Semi-supervised Deep Learning Research Articles

Semi-Supervised Training for (Pre-Stack) Seismic Data Analysis

A lack of labeled examples is a problem in different domains, such as text and image processing, medicine, and static reservoir characterization, because supervised learning relies on vast volumes of these data to perform successfully, but this is quite expensive. However, large amounts of unlabeled data exist in these domains. The deep semi-supervised learning (DSSL) approach leverages unlabeled data to improve supervised learning performance using deep neural networks. This approach has succeeded in image recognition, text classification, and speech recognition. Nevertheless, there have been few works on pre-stack seismic reservoir characterization, in which knowledge of rock and fluid properties is fundamental for oil exploration. This paper proposes a methodology to estimate acoustic impedance using pre-stack seismic data and DSSL with a recurrent neural network. The few labeled datasets for training were pre-processed from raw seismic and acoustic impedance data from five borehole logs. The results showed that the acoustic impedance estimation at the well location and outside it was better predicted by the DSSL compared to the supervised version of the same neural network. Therefore, employing a large amount of unlabeled data can be helpful in the development of seismic data interpretation systems.

Semi-supervised PolSAR Image Change Detection using Similarity Matching

Abstract. The lack of precisely labeled data limits the development of supervised polarimetric synthetic aperture radar (PolSAR) image change detection. Therefore, semi-supervised deep learning methods have recently demonstrated their significant capability for PolSAR image change detection. Similarity Matching (SimMatch) improves the performance of semi-supervised learning tasks across different benchmark datasets and different settings. Introducing SimMatch into the field of PolSAR image change detection can improve the performance of semi-supervised PolSAR image change detection under limited labeled data conditions. Usually, semi-supervision solves the problem of insufficient labeled data by generating pseudo-labels. However, when the pseudo-label method is simply applied, the model will fit on the confident but wrong pseudo-labels, resulting in poor performance. SimMatch offers a solution by requiring the strongly augmented view to share the same semantic similarity (i.e. label prediction) and instance characteristics (i.e. similarity between instances) with a weak augmented view for more intrinsic feature matching. Besides, by using a labeled memory buffer, the two similarities can be isomorphically transformed with each other by introducing the aggregating and unfolding techniques. Therefore, the semantic and instance pseudo-labels can be mutually propagated, and then, the detection performance of the PolSAR image change detection is improved. Experimental results on real PolSAR datasets demonstrated that SimMatch is an effective semi-supervised PolSAR change detection method and its performance surpasses some well-known change detection methods. Compared to the fully-supervised algorithm CWNN, the semi-supervised SimMatch algorithm can improve accuracy by up to 14.4%.

Deep semi-supervised learning for recovering traceability links between issues and commits

Traceability links between issues and commits record valuable information about the evolutionary history of software projects. Unfortunately, these links are often missing. While deep learning stands as the current state-of-the-art (SOTA) in automated traceability link recovery (TLR), its effectiveness is faced with the practical problem of limited labeled data during training. To overcome this challenge, in this paper, we propose DSSLink, a novel method based on deep semi-supervised learning, enhancing deep-learning-based link recovery tasks. DSSLink first learns knowledge from labeled data through pre-trained model and then leverages deep semi-supervised learning to infer pseudo-labels on unlabeled data. The extended dataset of pseudo-labeled and labeled data re-trains the deep learning model in an iterative process. Our extensive evaluations are conducted on two SOTA traceability methods (T-BERT and BTLink) across four GitHub projects and 11 Apache projects. Specifically, the maximum F1-score improvements for GitHub and Apache projects reached 22.9% and 43.5%, respectively. Evaluation results show that DSSLink is effective in enhancing TLR performance and outperforms TraceFUN, a recent approach that utilizes unlabeled data for TLR. The source code of DSSLink is available at https://github.com/DSSLink.Editor’s note: Open Science material was validated by the Journal of Systems and Software Open Science Board.

Real-time coronary artery segmentation in CAG images: A semi-supervised deep learning strategy

BackgroundWhen treating patients with coronary artery disease and concurrent renal concerns, we often encounter a conundrum: how to achieve a clearer view of vascular details while minimizing the contrast and radiation doses during percutaneous coronary intervention (PCI). Our goal is to use deep learning (DL) to create a real-time roadmap for guiding PCI. To this end, segmentation, a critical first step, paves the way for detailed vascular analysis. Unlike traditional supervised learning, which demands extensive labeling time and manpower, our strategy leans toward semi-supervised learning. This method not only economizes on labeling efforts but also aims at reducing contrast and radiation exposure. Methods and resultsCAG data sourced from eight tertiary centers in Taiwan, comprising 500 labeled and 8952 unlabeled images. Employing 400 labels for training and reserving 100 for validation, we built a U-Net based network within a teacher-student architecture. The initial teacher model was updated with 8952 unlabeled images inputted, employing a quality control strategy involving consistency regularization and RandAugment. The optimized teacher model produced pseudo-labels for label expansion, which were then utilized to train the final student model. We attained an average dice similarity coefficient of 0.9003 for segmentation, outperforming supervised learning methods with the same label count. Even with only 5 % labels for semi-supervised training, the results surpassed a supervised method with 100 % labels inputted. This semi-supervised approach's advantage extends beyond single-frame prediction, yielding consistently superior results in continuous angiography films. ConclusionsHigh labeling cost hinders DL training. Semi-supervised learning, quality control, and pseudo-label expansion can overcome this. DL-assisted segmentation potentially provides a real-time PCI roadmap and further diminishes radiation and contrast doses.

Semi-supervised learning in diagnosis of infant hip dysplasia towards multisource ultrasound images.

Automated diagnosis of infant hip dysplasia is heavily affected by the individual differences among infants and ultrasound machines. Hip sonographic images of 493 infants from various ultrasound machines were collected in the Department of Orthopedics in Yangzhou Maternal and Child Health Care Service Centre. Herein, we propose a semi-supervised learning method based on a feature pyramid network (FPN) and a contrastive learning scheme based on a Siamese architecture. A large amount of unlabeled data of ultrasound images was used via the Siamese network in the pre-training step, and then a small amount of annotated data for anatomical structures was adopted to train the model for landmark identification and standard plane recognition. The method was evaluated on our collected dataset. The method achieved a mean Dice similarity coefficient (DSC) of 0.7873 and a mean Hausdorff distance (HD) of 5.0102 in landmark identification, compared to the model without contrastive learning, which had a mean DSC of 0.7734 and a mean HD of 6.1586. The accuracy, precision, and recall of standard plane recognition were 95.4%, 91.64%, and 94.86%, respectively. The corresponding area under the curve (AUC) was 0.982. This study proposes a semi-supervised deep learning method following Graf's principle, which can better utilize a large volume of ultrasound images from various devices and infants. This method can identify the landmarks of infant hips more accurately than manual operators, thereby improving the efficiency of diagnosis of infant hip dysplasia.

Graph deep learning recognition of port ship behavior patterns from a network approach

The detection and recognition of ship mobile behavior patterns based on automatic identification system (AIS) data are crucial for ship navigation and maritime supervision. However, present methods, such as rule-based constraints and density clustering extraction, still suffer from restrictions imposed by artificial settings of parameters and thresholds. The lack of neighborhood context information in unstructured and serialized ship trajectory vector data makes it challenging to incorporate advanced methods such as deep learning for pattern learning. This study builds multiple semi-supervised deep learning graph neural network (GNN) models based on the message passing paradigm for recognizing ship mobile behavior patterns by constructing an AIS trajectory network and computing multi-order node features through Delaunay triangulation. It can get rid of the dependence on parameters of traditional rule-based and unsupervised learning methods. It also reduces the dependence on auxiliary information in trajectory qualitative by introducing multi-order node features. We conducted the experiments on cargo ship data from New York and Singapore to test our method. The model's classification accuracy is more than 99% for the New York dataset and more than 95% for the Singapore dataset. This method can be extended to other types of ships and traffic trajectory vector data with the excellent application potential of GNN in pattern recognition of vector mobile data. The code and experimental data are available at: https://github.com/destiny1103/DT-GNN.

A robust semi-supervised deep learning approach for emotion recognition using EEG signals

A robust semi-supervised deep learning approach for emotion recognition using EEG signals

A Critical Analysis of Deep Semi-Supervised Learning Approaches for Enhanced Medical Image Classification

Deep semi-supervised learning (DSSL) is a machine learning paradigm that blends supervised and unsupervised learning techniques to improve the performance of various models in computer vision tasks. Medical image classification plays a crucial role in disease diagnosis, treatment planning, and patient care. However, obtaining labeled medical image data is often expensive and time-consuming for medical practitioners, leading to limited labeled datasets. DSSL techniques aim to address this challenge, particularly in various medical image tasks, to improve model generalization and performance. DSSL models leverage both the labeled information, which provides explicit supervision, and the unlabeled data, which can provide additional information about the underlying data distribution. That offers a practical solution to resource-intensive demands of data annotation, and enhances the model’s ability to generalize across diverse and previously unseen data landscapes. The present study provides a critical review of various DSSL approaches and their effectiveness and challenges in enhancing medical image classification tasks. The study categorized DSSL techniques into six classes: consistency regularization method, deep adversarial method, pseudo-learning method, graph-based method, multi-label method, and hybrid method. Further, a comparative analysis of performance for six considered methods is conducted using existing studies. The referenced studies have employed metrics such as accuracy, sensitivity, specificity, AUC-ROC, and F1 score to evaluate the performance of DSSL methods on different medical image datasets. Additionally, challenges of the datasets, such as heterogeneity, limited labeled data, and model interpretability, were discussed and highlighted in the context of DSSL for medical image classification. The current review provides future directions and considerations to researchers to further address the challenges and take full advantage of these methods in clinical practices.

Triple-task mutual consistency for semi-supervised 3D medical image segmentation

Semi-supervised deep learning algorithm is an effective means of medical image segmentation. Among these methods, multi-task learning with consistency regularization has achieved outstanding results. However, most of the existing methods usually simply embed the Signed Distance Map (SDM) task into the network, which underestimates the potential ability of SDM in edge awareness and leads to excessive dependence between tasks. In this work, we propose a novel triple-task mutual consistency (TTMC) framework to enhance shape and edge awareness capabilities, and overcome the task dependence problem underestimated in previous work. Specifically, we innovatively construct the Signed Attention Map (SAM), a novel fusion image with attention mechanism, and use it as an auxiliary task for segmentation to enhance the edge awareness ability. Then we implement a triple-task deep network, which jointly predicts the voxel-wise classification map, the Signed Distance Map and the Signed Attention Map. In our proposed framework, an optimized differentiable transformation layer associates SDM with voxel-wise classification map and SAM prediction, while task-level consistency regularization utilizes unlabeled data in an unsupervised manner. Evaluated on the public Left Atrium dataset and NIH Pancreas dataset, our proposed framework achieves significant performance gains by effectively utilizing unlabeled data, outperforming recent state-of-the-art semi-supervised segmentation methods. Code is available at https://github.com/Saocent/TTMC.

EEG power spectra parameterization and adaptive channel selection towards semi-supervised seizure prediction

Background:The seizure prediction algorithms have demonstrated their potential in mitigating epilepsy risks by detecting the pre-ictal state using ongoing electroencephalogram (EEG) signals. However, most of them require high-density EEG, which is burdensome to the patients for daily monitoring. Moreover, prevailing seizure models require extensive training with significant labeled data which is very time-consuming and demanding for the epileptologists. Method:To address these challenges, here we propose an adaptive channel selection strategy and a semi-supervised deep learning model respectively to reduce the number of EEG channels and to limit the amount of labeled data required for accurate seizure prediction. Our channel selection module is centered on features from EEG power spectra parameterization that precisely characterize the epileptic activities to identify the seizure-associated channels for each patient. The semi-supervised model integrates generative adversarial networks and bidirectional long short-term memory networks to enhance seizure prediction. Results:Our approach is evaluated on the CHB-MIT and Siena epilepsy datasets. With utilizing only 4 channels, the method demonstrates outstanding performance with an AUC of 93.15% on the CHB-MIT dataset and an AUC of 88.98% on the Siena dataset. Experimental results also demonstrate that our selection approach reduces the model parameters and training time. Conclusions:Adaptive channel selection coupled with semi-supervised learning can offer the possible bases for a light weight and computationally efficient seizure prediction system, making the daily monitoring practical to improve patients’ quality of life.

Semi-supervised deep learning based on label propagation algorithm for debris flow susceptibility assessment in few-label scenarios

Semi-supervised deep learning based on label propagation algorithm for debris flow susceptibility assessment in few-label scenarios

A universal method for solar filament detection from H-alpha observations using semi-supervised deep learning

o he Observatory for Solar and Environmental Research (KSO; Austria) provide regular full-disk observations of the Sun in the core of the chromospheric Halpha absorption line. In this paper, we present a deep learning method that provides reliable extractions of solar filaments from Halpha filtergrams. First, we trained the object detection algorithm YOLOv5 with labeled filament data of ChroTel Halpha filtergrams. We used the trained model to obtain bounding boxes from the full GONG archive. In a second step, we applied a semi-supervised training approach where we used the bounding boxes of filaments to train the algorithm on a pixel-wise classification of solar filaments with u-net. We made use of the increased data set size, which avoids overfitting of spurious artifacts from the generated training masks. Filaments were predicted with an accuracy of 92<!PCT!>. With the resulting filament segmentations, physical parameters such as the area or tilt angle could be easily determined and studied. We demonstrated this in an example where we determined the rush-to-the pole for Solar Cycle 24 from the segmented GONG images. In a last step, we applied the filament detection to Halpha observations from KSO and demonstrated the general applicability of our method to Halpha filtergrams.

Deep Learning Models for Waterfowl Detection and Classification in Aerial Images

Waterfowl populations monitoring is essential for wetland conservation. Lately, deep learning techniques have shown promising advancements in detecting waterfowl in aerial images. In this paper, we present performance evaluation of several popular supervised and semi-supervised deep learning models for waterfowl detection in aerial images using four new image datasets containing 197,642 annotations. The best-performing model, Faster R-CNN, achieved 95.38% accuracy in terms of mAP. Semi-supervised learning models outperformed supervised models when the same amount of labeled data was used for training. Additionally, we present performance evaluation of several deep learning models on waterfowl classifications on aerial images using a new real-bird classification dataset consisting of 6,986 examples and a new decoy classification dataset consisting of about 10,000 examples per category of 20 categories. The best model achieved accuracy of 91.58% on the decoy dataset and 82.88% on the real-bird dataset.

Semi-supervised deep learning for molecular clump verification

Context. A reliable molecular clump detection algorithm is essential for studying these clumps. Existing detection algorithms for molecular clumps still require that detected candidates be verified manually, which is impractical for large-scale data. Semi-supervised learning methods, especially those based on deep features, have the potential to accomplish the task of molecular clump verification thanks to the powerful feature extraction capability of deep networks. Aims. Our main objective is to develop an automated method for the verification of molecular clump candidates. This method utilises a 3D convolutional neural network (3D CNN) to extract features of molecular clumps and employs semi-supervised learning to train the model, with the aim being to improve its generalisation ability and data utilisation. It addresses the issue of insufficient labelled samples in traditional supervised learning and enables the model to better adapt to new, unlabelled samples, achieving high accuracy in the verification of molecular clumps. Methods. We propose SS-3D-Clump, a semi-supervised deep clustering method that jointly learns the parameters of a 3D CNN and the cluster assignments of the generated features for automatic verification of molecular clumps. SS-3D-Clump iteratively classifies the features with the Constrained-KMeans and uses these class labels as supervision to update the weights of the entire network. Results. We used CO data from the Milky Way Imaging Scroll Painting project covering 350 square degrees in the Milky Way’s first, second, and third quadrants. The ClumpFind algorithm was applied to extract molecular clump candidates in these regions, which were subsequently verified using SS-3D-Clump. The SS-3D-Clump model, trained on a dataset comprising three different density regions, achieved an accuracy of 0.933, a recall rate of 0.955, a precision rate of 0.945, and an F1 score of 0.950 on the corresponding test dataset. These results closely align with those obtained through manual verification. Conclusions. Our experiments demonstrate that the SS-3D-Clump model achieves high accuracy in the automated verification of molecular clumps. It effectively captures the essential features of the molecular clumps and overcomes the challenge of limited labelled samples in supervised learning by using unlabelled samples through semi-supervised learning. This enhancement significantly improves the generalisation capability of the SS-3D-Clump model, allowing it to adapt effectively to new and unlabelled samples. Consequently, SS-3D-Clump can be integrated with any detection algorithm to create a comprehensive framework for the automated detection and verification of molecular clumps.

Gearbox fault diagnosis method based on improved semi-supervised MTDL and GAF

Aiming at the problem that it takes a long time and high cost to obtain complete labeled data under intelligent fault diagnosis and unlabeled data is not used. This paper proposes an improved semi-supervised mean teacher deep learning (MTDL) and Gramian angle field (GAF) fusion diagnostic method. This method fully utilizes a small number of labeled samples and a large number of unlabeled samples to deeply mine invisible fault features and potential physical correlations. At the same time, it solves the problem of losing the inter-data correlation structure when one-dimensional time series signals are used as inputs for neural networks. The GAF-MTDL method uses consistency regularization and modifies the network structure in the mean teacher algorithm into a semi-supervised deep learning model enhanced by WideResNet. The experimental results show that the proposed GAF-MTDL method saves a lot of manual labeling costs, improves the recognition accuracy and generalization ability, and can achieve excellent prediction accuracy with very little labeled data. In the end, the accuracy of planetary gear fault identification reached 98.22% under the labeling rate of 20%, and the accuracy of fault identification reached 99.98% through the verification of the bearing data set of Case Western Reserve University. The value of this research is to bring an efficient and low-cost technology to the field of industrial intelligent fault diagnosis, which can significantly improve the accuracy of fault identification.

CycleGAN-Based Data Augmentation for Subgrade Disease Detection in GPR Images with YOLOv5

Vehicle-mounted ground-penetrating radar (GPR) technology is an effective means of detecting railway subgrade diseases. However, existing methods of GPR data interpretation largely rely on manual identification, which is not only inefficient but also highly subjective. This paper proposes a semi-supervised deep learning method to identify railway subgrade diseases. This method addresses the sample imbalance problem in the defect dataset by utilizing a data augmentation method based on a generative adversarial network model. An initial network model for disease identification is obtained by training the YOLOv5 network with a small number of existing samples. The intelligently extended samples are then labeled to achieve a balance in the disease samples. The network is trained to improve the recognition accuracy of the intelligent model using a more complete dataset. The experimental results show that the accuracy of the proposed method can reach up to 94.53%, which is 23.85% higher than that of the supervised learning model without an extended dataset. This has strong industrial application value for railway subgrade disease detection as the potential learning ability of the model can be explored to a greater extent, thereby improving the recognition accuracy of subgrade diseases.

A deep semi-supervised machine learning algorithm for detecting transportation modes based on GPS tracking data

AbstractTransportation research has benefited from GPS tracking devices since a higher volume of data can be acquired. Trip information such as travel speed, time, and most visited locations can be easily extracted from raw GPS tracking data. However, transportation modes cannot be extracted directly and require more complex analytical processes. Common approaches for detecting travel modes heavily depend on manual labelling of trajectories with accurate trip information, which is inefficient in many aspects. This paper proposes a method of semi-supervised machine learning by using minimal labelled data. The method can accept GPS trajectory with adjustable length and extract latent information with long short-term memory (LSTM) Autoencoder. The method adopts a deep neural network architecture with three hidden layers to map the latent information to detect transportation mode. The proposed method is assessed by applying it to the case study where an accuracy of 93.94% can be achieved, which significantly outperforms similar studies.

GeoPDNN 1.0: a semi-supervised deep learning neural network using pseudo-labels for three-dimensional shallow strata modelling and uncertainty analysis in urban areas from borehole data

Abstract. Borehole data are essential for conducting precise urban geological surveys and large-scale geological investigations. Traditionally, explicit modelling and implicit modelling have been the primary methods for visualizing borehole data and constructing 3D geological models. However, explicit modelling requires substantial manual labour, while implicit modelling faces problems related to uncertainty analysis. Recently, machine learning approaches have emerged as effective solutions for addressing these issues in 3D geological modelling. Nevertheless, the use of machine learning methods for constructing 3D geological models is often limited by insufficient training data. In this paper, we propose the semi-supervised deep learning using pseudo-labels (SDLP) algorithm to overcome the issue of insufficient training data. Specifically, we construct the pseudo-labels in the training dataset using the triangular irregular network (TIN) method. A 3D geological model is constructed using borehole data obtained from a real building engineering project in Shenyang, Liaoning Province, NE China. Then, we compare the results of the 3D geological model constructed based on SDLP with those constructed by a support vector machine (SVM) method and an implicit Hermite radial basis function (HRBF) modelling method. Compared to the 3D geological models constructed using the HRBF algorithm and the SVM algorithm, the 3D geological model constructed based on the SDLP algorithm better conforms to the sedimentation patterns of the region. The findings demonstrate that our proposed method effectively resolves the issues of insufficient training data when using machine learning methods and the inability to perform uncertainty analysis when using the implicit method. In conclusion, the semi-supervised deep learning method with pseudo-labelling proposed in this paper provides a solution for 3D geological modelling in engineering project areas with borehole data.

SSDL-an automated semi-supervised deep learning approach for patient-specific 3D reconstruction of proximal femur from QCT images.

Deep Learning (DL) techniques have recently been used in medical image segmentation and the reconstruction of 3D anatomies of a human body. In this work, we propose a semi-supervised DL (SSDL) approach utilizing a CNN-based 3D U-Net model for femur segmentation from sparsely annotated quantitative computed tomography (QCT) slices. Specifically, QCT slices at the proximal end of the femur forming ball and socket joint with acetabulum were annotated for precise segmentation, where a segmenting binary mask was generated using a3D U-Net model to segment the femur accurately. A total of 5474 QCT slices were considered for training among which 2316 slices were annotated. 3D femurs were further reconstructed from segmented slices employing polynomial spline interpolation. Both qualitative and quantitative performance of segmentation and 3D reconstruction were satisfactory with more than 90% accuracy achieved for allof the standardperformance metrics considered. The spatial overlap index and reproducibility validation metric for segmentation-Dice Similarity Coefficient was 91.8% for unseen patients and 99.2% for validated patients. An average relative error of 12.02% and 10.75% for volume and surface area, respectively, were computed for 3D reconstructed femurs. The proposed approach demonstrates its effectiveness in accurately segmenting and reconstructing 3D femur from QCT slices.

Feature similarity learning based on fuzziness minimization for semi-supervised medical image segmentation

Deep learning has advanced the automation and intelligence levels of medical image segmentation, but the acquisition of annotations for medical images proves to be very challenging. Deep semi-supervised learning provides an effective approach to tackling this challenge. In this paper, we propose a method of feature similarity learning for deep semi-supervised image segmentation. The method performs a weighted transformation on the latent features of deep convolution network, and the objective of this transformation is designed to minimize the uncertainty of feature similarity. Based on fuzzy sets of similarity, we define the fuzziness to measure the uncertainty in the objective, and demonstrate that the optimization objective is equivalent to strengthening the discriminant property of those transformed features. This similarity learning is independent of the pseudo-labels, boosting the robustness against noisy pseudo-labels. Our method is implemented through a plug-and-play neural network layer, which can be optimized alongside the segmentation model in an end-to-end manner. The model validation focuses on the left atrium segmentation which is a challenging task for atrial fibrillation treatment. Comprehensive experiments demonstrate the effectiveness of the uncertainty-guided similarity learning on semi-supervised segmentation tasks.