Interactive Few-Shot Learning: Limited Supervision, Better Medical Image Segmentation

A systematic review of few-shot learning in medical imaging

The lack of annotated medical images limits the performance of deep learning models, which usually need large-scale labelled datasets. Few-shot learning techniques can reduce data scarcity issues and enhance medical image analysis speed and robustness. This systematic review gives a comprehensive overview of few-shot learning methods for medical image analysis, aiming to establish a standard methodological pipeline for future research reference. With a particular emphasis on the role of meta-learning, we analysed 80 relevant articles published from 2018 to 2023, conducting a risk of bias assessment and extracting relevant information, especially regarding the employed learning techniques. From this, we delineated a comprehensive methodological pipeline shared among all studies. In addition, we performed a statistical analysis of the studies' results concerning the clinical task and the meta-learning method employed while also presenting supplemental information such as imaging modalities and model robustness evaluation techniques. We discussed the findings of our analysis, providing a deep insight into the limitations of the state-of-the-art methods and the most promising approaches. Drawing on our investigation, we yielded recommendations on potential future research directions aiming to bridge the gap between research and clinical practice.

Reducing Manual Annotation Costs for Cell Segmentation by Upgrading Low-Quality Annotations

Deep-learning algorithms for cell segmentation typically require large data sets with high-quality annotations to be trained with. However, the annotation cost for obtaining such sets may prove to be prohibitively expensive. Our work aims to reduce the time necessary to create high-quality annotations of cell images by using a relatively small well-annotated data set for training a convolutional neural network to upgrade lower-quality annotations, produced at lower annotation costs. We investigate the performance of our solution when upgrading the annotation quality for labels affected by three types of annotation error: omission, inclusion, and bias. We observe that our method can upgrade annotations affected by high error levels from 0.3 to 0.9 Dice similarity with the ground-truth annotations. We also show that a relatively small well-annotated set enlarged with samples with upgraded annotations can be used to train better-performing cell segmentation networks compared to training only on the well-annotated set. Moreover, we present a use case where our solution can be successfully employed to increase the quality of the predictions of a segmentation network trained on just 10 annotated samples.

Few-Shot Learning for Medical Image Segmentation Using 3D U-Net and Model-Agnostic Meta-Learning (MAML)

Deep learning has attained state-of-the-art results in general image segmentation problems; however, it requires a substantial number of annotated images to achieve the desired outcomes. In the medical field, the availability of annotated images is often limited. To address this challenge, few-shot learning techniques have been successfully adapted to rapidly generalize to new tasks with only a few samples, leveraging prior knowledge. In this paper, we employ a gradient-based method known as Model-Agnostic Meta-Learning (MAML) for medical image segmentation. MAML is a meta-learning algorithm that quickly adapts to new tasks by updating a model's parameters based on a limited set of training samples. Additionally, we use an enhanced 3D U-Net as the foundational network for our models. The enhanced 3D U-Net is a convolutional neural network specifically designed for medical image segmentation. We evaluate our approach on the TotalSegmentator dataset, considering a few annotated images for four tasks: liver, spleen, right kidney, and left kidney. The results demonstrate that our approach facilitates rapid adaptation to new tasks using only a few annotated images. In 10-shot settings, our approach achieved mean dice coefficients of 93.70%, 85.98%, 81.20%, and 89.58% for liver, spleen, right kidney, and left kidney segmentation, respectively. In five-shot sittings, the approach attained mean Dice coefficients of 90.27%, 83.89%, 77.53%, and 87.01% for liver, spleen, right kidney, and left kidney segmentation, respectively. Finally, we assess the effectiveness of our proposed approach on a dataset collected from a local hospital. Employing five-shot sittings, we achieve mean Dice coefficients of 90.62%, 79.86%, 79.87%, and 78.21% for liver, spleen, right kidney, and left kidney segmentation, respectively.

Few-Shot Medical Image Segmentation via Generating Multiple Representative Descriptors

Automatic medical image segmentation has witnessed significant development with the success of large models on massive datasets. However, acquiring and annotating vast medical image datasets often proves to be impractical due to the time consumption, specialized expertise requirements, and compliance with patient privacy standards, etc. As a result, Few-shot Medical Image Segmentation (FSMIS) has become an increasingly compelling research direction. Conventional FSMIS methods usually learn prototypes from support images and apply nearest-neighbor searching to segment the query images. However, only a single prototype cannot well represent the distribution of each class, thus leading to restricted performance. To address this problem, we propose to Generate Multiple Representative Descriptors (GMRD), which can comprehensively represent the commonality within the corresponding class distribution. In addition, we design a Multiple Affinity Maps based Prediction (MAMP) module to fuse the multiple affinity maps generated by the aforementioned descriptors. Furthermore, to address intra-class variation and enhance the representativeness of descriptors, we introduce two novel losses. Notably, our model is structured as a dual-path design to achieve a balance between foreground and background differences in medical images. Extensive experiments on four publicly available medical image datasets demonstrate that our method outperforms the state-of-the-art methods, and the detailed analysis also verifies the effectiveness of our designed module.

Linear semantic transformation for semi-supervised medical image segmentation

Medical image segmentation is a focus research and foundation in developing intelligent medical systems. Recently, deep learning for medical image segmentation has become a standard process and succeeded significantly, promoting the development of reconstruction, and surgical planning of disease diagnosis. However, semantic learning is often inefficient owing to the lack of supervision of feature maps, resulting in that high-quality segmentation models always rely on numerous and accurate data annotations. Learning robust semantic representation in latent spaces remains a challenge. In this paper, we propose a novel semi-supervised learning framework to learn vital attributes in medical images, which constructs generalized representation from diverse semantics to realize medical image segmentation. We first build a self-supervised learning part that achieves context recovery by reconstructing space and intensity of medical images, which conduct semantic representation for feature maps. Subsequently, we combine semantic-rich feature maps and utilize simple linear semantic transformation to convert them into image segmentation. The proposed framework was tested using five medical segmentation datasets. Quantitative assessments indicate the highest scores of our method on IXI (73.78%), ScaF (47.50%), COVID-19-Seg (50.72%), PC-Seg (65.06%), and Brain-MR (72.63%) datasets. Finally, we compared our method with the latest semi-supervised learning methods and obtained 77.15% and 75.22% DSC values, respectively, ranking first on two representative datasets. The experimental results not only proved that the proposed linear semantic transformation was effectively applied to medical image segmentation, but also presented its simplicity and ease-of-use to pursue robust segmentation in semi-supervised learning. Our code is now open at: https://github.com/QingYunA/Linear-Semantic-Transformation-for-Semi-Supervised-Medical-Image-Segmentation.

The Role of AI in Hospitals and Clinics: Transforming Healthcare in the 21st Century

As healthcare systems around the world face challenges such as escalating costs, limited access, and growing demand for personalized care, artificial intelligence (AI) is emerging as a key force for transformation. This review is motivated by the urgent need to harness AI's potential to mitigate these issues and aims to critically assess AI's integration in different healthcare domains. We explore how AI empowers clinical decision-making, optimizes hospital operation and management, refines medical image analysis, and revolutionizes patient care and monitoring through AI-powered wearables. Through several case studies, we review how AI has transformed specific healthcare domains and discuss the remaining challenges and possible solutions. Additionally, we will discuss methodologies for assessing AI healthcare solutions, ethical challenges of AI deployment, and the importance of data privacy and bias mitigation for responsible technology use. By presenting a critical assessment of AI's transformative potential, this review equips researchers with a deeper understanding of AI's current and future impact on healthcare. It encourages an interdisciplinary dialogue between researchers, clinicians, and technologists to navigate the complexities of AI implementation, fostering the development of AI-driven solutions that prioritize ethical standards, equity, and a patient-centered approach.

Delicately Reinforced k-Nearest Neighbor Classifier Combined With Expert Knowledge Applied to Abnormity Forecast in Electrolytic Cell

As the profit and safety requirements become higher and higher, it is more and more necessary to realize an advanced intelligent analysis for abnormity forecast of the synthetical balance of material and energy (AF-SBME) on aluminum reduction cells (ARCs). Without loss of generality, AF-SBME belongs to classification problems. Its advanced intelligent analysis can be realized by high-performance data-driven classifiers. However, AF-SBME has some difficulties, including a high requirement for interpretability of data-driven classifiers, a small number, and decreasing-over-time correctness of training samples. In this article, based on a preferable data-driven classifier, which is called a reinforced k -nearest neighbor (R-KNN) classifier, a delicately R-KNN combined with expert knowledge (DR-KNN/CE) is proposed. It improves R-KNN in two ways, including using expert knowledge as external assistance and enhancing self-ability to mine and synthesize data knowledge. The related experiments on AF-SBME, where the relevant data are directly sampled from practical production, have demonstrated that the proposed DR-KNN/CE not only makes an effective improvement for R-KNN, but also has a more advanced performance compared with other existing high-performance data-driven classifiers.

Few-shot biomedical image segmentation using diffusion models: Beyond image generation

BACKGROUND

Medical image analysis pipelines often involve segmentation, which requires a large amount of annotated training data, which is time-consuming and costly. To address this issue, we proposed leveraging generative models to achieve few-shot image segmentation.

METHODS

We trained a denoising diffusion probabilistic model (DDPM) on 480,407 pelvis radiographs to generate 256 ✕ 256 px synthetic images. The DDPM was conditioned on demographic and radiologic characteristics and was rigorously validated by domain experts and objective image quality metrics (Frechet inception distance [FID] and inception score [IS]). For the next step, three landmarks (greater trochanter [GT], lesser trochanter [LT], and obturator foramen [OF]) were annotated on 45 real-patient radiographs; 25 for training and 20 for testing. To extract features, each image was passed through the pre-trained DDPM at three timesteps and for each pass, features from specific blocks were extracted. The features were concatenated with the real image to form an image with 4225 channels. The feature-set was broken into random patches, which were fed to a U-Net. Dice Similarity Coefficient (DSC) was used to compare the performance with a vanilla U-Net trained on radiographs.

RESULTS

Expert accuracy was 57.5 % in determining real versus generated images, while the model reached an FID = 7.2 and IS = 210. The segmentation UNet trained on the 20 feature-sets achieved a DSC of 0.90, 0.84, and 0.61 for OF, GT, and LT segmentation, respectively, which was at least 0.30 points higher than the naively trained model.

CONCLUSION

We demonstrated the applicability of DDPMs as feature extractors, facilitating medical image segmentation with few annotated samples.

Dynamic feature splicing for few-shot rare disease diagnosis

Annotated images for rare disease diagnosis are extremely hard to collect. Therefore, identifying rare diseases under a few-shot learning (FSL) setting is significant. Existing FSL methods transfer useful and global knowledge from base classes with abundant training samples to enrich features of novel classes with few training samples, but still face difficulties when being applied to medical images due to the complex lesion characteristics and large intra-class variance. In this paper, we propose a dynamic feature splicing (DNFS) framework for few-shot rare disease diagnosis. Under DNFS, both low-level features (i.e., the output of three convolutional blocks) and high-level features (i.e., the output of the last fully connected layer) of novel classes are dynamically enriched. We construct the position coherent DNFS (P-DNFS) module to perform low-level feature splicing, where a lesion-oriented Transformer is designed to detect lesion regions. Thus, novel-class channels are replaced by similar base-class channels within the detected lesion regions to achieve disease-related feature enrichment. We also devise a semantic coherent DNFS (S-DNFS) module to perform high-level feature splicing. It explores cross-image channel relations and selects base-class channels with semantic consistency for explicit knowledge transfer. Both low-level and high-level feature splicings are performed dynamically and iteratively. Consequently, abundant spliced features are generated for disease diagnosis, leading to more accurate decision boundary and improved diagnosis performance. Extensive experiments have been conducted on three medical image classification datasets. Our results suggest that the proposed DNFS achieves superior performance against state-of-the-art approaches.

Continual Nuclei Segmentation via Prototype-Wise Relation Distillation and Contrastive Learning

Deep learning models have achieved remarkable success in multi-type nuclei segmentation. These models are mostly trained at once with the full annotation of all types of nuclei available, while lack the ability of continually learning new classes due to the problem of catastrophic forgetting. In this paper, we study the practical and important class-incremental continual learning problem, where the model is incrementally updated to new classes without accessing to previous data. We propose a novel continual nuclei segmentation method, to avoid forgetting knowledge of old classes and facilitate the learning of new classes, by achieving feature-level knowledge distillation with prototype-wise relation distillation and contrastive learning. Concretely, prototype-wise relation distillation imposes constraints on the inter-class relation similarity, encouraging the encoder to extract similar class distribution for old classes in the feature space. Prototype-wise contrastive learning with a hard sampling strategy enhances the intra-class compactness and inter-class separability of features, improving the performance on both old and new classes. Experiments on two multi-type nuclei segmentation benchmarks, i.e., MoNuSAC and CoNSeP, demonstrate the effectiveness of our method with superior performance over many competitive methods. Codes are available at https://github.com/zzw-szu/CoNuSeg.

Self-supervised-RCNN for medical image segmentation with limited data annotation

Many successful methods developed for medical image analysis based on machine learning use supervised learning approaches, which often require large datasets annotated by experts to achieve high accuracy. However, medical data annotation is time-consuming and expensive, especially for segmentation tasks. To overcome the problem of learning with limited labeled medical image data, an alternative deep learning training strategy based on self-supervised pretraining on unlabeled imaging data is proposed in this work. For the pretraining, different distortions are arbitrarily applied to random areas of unlabeled images. Next, a Mask-RCNN architecture is trained to localize the distortion location and recover the original image pixels. This pretrained model is assumed to gain knowledge of the relevant texture in the images from the self-supervised pretraining on unlabeled imaging data. This provides a good basis for fine-tuning the model to segment the structure of interest using a limited amount of labeled training data. The effectiveness of the proposed method in different pretraining and fine-tuning scenarios was evaluated based on the Osteoarthritis Initiative dataset with the aim of segmenting effusions in MRI datasets of the knee. Applying the proposed self-supervised pretraining method improved the Dice score by up to 18% compared to training the models using only the limited annotated data. The proposed self-supervised learning approach can be applied to many other medical image analysis tasks including anomaly detection, segmentation, and classification.

Robust Prototypical Few-Shot Organ Segmentation With Regularized Neural-ODEs

Despite the tremendous progress made by deep learning models in image semantic segmentation, they typically require large annotated examples, and increasing attention is being diverted to problem settings like Few-Shot Learning (FSL) where only a small amount of annotation is needed for generalisation to novel classes. This is especially seen in medical domains where dense pixel-level annotations are expensive to obtain. In this paper, we propose Regularized Prototypical Neural Ordinary Differential Equation (R-PNODE), a method that leverages intrinsic properties of Neural-ODEs, assisted and enhanced by additional cluster and consistency losses to perform Few-Shot Segmentation (FSS) of organs. R-PNODE constrains support and query features from the same classes to lie closer in the representation space thereby improving the performance over the existing Convolutional Neural Network (CNN) based FSS methods. We further demonstrate that while many existing Deep CNN-based methods tend to be extremely vulnerable to adversarial attacks, R-PNODE exhibits increased adversarial robustness for a wide array of these attacks. We experiment with three publicly available multi-organ segmentation datasets in both in-domain and cross-domain FSS settings to demonstrate the efficacy of our method. In addition, we perform experiments with seven commonly used adversarial attacks in various settings to demonstrate R-PNODE's robustness. R-PNODE outperforms the baselines for FSS by significant margins and also shows superior performance for a wide array of attacks varying in intensity and design.

Semi-automatic muscle segmentation in MR images using deep registration-based label propagation

Fully automated approaches based on convolutional neural networks have shown promising performances on muscle segmentation from magnetic resonance (MR) images, but still rely on an extensive amount of training data to achieve valuable results. Muscle segmentation for pediatric and rare diseases cohorts is therefore still often done manually. Producing dense delineations over 3D volumes remains a time-consuming and tedious task, with significant redundancy between successive slices. In this work, we propose a segmentation method relying on registration-based label propagation, which provides 3D muscle delineations from a limited number of annotated 2D slices. Based on an unsupervised deep registration scheme, our approach ensures the preservation of anatomical structures by penalizing deformation compositions that do not produce consistent segmentation from one annotated slice to another. Evaluation is performed on MR data from lower leg and shoulder joints. Results demonstrate that the proposed few-shot multi-label segmentation model outperforms state-of-the-art techniques.

Defensive Few-Shot Learning

This article investigates a new challenging problem called defensive few-shot learning in order to learn a robust few-shot model against adversarial attacks. Simply applying the existing adversarial defense methods to few-shot learning cannot effectively solve this problem. This is because the commonly assumed sample-level distribution consistency between the training and test sets can no longer be met in the few-shot setting. To address this situation, we develop a general defensive few-shot learning (DFSL) framework to answer the following two key questions: (1) how to transfer adversarial defense knowledge from one sample distribution to another? (2) how to narrow the distribution gap between clean and adversarial examples under the few-shot setting? To answer the first question, we propose an episode-based adversarial training mechanism by assuming a task-level distribution consistency to better transfer the adversarial defense knowledge. As for the second question, within each few-shot task, we design two kinds of distribution consistency criteria to narrow the distribution gap between clean and adversarial examples from the feature-wise and prediction-wise perspectives, respectively. Extensive experiments demonstrate that the proposed framework can effectively make the existing few-shot models robust against adversarial attacks. Code is available at https://github.com/WenbinLee/DefensiveFSL.git.

Learning what and where to segment: A new perspective on medical image few-shot segmentation

Traditional medical image segmentation methods based on deep learning require experts to provide extensive manual delineations for model training. Few-shot learning aims to reduce the dependence on the scale of training data but usually shows poor generalizability to the new target. The trained model tends to favor the training classes rather than being absolutely class-agnostic. In this work, we propose a novel two-branch segmentation network based on unique medical prior knowledge to alleviate the above problem. Specifically, we explicitly introduce a spatial branch to provide the spatial information of the target. In addition, we build a segmentation branch based on the classical encoder-decoder structure in supervised learning and integrate prototype similarity and spatial information as prior knowledge. To achieve effective information integration, we propose an attention-based fusion module (AF) that enables the content interaction of decoder features and prior knowledge. Experiments on an echocardiography dataset and an abdominal MRI dataset show that the proposed model achieves substantial improvements over state-of-the-art methods. Moreover, some results are comparable to those of the fully supervised model. The source code is available at github.com/warmestwind/RAPNet.

Enhanced Patient-Centricity: How the Biopharmaceutical Industry Is Optimizing Patient Care through AI/ML/DL

Technologies utilizing cutting-edge methodologies, including artificial intelligence (AI), machine learning (ML) and deep learning (DL), present powerful opportunities to help evaluate, predict, and improve patient outcomes by drawing insights from real-world data (RWD) generated during medical care. They played a role during and following the Coronavirus Disease 2019 (COVID-19) pandemic by helping protect healthcare providers, prioritize care for vulnerable populations, predict disease trends, and find optimal therapies. Potential applications across therapeutic areas include diagnosis, disease management and patient journey mapping. Use of fit-for-purpose datasets for ML models is seeing growth and may potentially help additional enterprises develop AI strategies. However, biopharmaceutical companies often face specific challenges, including multi-setting data, system interoperability, data governance, and patient privacy requirements. There remains a need for evolving regulatory frameworks, operating models, and data governance to enable further developments and additional research. We explore recent literature and examine the hurdles faced by researchers in the biopharmaceutical industry to fully realize the promise of AI/ML/DL for patient-centric purposes.

Explainable deep learning algorithm for distinguishing incomplete Kawasaki disease by coronary artery lesions on echocardiographic imaging

BACKGROUND AND OBJECTIVE

Incomplete Kawasaki disease (KD) has often been misdiagnosed due to a lack of the clinical manifestations of classic KD. However, it is associated with a markedly higher prevalence of coronary artery lesions. Identifying coronary artery lesions by echocardiography is important for the timely diagnosis of and favorable outcomes in KD. Moreover, similar to KD, coronavirus disease 2019, currently causing a worldwide pandemic, also manifests with fever; therefore, it is crucial at this moment that KD should be distinguished clearly among the febrile diseases in children. In this study, we aimed to validate a deep learning algorithm for classification of KD and other acute febrile diseases.

METHODS

We obtained coronary artery images by echocardiography of children (n = 138 for KD; n = 65 for pneumonia). We trained six deep learning networks (VGG19, Xception, ResNet50, ResNext50, SE-ResNet50, and SE-ResNext50) using the collected data.

RESULTS

SE-ResNext50 showed the best performance in terms of accuracy, specificity, and precision in the classification. SE-ResNext50 offered a precision of 81.12%, a sensitivity of 84.06%, and a specificity of 58.46%.

CONCLUSIONS

The results of our study suggested that deep learning algorithms have similar performance to an experienced cardiologist in detecting coronary artery lesions to facilitate the diagnosis of KD.

Deep learning kidney segmentation with very limited training data using a cascaded convolution neural network

BACKGROUND

Deep learning segmentation requires large datasets with ground truth. Image annotation is time consuming and leads to shortages of ground truth data for clinical imaging. This study is to investigate the feasibility of kidney segmentation using deep learning convolution neural network (CNN) models trained with MR images from only a few subjects.

METHODS

A total of 60 subjects from two cohorts were included in this study. The first cohort of 20 subjects from publicly available data was used for training and testing. The second cohort of 40 subjects with renal masses from our institution was used for testing only. A few-shot deep learning approach using 3D augmentation was investigated. T1-weighted images in the first cohort were used for training and testing. Cascaded CNN networks were trained using images from one, three, and six subjects, respectively. Images for the remaining subjects were used for testing. Images in the second cohort were utilized for testing only. Dice and Jaccard coefficients were generated to evaluate the performance of CNN models. Statistical analyses for segmentation metrics among different approaches were performed.

RESULTS

Our approach achieved mean Dice coefficients of 0.85 using a single training subject and 0.91 with six training subjects. Compared to a single Unet, the cascaded network significantly improved the results using a single training subject (Dice, 0.759 vs. 0.835; p<0.001) and three subjects (0.864 vs. 0.893; p = 0.015) in the first cohort, and the results for the second cohort (0.821 vs. 0.873; p = 0.008).

CONCLUSION

Our few-shot kidney segmentation approach using 3D augmentation achieved a good performance even using a single Unet. Furthermore, the cascaded network significantly improved the performance of segmentation and was superior to a single Unet in certain cases. Our approach provides a promising solution to segmentation in medical imaging when the number of ground truth masks is limited.