201905019001)the China Scholarship Council(No.202106240078).
文摘A common necessity for prior unsupervised domain adaptation methods that can improve the domain adaptation in unlabeled target domain dataset is access to source domain data-set and target domain dataset simultaneously.However,data privacy makes it not always possible to access source domain dataset and target domain dataset in actual industrial equipment simulta-neously,especially for aviation component like Electro-Mechanical Actuator(EMA)whose dataset are often not shareable due to the data copyright and confidentiality.To address this problem,this paper proposes a source free unsupervised domain adaptation framework for EMA fault diagnosis.The proposed framework is a combination of feature network and classifier.Firstly,source domain datasets are only applied to train a source model.Secondly,the well-trained source model is trans-ferred to target domain and classifier is frozen based on source domain hypothesis.Thirdly,nearest centroid filtering is introduced to filter the reliable pseudo labels for unlabeled target domain data-set,and finally,supervised learning and pseudo label clustering are applied to fine-tune the trans-ferred model.In comparison with several traditional unsupervised domain adaptation methods,case studies based on low-and high-frequency monitoring signals on EMA indicate the effectiveness of the proposed method.
基金Supported by the Fund for Shanxi“1331 Project”and Supported by Fundamental Research Program of Shanxi Province(No.202203021211006)the Key Research,Development Program of Shanxi Province(No.201903D311009)+4 种基金the Key Research Program of Taiyuan University(No.21TYKZ01)the Open Fund of Shanxi Province Key Laboratory of Ophthalmology(No.2023SXKLOS04)Shenzhen Fund for Guangdong Provincial High-Level Clinical Key Specialties(No.SZGSP014)Sanming Project of Medicine in Shenzhen(No.SZSM202311012)Shenzhen Science and Technology Planning Project(No.KCXFZ20211020163813019).
文摘AIM:To address the challenges of data labeling difficulties,data privacy,and necessary large amount of labeled data for deep learning methods in diabetic retinopathy(DR)identification,the aim of this study is to develop a source-free domain adaptation(SFDA)method for efficient and effective DR identification from unlabeled data.METHODS:A multi-SFDA method was proposed for DR identification.This method integrates multiple source models,which are trained from the same source domain,to generate synthetic pseudo labels for the unlabeled target domain.Besides,a softmax-consistence minimization term is utilized to minimize the intra-class distances between the source and target domains and maximize the inter-class distances.Validation is performed using three color fundus photograph datasets(APTOS2019,DDR,and EyePACS).RESULTS:The proposed model was evaluated and provided promising results with respectively 0.8917 and 0.9795 F1-scores on referable and normal/abnormal DR identification tasks.It demonstrated effective DR identification through minimizing intra-class distances and maximizing inter-class distances between source and target domains.CONCLUSION:The multi-SFDA method provides an effective approach to overcome the challenges in DR identification.The method not only addresses difficulties in data labeling and privacy issues,but also reduces the need for large amounts of labeled data required by deep learning methods,making it a practical tool for early detection and preservation of vision in diabetic patients.
基金This work was supported by the National Natural Science Foundation of China(Grant Nos.61876091,61772284)the China Postdoctoral Science Foundation(2019M651918)the Open Foundation of MIIT Key Laboratory of Pattern Analysis and Machine Intelligence.
文摘Unsupervised domain adaptation(UDA),which aims to use knowledge from a label-rich source domain to help learn unlabeled target domain,has recently attracted much attention.UDA methods mainly concentrate on source classification and distribution alignment between domains to expect the correct target prediction.While in this paper,we attempt to learn the target prediction end to end directly,and develop a Self-corrected unsupervised domain adaptation(SCUDA)method with probabilistic label correction.SCUDA adopts a probabilistic label corrector to learn and correct the target labels directly.Specifically,besides model parameters,those target pseudo-labels are also updated in learning and corrected by the anchor-variable,which preserves the class candidates for samples.Experiments on real datasets show the competitiveness of SCUDA.
