DS-SIAUG: A Self-Training Approach Using a Disrupted Student Model for Enhanced Side-Scan Sonar Image Augmentation.

Side-scan sonar is a principal technique for subsea target detection, where the quantity of sonar images of seabed targets significantly influences the accuracy of intelligent target recognition. To expand the number of representative side-scan sonar target image samples, a novel augmentation method employing self-training with a Disrupted Student model is designed (DS-SIAUG). The process begins by inputting a dataset of side-scan sonar target images, followed by augmenting the samples through an adversarial network consisting of the DDPM (Denoising Diffusion Probabilistic Model) and the YOLO (You Only Look Once) detection model. Subsequently, the Disrupted Student model is used to filter out representative target images. These selected images are then reused as a new dataset to repeat the adversarial filtering process. Experimental results indicate that using the Disrupted Student model for selection achieves a target recognition accuracy comparable to manual selection, improving the accuracy of intelligent target recognition by approximately 5% over direct adversarial network augmentation.

Gradual Self-Training via Confidence and Volume Based Domain Adaptation for Multi Dataset Deep Learning-Based Brain Metastases Detection Using Nonlocal Networks on MRI Images.

BACKGROUND: Research suggests that treatment of multiple brain metastases (BMs) with stereotactic radiosurgery shows improvement when metastases are detected early, providing a case for BM detection capabilities on small lesions. PURPOSE: To demonstrate automatic detection of BM on three MRI datasets using a deep learning-based approach. To improve the performance of the network is iteratively co-trained with datasets from different domains. A systematic approach is proposed to prevent catastrophic forgetting during co-training. STUDY TYPE: Retrospective. POPULATION: A total of 156 patients (105 ground truth and 51 pseudo labels) with 1502 BM (BrainMetShare); 121 patients with 722 BM (local); 400 patients with 447 primary gliomas (BrATS). Training/pseudo labels/validation data were distributed 84/51/21 (BrainMetShare). Training/validation data were split: 121/23 (local) and 375/25 (BrATS). FIELD STRENGTH/SEQUENCE: A 5 T and 3 T/T1 spin-echo postcontrast (T1-gradient echo) (BrainMetShare), 3 T/T1 magnetization prepared rapid acquisition gradient echo postcontrast (T1-MPRAGE) (local), 0.5 T, 1 T, and 1.16 T/T1-weighted-fluid-attenuated inversion recovery (T1-FLAIR) (BrATS). ASSESSMENT: The ground truth was manually segmented by two (BrainMetShare) and four (BrATS) radiologists and manually annotated by one (local) radiologist. Confidence and volume based domain adaptation (CAVEAT) method of co-training the three datasets on a 3D nonlocal convolutional neural network (CNN) architecture was implemented to detect BM. STATISTICAL TESTS: The performance was evaluated using sensitivity and false positive rates per patient (FP/patient) and free receiver operating characteristic (FROC) analysis at seven predefined (1/8, 1/4, 1/2, 1, 2, 4, and 8) FPs per scan. RESULTS: The sensitivity and FP/patient from a held-out set registered 0.811 at 2.952 FP/patient (BrainMetShare), 0.74 at 3.130 (local), and 0.723 at 2.240 (BrATS) using the CAVEAT approach with lesions as small as 1 mm being detected. DATA CONCLUSION: Improved sensitivities at lower FP can be achieved by co-training datasets via the CAVEAT paradigm to address the problem of data sparsity. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY STAGE: 2.

Graph-Based Self-Training for Semi-Supervised Deep Similarity Learning.

Semi-supervised learning is a learning pattern that can utilize labeled data and unlabeled data to train deep neural networks. In semi-supervised learning methods, self-training-based methods do not depend on a data augmentation strategy and have better generalization ability. However, their performance is limited by the accuracy of predicted pseudo-labels. In this paper, we propose to reduce the noise in the pseudo-labels from two aspects: the accuracy of predictions and the confidence of the predictions. For the first aspect, we propose a similarity graph structure learning (SGSL) model that considers the correlation between unlabeled and labeled samples, which facilitates the learning of more discriminative features and, thus, obtains more accurate predictions. For the second aspect, we propose an uncertainty-based graph convolutional network (UGCN), which can aggregate similar features based on the learned graph structure in the training phase, making the features more discriminative. It can also output the uncertainty of predictions in the pseudo-label generation phase, generating pseudo-labels only for unlabeled samples with low uncertainty; thus, reducing the noise in the pseudo-labels. Further, a positive and negative self-training framework is proposed, which combines the proposed SGSL model and UGCN into the self-training framework for end-to-end training. In addition, in order to introduce more supervised signals in the self-training process, negative pseudo-labels are generated for unlabeled samples with low prediction confidence, and then the positive and negative pseudo-labeled samples are trained together with a small number of labeled samples to improve the performance of semi-supervised learning. The code is available upon request.

Cross-Referencing Self-Training Network for Sound Event Detection in Audio Mixtures.

Sound event detection is an important facet of audio tagging that aims to identify sounds of interest and define both the sound category and time boundaries for each sound event in a continuous recording. With advances in deep neural networks, there has been tremendous improvement in the performance of sound event detection systems, although at the expense of costly data collection and labeling efforts. In fact, current state-of-the-art methods employ supervised training methods that leverage large amounts of data samples and corresponding labels in order to facilitate identification of sound category and time stamps of events. As an alternative, the current study proposes a semi-supervised method for generating pseudo-labels from unsupervised data using a student-teacher scheme that balances self-training and cross-training. Additionally, this paper explores post-processing which extracts sound intervals from network prediction, for further improvement in sound event detection performance. The proposed approach is evaluated on sound event detection task for the DCASE2020 challenge. The results of these methods on both "validation" and "public evaluation" sets of DESED database show significant improvement compared to the state-of-the art systems in semi-supervised learning.

Neural effects of a short-term virtual reality self-training program to reduce social anxiety.

BACKGROUND: Social anxiety disorder (SAD) is characterized by anxiety regarding social situations, avoidance of external social stimuli, and negative self-beliefs. Virtual reality self-training (VRS) at home may be a good interim modality for reducing social fears before formal treatment. This study aimed to find neurobiological evidence for the therapeutic effect of VRS. METHODS: Fifty-two patients with SAD were randomly assigned to a VRS or waiting list (WL) group. The VRS group received an eight-session VRS program for 2 weeks, whereas the WL group received no intervention. Clinical assessments and functional magnetic resonance imaging scanning with the distress and speech evaluation tasks were repeatedly performed at baseline and after 3 weeks. RESULTS: The post-VRS assessment showed significantly decreased anxiety and avoidance scores, distress index, and negative evaluation index for 'self', but no change in the negative evaluation index for 'other'. Patients showed significant responses to the distress task in various regions, including both sides of the prefrontal regions, occipital regions, insula, and thalamus, and to the speech evaluation task in the bilateral anterior cingulate cortex. Among these, significant neuronal changes after VRS were observed only in the right lingual gyrus and left thalamus. CONCLUSIONS: VRS-induced improvements in the ability to pay attention to social stimuli without avoidance and even positively modulate emotional cues are based on functional changes in the visual cortices and thalamus. Based on these short-term neuronal changes, VRS can be a first intervention option for individuals with SAD who avoid society or are reluctant to receive formal treatment.

A self-training deep neural network for early prediction of cognitive deficits in very preterm infants using brain functional connectome data.

BACKGROUND: Deep learning has been employed using brain functional connectome data for evaluating the risk of cognitive deficits in very preterm infants. Although promising, training these deep learning models typically requires a large amount of labeled data, and labeled medical data are often very difficult and expensive to obtain. OBJECTIVE: This study aimed to develop a self-training deep neural network (DNN) model for early prediction of cognitive deficits at 2 years of corrected age in very preterm infants (gestational age ≤32 weeks) using both labeled and unlabeled brain functional connectome data. MATERIALS AND METHODS: We collected brain functional connectome data from 343 very preterm infants at a mean (standard deviation) postmenstrual age of 42.7 (2.5) weeks, among whom 103 children had a cognitive assessment at 2 years (i.e. labeled data), and the remaining 240 children had not received 2-year assessments at the time this study was conducted (i.e. unlabeled data). To develop a self-training DNN model, we built an initial student model using labeled brain functional connectome data. Then, we applied the trained model as a teacher model to generate pseudo-labels for unlabeled brain functional connectome data. Next, we combined labeled and pseudo-labeled data to train a new student model. We iterated this procedure to obtain the best student model for the early prediction task in very preterm infants. RESULTS: In our cross-validation experiments, the proposed self-training DNN model achieved an accuracy of 71.0%, a specificity of 71.5%, a sensitivity of 70.4% and an area under the curve of 0.75, significantly outperforming transfer learning models through pre-training approaches. CONCLUSION: We report the first self-training prognostic study in very preterm infants, efficiently utilizing a small amount of labeled data with a larger share of unlabeled data to aid the model training. The proposed technique is expected to facilitate deep learning with insufficient training data.

Self-Trained Deep Forest with Limited Samples for Urban Impervious Surface Area Extraction in Arid Area Using Multispectral and PolSAR Imageries.

Impervious surface area (ISA) has been recognized as a significant indicator for evaluating levels of urbanization and the quality of urban ecological environments. ISA extraction methods based on supervised classification usually rely on a large number of manually labeled samples, the production of which is a time-consuming and labor-intensive task. Furthermore, in arid areas, man-made objects are easily confused with bare land due to similar spectral responses. To tackle these issues, a self-trained deep-forest (STDF)-based ISA extraction method is proposed which exploits the complementary information contained in multispectral and polarimetric synthetic aperture radar (PolSAR) images using limited numbers of samples. In detail, this method consists of three major steps. First, multi-features, including spectral, spatial and polarimetric features, are extracted from Sentinel-2 multispectral and Chinese GaoFen-3 (GF-3) PolSAR images; secondly, a deep forest (DF) model is trained in a self-training manner using a limited number of samples for ISA extraction; finally, ISAs (in this case, in three major cities located in Central Asia) are extracted and comparatively evaluated. The experimental results from the study areas of Bishkek, Tashkent and Nursultan demonstrate the effectiveness of the proposed method, with an overall accuracy (OA) above 95% and a Kappa coefficient above 0.90.

Deep neural-network based optimization for the design of a multi-element surface magnet for MRI applications.

We present a combination of a CNN-based encoder with an analytical forward map for solving inverse problems. We call it an encoder-analytic (EA) hybrid model. It does not require a dedicated training dataset and can train itself from the connected forward map in a direct learning fashion. A separate regularization term is not required either, since the forward map also acts as a regularizer. As it is not a generalization model it does not suffer from overfitting. We further show that the model can be customized to either find a specific target solution or one that follows a given heuristic. As an example, we apply this approach to the design of a multi-element surface magnet for low-field magnetic resonance imaging (MRI). We further show that the EA model can outperform the benchmark genetic algorithm model currently used for magnet design in MRI, obtaining almost 10 times better results.

Attitudes toward overtime work and self-training: A survey on obstetricians and gynecologists in Japan.

AIM: The Ministry of Health, Labour, and Welfare of Japan proposed a regulation of overtime work as a reform in work style. However, the regulation may deteriorate the quality of medical services due to the reduction in training time. Thus, the study aimed to reveal perceptions in terms of generation gaps in views on self-training and overtime work, among members of the Japan Society of Obstetrics and Gynecology (JSOG). METHODS: A web-based, self-administered questionnaire survey was conducted among members of the JSOG. In total, 1256 respondents were included in the analysis. Data were collected on age, sex, experience as a medical doctor, location of workplace, work style, the type of main workplace, and number of full-time doctors in the main workplace. The study examined the attitudes of the respondents toward overtime work and self-training. The respondents were categorized based on experience as a medical doctor. RESULTS: According to years of experience, 112 (8.9%), 226 (18.0%), 383 (30.5%), 535 (42.6%) doctors have been working for ≤5, 6-10, 11-19, and ≥ 20 years, respectively. Although 54.5% of doctors with ≤5 years of experience expected the regulation on working hours to improve the quality of medical services, those with ≥20 years of experience expressed potential deterioration. After adjusting for covariates, more years of experience were significantly related with the expectation of deterioration in the quality of medical services. CONCLUSIONS: The study revealed a generation gap in the views about self-training and overtime work among obstetricians and gynecologists in Japan.

An Improved Self-Training Method for Positive Unlabeled Time Series Classification Using DTW Barycenter Averaging.

Traditional supervised time series classification (TSC) tasks assume that all training data are labeled. However, in practice, manually labelling all unlabeled data could be very time-consuming and often requires the participation of skilled domain experts. In this paper, we concern with the positive unlabeled time series classification problem (PUTSC), which refers to automatically labelling the large unlabeled set U based on a small positive labeled set PL. The self-training (ST) is the most widely used method for solving the PUTSC problem and has attracted increased attention due to its simplicity and effectiveness. The existing ST methods simply employ the one-nearest-neighbor (1NN) formula to determine which unlabeled time-series should be labeled. Nevertheless, we note that the 1NN formula might not be optimal for PUTSC tasks because it may be sensitive to the initial labeled data located near the boundary between the positive and negative classes. To overcome this issue, in this paper we propose an exploratory methodology called ST-average. Unlike conventional ST-based approaches, ST-average utilizes the average sequence calculated by DTW barycenter averaging technique to label the data. Compared with any individuals in PL set, the average sequence is more representative. Our proposal is insensitive to the initial labeled data and is more reliable than existing ST-based methods. Besides, we demonstrate that ST-average can naturally be implemented along with many existing techniques used in original ST. Experimental results on public datasets show that ST-average performs better than related popular methods.

Robot-Assisted Gait Self-Training: Assessing the Level Achieved.

This paper presents the technological status of robot-assisted gait self-training under real clinical environment conditions. A successful rehabilitation after surgery in hip endoprosthetics comprises self-training of the lessons taught by physiotherapists. While doing this, immediate feedback to the patient about deviations from the expected physiological gait pattern during training is important. Hence, the Socially Assistive Robot (SAR) developed for this type of training employs task-specific, user-centered navigation and autonomous, real-time gait feature classification techniques to enrich the self-training through companionship and timely corrective feedback. The evaluation of the system took place during user tests in a hospital from the point of view of technical benchmarking, considering the therapists' and patients' point of view with regard to training motivation and from the point of view of initial findings on medical efficacy as a prerequisite from an economic perspective. In this paper, the following research questions were primarily considered: Does the level of technology achieved enable autonomous use in everyday clinical practice? Has the gait pattern of patients who used additional robot-assisted gait self-training for several days been changed or improved compared to patients without this training? How does the use of a SAR-based self-training robot affect the motivation of the patients?

[Development of Self Training Device for Lung Function Compliance Guided by Meridians].

OBJECTIVE: To develop a self deep breathing training device which can improve lung function compliance and blood oxygen saturation. METHODS: The device consists of four parts:flow tube, measuring cylinder, mobile phone holder and meridian guidance audio-visual synthesis training software. The flow tube measures the flow rate of inhaled gas, the metering cylinder measures the total amount of inhaled gas, and the mobile phone rack is equipped with a mobile phone storing the meridian guidance audio-visual synthesis training software. RESULTS: The device is reasonable in structure and flexible in operation, which can meet the requirements of self deep inspiration training under the guidance of training module. CONCLUSIONS: Deep inspiration training under the guidance of guidance training module can form "deep and slow" abdominal breathing, and then improve lung function.

Robust Semi-Supervised Traffic Sign Recognition via Self-Training and Weakly-Supervised Learning.

Traffic sign recognition is a classification problem that poses challenges for computer vision and machine learning algorithms. Although both computer vision and machine learning techniques have constantly been improved to solve this problem, the sudden rise in the number of unlabeled traffic signs has become even more challenging. Large data collation and labeling are tedious and expensive tasks that demand much time, expert knowledge, and fiscal resources to satisfy the hunger of deep neural networks. Aside from that, the problem of having unbalanced data also poses a greater challenge to computer vision and machine learning algorithms to achieve better performance. These problems raise the need to develop algorithms that can fully exploit a large amount of unlabeled data, use a small amount of labeled samples, and be robust to data imbalance to build an efficient and high-quality classifier. In this work, we propose a novel semi-supervised classification technique that is robust to small and unbalanced data. The framework integrates weakly-supervised learning and self-training with self-paced learning to generate attention maps to augment the training set and utilizes a novel pseudo-label generation and selection algorithm to generate and select pseudo-labeled samples. The method improves the performance by: (1) normalizing the class-wise confidence levels to prevent the model from ignoring hard-to-learn samples, thereby solving the imbalanced data problem; (2) jointly learning a model and optimizing pseudo-labels generated on unlabeled data; and (3) enlarging the training set to satisfy the hunger of deep learning models. Extensive evaluations on two public traffic sign recognition datasets demonstrate the effectiveness of the proposed technique and provide a potential solution for practical applications.

A Semi-Automatic Annotation Approach for Human Activity Recognition.

Modern smartphones and wearables often contain multiple embedded sensors which generate significant amounts of data. This information can be used for body monitoring-based areas such as healthcare, indoor location, user-adaptive recommendations and transportation. The development of Human Activity Recognition (HAR) algorithms involves the collection of a large amount of labelled data which should be annotated by an expert. However, the data annotation process on large datasets is expensive, time consuming and difficult to obtain. The development of a HAR approach which requires low annotation effort and still maintains adequate performance is a relevant challenge. We introduce a Semi-Supervised Active Learning (SSAL) based on Self-Training (ST) approach for Human Activity Recognition to partially automate the annotation process, reducing the annotation effort and the required volume of annotated data to obtain a high performance classifier. Our approach uses a criterion to select the most relevant samples for annotation by the expert and propagate their label to the most confident samples. We present a comprehensive study comparing supervised and unsupervised methods with our approach on two datasets composed of daily living activities. The results showed that it is possible to reduce the required annotated data by more than 89% while still maintaining an accurate model performance.

High-fidelity simulation self-training enables novice bronchoscopists to acquire basic bronchoscopy skills comparable to their moderately and highly experienced counterparts.

BACKGROUND: We sought to determine whether a self-training program on a high-fidelity flexible bronchoscopy (FB) simulator would allow residents who were novices in bronchoscopy to acquire competencies similar to those of experienced bronchoscopists as concerns the visualization of the bronchial tree and the identification of its anatomical elements. METHODS: We performed a prospective cohort study, categorizing bronchoscopists into three groups according to their experience level: novice (Group A, no FBs performed, n = 8), moderate (Group B, 30 ≤ FBs performed ≤200, n = 17) or high (Group C, > 200 FBs performed, n = 9). All were initially evaluated on their ability to perform on a high-fidelity FB simulator a complete visualization/identification of the bronchial tree in the least amount of time possible. The residents in Group A then completed a simulation-based self-training program and underwent a final evaluation thereafter. RESULTS: The median total procedure time for Group A fell from 561 s (IQR = 134) in the initial evaluation to 216 s (IQR = 257) in the final evaluation (P = 0.002). The visualization and identification scores for Group A also improved significantly in the final evaluation. Resultantly, the overall performance score for Group A climbed from 5.9% (IQR = 5.1) before self-training to 25.5% (IQR = 26.3) after (P = 0.002), thus becoming comparable to the overall performance scores of Group B (25.3%, IQR = 13.8) and Group C (22.2%, IQR = 5.5). CONCLUSIONS: Novice bronchoscopists who self-train on a high-fidelity simulator acquire basic competencies similar to those of moderately or even highly experienced bronchoscopists. High-fidelity simulation should be rapidly integrated within the learning curriculum and replace traditional, in-patient learning methods.

Experiences of Older Adults with Mild Cognitive Impairment from Cognitive Self-Training Using Touchscreen Tablets.

Background: "Tablet Enhancement of Cognition and Health" (TECH) is a cognitive intervention that includes two components: 5 weeks of daily self-training using puzzle-game apps on a touch screen tablet and weekly group sessions. This study aimed to (i) explore experiences of older adults with mild cognitive impairment (MCI) following their participation in TECH, (ii) identify hindering and enabling factors to self-training, and (iii) describe participants' perceived and objective cognitive changes and examine factors associated with their satisfaction from TECH. Materials and Methods: We used quantitative and qualitative measures; a phenomenological qualitative design using focus groups and interviews of 14 older adults with MCI and a focus group of the TECH facilitators. Satisfaction with TECH, self-training time, and perceived and objective cognitive changes (using the Montreal Cognitive Assessment) were evaluated. Results: Qualitative data were classified into three categories: Memory problems, Hindering and enabling factors to self-training, and Meaningful group sessions. The TECH facilitators reported positive changes, less cognitive complaints, and commitment and satisfaction of the participants. Participants reported overall satisfaction from TECH and performed a median interquartile range of 22.6 (19.9-42.8) self-training hours. Higher satisfaction was correlated with a higher objective cognitive change (r = 0.95, P < 0.01) and less training time (r = -0.91, P < 0.01). Discussion and Conclusions: Participants in the current study actively engaged in daily self-training using touch screen-tablet-puzzle-game and functional apps, driven by both internal and external motivators. Despite the lack of cognitive improvement, they expressed satisfaction with their participation in TECH. Therefore, encouraging older adults to engage in meaningful cognitive stimulating activities is recommended.

Image-level supervision and self-training for transformer-based cross-modality tumor segmentation.

Deep neural networks are commonly used for automated medical image segmentation, but models will frequently struggle to generalize well across different imaging modalities. This issue is particularly problematic due to the limited availability of annotated data, both in the target as well as the source modality, making it difficult to deploy these models on a larger scale. To overcome these challenges, we propose a new semi-supervised training strategy called MoDATTS. Our approach is designed for accurate cross-modality 3D tumor segmentation on unpaired bi-modal datasets. An image-to-image translation strategy between modalities is used to produce synthetic but annotated images and labels in the desired modality and improve generalization to the unannotated target modality. We also use powerful vision transformer architectures for both image translation (TransUNet) and segmentation (Medformer) tasks and introduce an iterative self-training procedure in the later task to further close the domain gap between modalities, thus also training on unlabeled images in the target modality. MoDATTS additionally allows the possibility to exploit image-level labels with a semi-supervised objective that encourages the model to disentangle tumors from the background. This semi-supervised methodology helps in particular to maintain downstream segmentation performance when pixel-level label scarcity is also present in the source modality dataset, or when the source dataset contains healthy controls. The proposed model achieves superior performance compared to other methods from participating teams in the CrossMoDA 2022 vestibular schwannoma (VS) segmentation challenge, as evidenced by its reported top Dice score of 0.87±0.04 for the VS segmentation. MoDATTS also yields consistent improvements in Dice scores over baselines on a cross-modality adult brain gliomas segmentation task composed of four different contrasts from the BraTS 2020 challenge dataset, where 95% of a target supervised model performance is reached when no target modality annotations are available. We report that 99% and 100% of this maximum performance can be attained if 20% and 50% of the target data is additionally annotated, which further demonstrates that MoDATTS can be leveraged to reduce the annotation burden.

STS-TransUNet: Semi-supervised Tooth Segmentation Transformer U-Net for dental panoramic image.

In this paper, we introduce a novel deep learning method for dental panoramic image segmentation, which is crucial in oral medicine and orthodontics for accurate diagnosis and treatment planning. Traditional methods often fail to effectively combine global and local context, and struggle with unlabeled data, limiting performance in varied clinical settings. We address these issues with an advanced TransUNet architecture, enhancing feature retention and utilization by connecting the input and output layers directly. Our architecture further employs spatial and channel attention mechanisms in the decoder segments for targeted region focus, and deep supervision techniques to overcome the vanishing gradient problem for more efficient training. Additionally, our network includes a self-learning algorithm using unlabeled data, boosting generalization capabilities. Named the Semi-supervised Tooth Segmentation Transformer U-Net (STS-TransUNet), our method demonstrated superior performance on the MICCAI STS-2D dataset, proving its effectiveness and robustness in tooth segmentation tasks.

Enhanced Domain Adaptation for Foot Ulcer Segmentation Through Mixing Self-Trained Weak Labels.

Wound management requires the measurement of the wound parameters such as its shape and area. However, computerized analysis of the wound suffers the challenge of inexact segmentation of the wound images due to limited or inaccurate labels. It is a common scenario that the source domain provides an abundance of labeled data, while the target domain provides only limited labels. To overcome this, we propose a novel approach that combines self-training learning and mixup augmentation. The neural network is trained on the source domain to generate weak labels on the target domain via the self-training process. In the second stage, generated labels are mixed up with labels from the source domain to retrain the neural network and enhance generalization across diverse datasets. The efficacy of our approach was evaluated using the DFUC 2022, FUSeg, and RMIT datasets, demonstrating substantial improvements in segmentation accuracy and robustness across different data distributions. Specifically, in single-domain experiments, segmentation on the DFUC 2022 dataset scored a dice score of 0.711, while the score on the FUSeg dataset achieved 0.859. For domain adaptation, when these datasets were used as target datasets, the dice scores were 0.714 for DFUC 2022 and 0.561 for FUSeg.

A 2.5D Self-Training Strategy for Carotid Artery Segmentation in T1-Weighted Brain Magnetic Resonance Images.

Precise annotations for large medical image datasets can be time-consuming. Additionally, when dealing with volumetric regions of interest, it is typical to apply segmentation techniques on 2D slices, compromising important information for accurately segmenting 3D structures. This study presents a deep learning pipeline that simultaneously tackles both challenges. Firstly, to streamline the annotation process, we employ a semi-automatic segmentation approach using bounding boxes as masks, which is less time-consuming than pixel-level delineation. Subsequently, recursive self-training is utilized to enhance annotation quality. Finally, a 2.5D segmentation technique is adopted, wherein a slice of a volumetric image is segmented using a pseudo-RGB image. The pipeline was applied to segment the carotid artery tree in T1-weighted brain magnetic resonance images. Utilizing 42 volumetric non-contrast T1-weighted brain scans from four datasets, we delineated bounding boxes around the carotid arteries in the axial slices. Pseudo-RGB images were generated from these slices, and recursive segmentation was conducted using a Res-Unet-based neural network architecture. The model's performance was tested on a separate dataset, with ground truth annotations provided by a radiologist. After recursive training, we achieved an Intersection over Union (IoU) score of (0.68 ± 0.08) on the unseen dataset, demonstrating commendable qualitative results.