Learning with limited annotations: A survey on deep semi-supervised learning for medical image segmentation

Medical image segmentation is a critical application of computer vision in the analysis of medical images. Its primary objective is to isolate regions of interest in medical images from the background, thereby assisting clinicians in accurately identifying lesions, their sizes, locations, and their relationships with surrounding tissues. However, compared to natural images, medical images present unique challenges, such as low resolution, poor contrast, inconsistency, and scattered target regions. Furthermore, the accuracy and stability of segmentation results are subject to more stringent requirements. In recent years, with the widespread application of Convolutional Neural Networks (CNNs) in computer vision, deep learning-based methods for medical image segmentation have become a focal point of research. This paper categorizes, reviews, and summarizes the current representative methods and research status in the field of medical image segmentation. A comparative analysis of relevant experiments is presented, along with an introduction to commonly used public datasets, performance evaluation metrics, and loss functions in medical image segmentation. Finally, potential future research directions and development trends in this field are predicted and analyzed.

Active learning is considered a viable solution to alleviate the contradiction between the high dependency of deep learning-based segmentation methods on annotated data and the expensive pixel-level annotation cost of medical images. However, most existing methods suffer from unreliable uncertainty assessment and the struggle to balance diversity and informativeness, leading to poor performance in segmentation tasks. In response, we propose an efficient Predictive Accuracy-based Active Learning (PAAL) method for medical image segmentation, first introducing predictive accuracy to define uncertainty. Specifically, PAAL mainly consists of an Accuracy Predictor (AP) and a Weighted Polling Strategy (WPS). The former is an attached learnable module that can accurately predict the segmentation accuracy of unlabeled samples relative to the target model with the predicted posterior probability. The latter provides an efficient hybrid querying scheme by combining predicted accuracy and feature representation, aiming to ensure the uncertainty and diversity of the acquired samples. Extensive experiment results on multiple datasets demonstrate the superiority of PAAL. PAAL achieves comparable accuracy to fully annotated data while reducing annotation costs by approximately 50% to 80%, showcasing significant potential in clinical applications. The code is available at https://github.com/shijun18/PAAL-MedSeg.

Distributed contrastive learning for medical image segmentation.

Medical image segmentation using deep learning typically requires a large quantity of well-annotated data. However, the acquisition of pixel-level annotations is arduous and expensive, often requiring the expertise of experienced medical professionals. Recent advancements in few-shot learning can help address label scarcity in medical image segmentation by leveraging a small amount of labeled data. Most research on prototypical learning for medical image segmentation involves prototypes averaged across the entire class, ignoring fine details. In this work, we propose a novel dual prototype network and introduce part prototypes to supplement local prototypes for extracting fine-grained features and enhancing model performance in one-shot medical image segmentation. Extensive experiments on the CHAOS dataset demonstrate the effectiveness of the proposed method on one-shot segmentation, outperforming other methods by a significant margin.Clinical relevance- The proposed method can address the problem of prototypical self-supervised learning in the one-shot medical image segmentation scenario, saving annotation costs in clinical practice.

Aggregating multi-level feature representation plays a critical role in achieving robust volumetric medical image segmentation, which is important for the auxiliary diagnosis and treatment. Unlike the recent neural architecture search (NAS) methods that typically searched the optimal operators in each network layer, but missed a good strategy to search for feature aggregations, this paper proposes a novel NAS method for 3D medical image segmentation, named UXNet, which searches both the scale-wise feature aggregation strategies as well as the block-wise operators in the encoder-decoder network. UXNet has several appealing benefits. (1) It significantly improves flexibility of the classical UNet architecture, which only aggregates feature representations of encoder and decoder in equivalent resolution. (2) A continuous relaxation of UXNet is carefully designed, enabling its searching scheme performed in an efficient differentiable manner. (3) Extensive experiments demonstrate the effectiveness of UXNet compared with recent NAS methods for medical image segmentation. The architecture discovered by UXNet outperforms existing state-of-the-art models in terms of Dice on several public 3D medical image segmentation benchmarks, especially for the boundary locations and tiny tissues. The searching computational complexity of UXNet is cheap, enabling to search a network with best performance less than 1.5 days on two TitanXP GPUs.

Pixel accurate 2-D, 3-D medical image segmentation to identify abnormalities for further analysis is on high demand for computer-aided medical imaging applications. Various segmentation algorithms have been studied and applied in medical imaging for many years, but the problem remains challenging due to growing a large number of variety of applications starting from lung disease diagnosis based on x-ray images, nucleus detection, and segmentation based on microscopic pictures to kidney tumour segmentation. The recent innovation in deep learning brought revolutionary advances in computer vision. Image segmentation is one such area where deep learning shows its capacity and improves the performance by a larger margin than its successor. This chapter overviews the most popular deep learning-based image segmentation techniques and discusses their capabilities and basic advantages and limitations in the domain of medical imaging.

Semi-supervised learning (SSL) is an effective method for medical image segmentation as it alleviates the dependence on clinical pixel-level annotations. Among the SSL methods, pseudo-labels and consistency regularization play a key role as the dominant paradigm. However, current consistency regularization methods based on shared encoder structures are prone to trap the model in cognitive bias, which impairs the segmentation performance. Furthermore, traditional fixed-threshold-based pseudo-label selection methods lack the utilization of low-confidence pixels, making the model’s initial segmentation capability insufficient, especially for confusing regions. To this end, we propose a multi-perspective dynamic consistency (MPDC) framework to mitigate model cognitive bias and to fully utilize the low-confidence pixels. Specially, we propose a novel multi-perspective collaborative learning strategy that encourages the sub-branch networks to learn discriminative features from multiple perspectives, thus avoiding the problem of model cognitive bias and enhancing boundary perception. In addition, we further employ a dynamic decoupling consistency scheme to fully utilize low-confidence pixels. By dynamically adjusting the threshold, more pseudo-labels are involved in the early stages of training. Extensive experiments on several challenging medical image segmentation datasets show that our method achieves state-of-the-art performance, especially on boundaries, with significant improvements.

MSMCNet: Differential context drives accurate localization and edge smoothing of lesions for medical image segmentation

The purpose of medical image segmentation is to divides the lesions image with a special meaning and background regional. The characteristics of medical images are generally more complex. There is often overlap between different regions, the edge of the area is vague. Research based on fuzzy clustering method for medical image segmentation, fuzzy C-means clustering and fuzzy kernel clustering methods were studied and discussed fully considered the spatial information, the single objective problem conversion into multi-objective problem. The experimental results show that the method use in medical image segmentation achieved good results.

Medical image segmentation is a fundamental task in medical image analysis, playing a crucial role in disease diagnosis, treatment planning, and clinical decision-making. Accurate segmentation of anatomical structures, such as blood vessels, organs, and lesions, is essential for reliable medical interpretations. However, existing segmentation models often face challenges in effectively capturing both local and global features within medical images. Traditional convolutional neural networks (CNNs) excel at extracting fine-grained local features but struggle to model long-range dependencies due to their limited receptive fields. Conversely, transformer-based models can capture global contextual relationships but often fail to preserve detailed local structures, leading to suboptimal segmentation performance. The purpose of this study is to develop SEFormer, a novel hybrid network that integrates SENet, ResNet, and Transformer to enhance medical image segmentation. SEFormer aims to effectively capture both local and global features, addressing the limitations of traditional CNNs and transformer-based models.SENet is incorporated to recalibrate feature maps by adaptively emphasizing informative features, thereby compensating for the global representation limitations of CNNs. Meanwhile, the ResNet backbone ensures deep feature extraction while maintaining computational efficiency. Additionally, the Transformer module is introduced to capture long-range dependencies and enhance contextual awareness, complementing the local feature extraction capabilities of CNNs. This hybrid approach allows SEFormer to balance fine-grained spatial details with broader contextual information, leading to more precise segmentation results.Furthermore, to mitigate information loss during feature extraction and ensure a more complete and hierarchical representation of image information, we introduce a feature pyramid structure inspired by multi-scale image pyramid models. By progressively increasing the receptive field across different scales and employing skip connections, SEFormer effectively fuses multi-scale local and global features at each stage of the pyramid. This hierarchical fusion mechanism ensures a richer and more robust feature representation, which is particularly beneficial for segmenting complex medical structures such as blood vessels. We evaluate SEFormer on the CHASE_DB1 dataset, a widely used benchmark for retinal vessel segmentation. Experimental results demonstrate that SEFormer outperforms existing state-of-the-art segmentation methods, achieving a 3.25% improvement in segmentation accuracy. Additionally, we conduct ablation studies to verify the contributions of different components within SEFormer. The results show that incorporating SENet enhances feature recalibration and channel attention, while the Transformer module significantly improves global context awareness. The feature pyramid structure further contributes to performance gains by ensuring a multi-scale representation of vascular structures. Compared to conventional CNN-based methods, our approach achieves better segmentation accuracy with improved robustness to variations in vessel thickness, noise, and image contrast. Furthermore, SEFormer maintains computational efficiency, making it suitable for real-world clinical applications where both accuracy and processing speed are critical. SEFormer provides an efficient and accurate approach for medical image segmentation by leveraging the complementary strengths of CNNs, SENet, and Transformers while integrating a feature pyramid structure to maximize feature representation. Future research directions include extending SEFormer to other medical imaging modalities, such as CT and MRI scans, as well as exploring further optimization techniques to reduce computational costs while maintaining high segmentation performance. Additionally, integrating SEFormer with active learning frameworks could further improve segmentation performance by leveraging expert-annotated data more efficiently.

3D brain tumor segmentation is essential for the diagnosis, monitoring, and treatment planning of brain diseases. In recent studies, the Deep Convolution Neural Network (DCNN) is one of the most potent methods for medical image segmentation. In this paper, we review the different kinds of tricks applied to 3D brain tumor segmentation with DNN. We divide such tricks into three main categories: data processing methods including data sampling, random patch-size training, and semi-supervised learning, model devising methods including architecture devising and result fusing, and optimizing processes including warming-up learning and multi-task learning. Most of these approaches are not particular to brain tumor segmentation, but applicable to other medical image segmentation problems as well. Evaluated on the BraTS2019 online testing set, we obtain Dice scores of 0.810, 0.883 and 0.861, and Hausdorff Distances (95th percentile) of 2.447, 4.792, and 5.581 for enhanced tumor core, whole tumor, and tumor core, respectively. Our method won the second place of the BraTS 2019 Challenge for the tumor segmentation.

FSA-Net: Rethinking the attention mechanisms in medical image segmentation from releasing global suppressed information

It needs set parameters on image segmentation based on PCNN (Pulse Coupled Neural Network) now. This paper points out the new method for medical image segmentation based on improved PCNN and Tsallis entropy. The new methods can automatically segment the medical images without selecting the PCNN parameters. It gets the best results with combining with the Tsallis entropy. The new method is very useful for PCNN application in the medical images segmentation.

Segmenting medical images via explicit–implicit attention aggregation

Segmentation of medical images is an important step in many clinical and diagnostic imaging applications. Medical images present many challenges for automated segmentation including poor contrast at tissue boundaries. Traditional segmentation methods based solely on information from the image do not work well in such cases. Statistical shape information for objects in medical images are easy to obtain. In this paper, we propose a graph cuts-based segmentation method for medical images that incorporates statistical shape priors to increase robustness. Our proposed method is able to deal with complex shapes and shape variations while taking advantage of the globally efficient optimization by graph cuts. We demonstrate the effectiveness of our method on kidney images without strong boundaries.