DSEU-net: A novel deep supervision SEU-net for medical ultrasound image segmentation

In this paper, a novel level set approach is proposed for segmentation of medical ultrasound images. Considering the speckle noise and low contrast of medical ultrasound images, we add an extra stopping term into the formulation of level set evolution without re-initialization (LSEWR). Compared with traditional active contour models, the proposed level set approach has more flexible initialization and larger capture range. It is insensitive to the initial contour and larger time step can be used. The initial contour can be easily initialized as a circle or rectangular, thus achieving semi-automatic segmentation of ultrasound medical images. The experimental results show that the proposed method can be used for semi-automatic and high-quality segmentation of medical ultrasound images.

ESKNet: An enhanced adaptive selection kernel convolution for ultrasound breast tumors segmentation

Medical ultrasound (US) image segmentation faces significant challenges due to limited labeled data and characteristic imaging artifacts, including speckle noise and low-contrast boundaries. While semi-supervised learning (SSL) approaches have emerged to address data scarcity, existing methods suffer from suboptimal unlabeled data utilization and lack robust feature representation mechanisms. In this article, we propose Switch, a novel SSL framework with two key innovations: 1) a multiscale switch (MSS) strategy that employs hierarchical patch mixing to achieve uniform spatial coverage; and 2) a frequency-domain switch (FDS) with contrastive learning that performs amplitude switching in Fourier space for robust feature representations. Our framework integrates these components within a teacher-student architecture to effectively leverage both labeled and unlabeled data. Comprehensive evaluation across six diverse US datasets (lymph nodes, breast lesions, thyroid nodules, and prostate) demonstrates consistent superiority over state-of-the-art (SOTA) methods. At a 5% labeling ratio, Switch achieves remarkable improvements: 80.04% Dice on LN-INT, 85.52% Dice on DDTI, and 83.48% Dice on Prostate datasets, with our semi-supervised approach even exceeding fully supervised baselines. The method maintains parameter efficiency (1.8 M parameters) while delivering superior performance, validating its effectiveness for resource-constrained medical imaging applications. The source code is publicly available at https://github.com/jinggqu/Switch.

Segmentation of medical ultrasound images (e.g., for the purpose of surgical or radiotherapy planning) is known to be a difficult task due to the relatively low resolution and reduced contrast of the images, as well as due to the discontinuity and uncertainty of segmentation boundaries caused by speckle noise. Under such conditions, useful segmentation results seem to be only achievable by means of relatively complex algorithms, which are usually computationally involved and/or require a prior learning. In this paper, a different approach to the problem of segmentation of medical ultrasound images is proposed. In particular, we propose to preprocess the images before they are subjected to a segmentation procedure. The proposed preprocessing modifies the images (without affecting their anatomic contents) so that the resulting images can be effectively segmented by relatively simple and computationally efficient means. The performance of the proposed method is tested in a series of both in silico and in vivo experiments.

Image segmentation plays an important role in both qualitative and quantitative analysis of medical ultrasound images. But the performance of the classical image-segmentation techniques degrades severely when they are applied to segment medical ultrasound images, for medical ultrasound images have features of poor contrast and strong speckle noise. Firstly, this article investigates and compiles some of the techniques mostly used in the segmentation of medical ultrasound images. Then a bibliographical survey of current research of medical ultrasound images segmentation is given in this paper. Finally, the general tendencies of medical ultrasound images segmentation are presented.

Object segmentation in medical images is an actively investigated research area. Segmentation techniques are a valuable tool in medical diagnostics for cancer tumours and cysts, for planning surgery operations and other medical treatment. In this paper, a Monte Carlo algorithm for extracting lesion contours in ultrasound medical images is proposed. An efficient multiple model particle filter for progressive contour growing (tracking) from a starting point is developed, accounting for convex, non-circular forms of delineated contour areas. The driving idea of the proposed particle filter consists in the incorporation of different image intensity inside and outside the contour into the filter likelihood function. The filter employs image intensity gradients as measurements and requires information about four manually selected points: a seed point, a starting point, arbitrarily selected on the contour, and two additional points, bounding the measurement formation area around the contour. The filter performance is studied by segmenting contours from a number of real and simulated ultrasound medical images. Accurate contour segmentation is achieved with the proposed approach in ultrasound images with a high level of speckle noise.

Ultrasound imaging can distinctly display the morphology and structure of internal organs within the human body, enabling the examination of organs like the breast, liver, and thyroid. It can identify the locations of tumors, nodules, and other lesions, thereby serving as an efficacious tool for treatment detection and rehabilitation evaluation. Typically, the attending physician is required to manually demarcate the boundaries of lesion locations, such as tumors, in ultrasound images. Nevertheless, several issues exist. The high noise level in ultrasound images, the degradation of image quality due to the impact of surrounding tissues, and the influence of the operator’s experience and proficiency on the determination of lesion locations can all contribute to a reduction in the accuracy of delineating the boundaries of lesion sites. In the wake of the advancement of deep learning, its application in medical image segmentation is becoming increasingly prevalent. For instance, while the U-Net model has demonstrated a favorable performance in medical image segmentation, the convolution layers of the traditional U-Net model are relatively simplistic, leading to suboptimal extraction of global information. Moreover, due to the significant noise present in ultrasound images, the model is prone to interference. In this research, we propose an Attention Residual Network model (ARU-Net). By incorporating residual connections within the encoder section, the learning capacity of the model is enhanced. Additionally, a spatial hybrid convolution module is integrated to augment the model’s ability to extract global information and deepen the vertical architecture of the network. During the feature fusion stage of the skip connections, a channel attention mechanism and a multi-convolutional self-attention mechanism are respectively introduced to suppress noisy points within the fused feature maps, enabling the model to acquire more information regarding the target region. Finally, the predictive efficacy of the model was evaluated using publicly accessible breast ultrasound and thyroid ultrasound data. The ARU-Net achieved mean Intersection over Union (mIoU) values of 82.59% and 84.88%, accuracy values of 97.53% and 96.09%, and F1-score values of 90.06% and 89.7% for breast and thyroid ultrasound, respectively.

Deep learning has achieved remarkable success in medical image segmentation, particularly in ultrasound imaging, where deep neural networks have demonstrated near-expert performance. However, these models typically assume that training and test data follow the same distribution-an assumption that often fails in real-world clinical practice due to domain shifts caused by variations in imaging devices, acquisition protocols, and operator techniques. These discrepancies can significantly degrade model performance. Existing solutions-such as supervised fine-tuning, unsupervised domain adaptation, and domain generalization-require either costly labeled data or access to source domain data, limiting their scalability and clinical applicability. To address domain shift in real-world ultrasound image segmentation, this study proposes a test-time adaptation (TTA) framework that eliminates the need for source data or target labels, while ensuring robustness against distributional drift and catastrophic forgetting. We present Prototype Bank-Driven Test-Time Adaptation (PBTTA), a novel TTA framework consisting of two key modules: (1) the Dynamic Statistics Fusion Module (DSFM), which enables domain-level adaptation by dynamically adjusting batch normalization layers using fused statistics from the test sample and source domain; and (2) the Prototype Bank-Guided Semantic Adaptation Module (PBSAM), which maintains a dynamic prototype bank for each semantic class to support semantic-level adaptation. PBTTA employs a dual-classifier strategy that combines a prototype-based classifier for stable semantic guidance and a parameter-based classifier for flexible decision-making. Notably, PBTTA does not require backpropagation to update model parameters during test time adaptation phase, ensuring efficient and stable adaptation. PBTTA achieves state-of-the-art performance across both ultrasound breast and thyroid tumor segmentation tasks. On average, it improves the Dice score by 15.04% (to 64.82%) for breast tumor segmentation and by 8.88% (to 57.45%) for thyroid tumor segmentation, compared to non-adaptive baselines. Moreover, PBTTA exhibits excellent robustness under continuous domain shifts and effectively mitigates catastrophic forgetting.

This paper presents a new extended Chan-Vese level set method for ultrasound image segmentation. The proposed method introduces wavelet multiresolution analysis to create an edge representing mask to spot edge information of the image. While the evolution of the level set function, average edge energy of zero level set curve is calculated from the edge representing mask to control the evolving speed. The proposed method has some advantages compared with traditional Chan-Vese method as follows. First, the proposed method is robust to the inherent speckle noise in ultrasound images. Second, the method is sensitive to intensity inhomogeneity of objects in images. Third, the method can solve the segmentation of ultrasound images with weak or missing boundaries. We apply the proposed method to synthetic and medical ultrasound images, and the results suggest that this method is superior to the traditional Chan-Vese level set method.

The segmentation of ultrasound image for breast cancer is an important task in the field of biomedical research. The traditional U-Net model, with its simple structure and remarkable performance, this approach has found extensive application in the segmentation of medical images. However, U-Net tends to be affected by background noise when handling images with complex backgrounds or blurry boundaries, which may impact the segmentation accuracy. To address this issue, the Attention U-Net model incorporates an attention mechanism, enabling the model to selectively focus on critical target areas within the image, thereby improving segmentation accuracy. This paper further optimizes the Attention U-Net architecture by increasing the depth of both the encoder and decoder sections, enhancing the model's capacity for feature extraction and image reconstruction. Consequently, both the accuracy and robustness of segmentation are enhanced. The experimental findings indicate that the proposed modified Attention U-Net model significantly outperforms traditional methods in breast ultrasound image segmentation tasks. It effectively handles various types of breast images, particularly those with complex backgrounds, blurred targets, or small sizes, maintaining high segmentation accuracy. This study offers an effective solution for the automated segmentation of breast ultrasound images, with substantial implications for enhancing both the automation and diagnostic efficiency in medical image analysis.

Image segmentation is one of the earliest and most important stages of image processing and plays an important role in both qualitative and quantitative analysis of medical ultrasound images but ultrasound images have low level of contrast and are corrupted with strong speckle noise. Due to these effects, segmentation of ultrasound images is very challenging and traditional image segmentation methods may not be leads to satisfactory results. The active contour method has been one of the widely used techniques for image segmentation; however, due to low quality of ultrasound images, it has encountered difficulties. In this paper, we presented a segmental method combined genetic algorithm and active contour with an energy minimization procedure based on genetic algorithms. This method have been proposed to overcome some limits of classical active contours, as con-tour initialization and local minima (speckle noise), and have been successfully applied on medical ultrasound images. Experimental result on medical ultrasound image show that our presented method only can correctly segment the circular tissue’s on ultra-sound images.

Though a popular imaging technique, ultrasound is known for producing images filled with noise, distortions and shadowing effects. As a result, segmentation of ultrasound images require significant prior knowledge, often inserted into algorithms interactively or through shape information of the region of interest. This type of prior knowledge puts limitations on current approaches. This paper presents a different approach to ultrasound image segmentation that relies mainly on the physical properties of ultrasonic imaging. Robust intensity-based external energy formulations are incorporated into an active contour framework that is tolerant of the noise common to ultrasound images. By initializing the contour through an ellipse fitting procedure, an autonomous ultrasound image segmentation system is created that that can generalize to objects of varying shapes and sizes. The segmentation system was tested on ultrasound images of neonatal kidneys with results comparable to current methods

A review of medical ocular image segmentation

Segmentation of medical ultrasound images is one of the most important functional components of medical ultrasonic instruments for computed aided diagnosis, such as breast lesion early detection and measurement. However, the segmentation of breast lesions from ultrasound images is still a challenging task due to the variance in shape of the lesions and interference from speckle noise. In this paper, a novel approach using tight frames of grouping bandlet(TFGB) and improved stochastic neighbor embedding algorithm(ISNE) is adopted for the segmentation of breast lesions. Novelty in this paper includes that grouplet transform is successively applied and adapted for image segmentation for the first time whereas all of previous works were focused on image inpainting, texture synthesis and image denoising. Moreover, an effective stopping criterion is proposed to improve the dimension reduction technique-stochastic neighbor embedding (SNE) in this paper. Applications to clinical ultrasound images with fibroadenoma and fibrocystic breast lesions contaminated by speckle noise are performed, respectively. Experimental results show that compared with the state-of-the-art approach to the segmentation of breast lesions in ultrasound images, the proposed approach demonstrates robustness to speckle noise and superior performance on the effectiveness.

A hybrid enhanced attention transformer network for medical ultrasound image segmentation