Source-free unsupervised domain adaptation for cross-modality abdominal multi-organ segmentation

Beyond pixel: Superpixel-based MRI segmentation through traditional machine learning and graph convolutional network

BACKGROUND AND OBJECTIVE

Tendon segmentation is crucial for studying tendon-related pathologies like tendinopathy, tendinosis, etc. This step further enables detailed analysis of specific tendon regions using automated or semi-automated methods. This study specifically aims at the segmentation of Achilles tendon, the largest tendon in the human body.

METHODS

This study proposes a comprehensive end-to-end tendon segmentation module composed of a preliminary superpixel-based coarse segmentation preceding the final segmentation task. The final segmentation results are obtained through two distinct approaches. In the first approach, the coarsely generated superpixels are subjected to classification using Random Forest (RF) and Support Vector Machine (SVM) classifiers to classify whether each superpixel belongs to a tendon class or not (resulting in tendon segmentation). In the second approach, the arrangements of superpixels are converted to graphs instead of being treated as conventional image grids. This classification process uses a graph-based convolutional network (GCN) to determine whether each superpixel corresponds to a tendon class or not.

RESULTS

All experiments are conducted on a custom-made ankle MRI dataset. The dataset comprises 76 subjects and is divided into two sets: one for training (Dataset 1, trained and evaluated using leave-one-group-out cross-validation) and the other as unseen test data (Dataset 2). Using our first approach, the final test AUC (Area Under the ROC Curve) scores using RF and SVM classifiers on the test data (Dataset 2) are 0.992 and 0.987, respectively, with sensitivities of 0.904 and 0.966. On the other hand, using our second approach (GCN-based node classification), the AUC score for the test set is 0.933 with a sensitivity of 0.899.

CONCLUSIONS

Our proposed pipeline demonstrates the efficacy of employing superpixel generation as a coarse segmentation technique for the final tendon segmentation. Whether utilizing RF, SVM-based superpixel classification, or GCN-based classification for tendon segmentation, our system consistently achieves commendable AUC scores, especially the non-graph-based approach. Given the limited dataset, our graph-based method did not perform as well as non-graph-based superpixel classifications; however, the results obtained provide valuable insights into how well the models can distinguish between tendons and non-tendons. This opens up opportunities for further exploration and improvement.

Hybrid-feature based spherical quasi-conformal registration for AD-induced hippocampal surface morphological changes

BACKGROUND AND OBJECTIVE

Establishing accurate one-to-one morphological correspondence between different hippocampal surfaces is a solid foundation for the analysis of AD-induced hippocampal morphological changes. However, owing to the large variations between hippocampal surfaces, exiting registration work either fails to obtain the accurate matching of local and overall morphological features or does not preserve the bijectivity during parametric mapping. For this reason, this study proposes a hybrid-feature based spherical quasi-conformal registration (HSQR) method that can effectively maintain the diffeomorphic property while meeting the hybrid-feature matching constraints in the spherical parameter domain.

METHODS

The HSQR algorithm is primarily achieved through hippocampal surface hybrid feature extraction and spherical quasi-conformal registration. First, hybrid features for a comprehensive morphological description of the hippocampal surface were established, which included essential anatomical features (landmarks) and mean curvature (intensity) features to ensure the accuracy of surface morphology alignment. Second, spherical parameterization was applied to genus-0 closed surfaces, such as the hippocampus, which maximized the preservation of the original local surface morphology through area-preserving properties. Third, a novel spherical quasi-conformal registration algorithm that can handle large deformations is established. It transforms a 3D spherical parameter domain into a 2D plane parameter domain using iterative local stereo projection to improve the efficiency of the registration algorithm. Subsequently, by controlling the Beltramin coefficient, the hybrid morphological features could be aligned while ensuring bijection before and after registration.

RESULTS

Using a cohort including 161 patients with amyloid-β (Aβ) positive Alzheimer disease (AD), 234 Aβ positive mild cognitive impairment (MCI) and 266 Aβ negative cognitively unimpaired (CU) individuals from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, we set up the experiment which indicated that the HSQR-based whole bilateral hippocampal atrophy features demonstrated the stronger statistical power for group morphological differences of CU vs. MCI with q-value: 0.0453 for left hippocampus and 0.0401 for right hippocampus and group morphological differences of AD vs. MCI with q-value: 0.0282 for left hippocampus and 0.0421 for right hippocampus.

CONCLUSIONS

Our registration algorithm may provide a solid foundation for the accurate quantification of hippocampal surface morphological changes for the differential diagnosis and tracking of AD.

A Comprehensive Survey on Source-Free Domain Adaptation

Over the past decade, domain adaptation has become a widely studied branch of transfer learning which aims to improve performance on target domains by leveraging knowledge from the source domain. Conventional domain adaptation methods often assume access to both source and target domain data simultaneously, which may not be feasible in real-world scenarios due to privacy and confidentiality concerns. As a result, the research of Source-Free Domain Adaptation (SFDA) has drawn growing attention in recent years, which only utilizes the source-trained model and unlabeled target data to adapt to the target domain. Despite the rapid explosion of SFDA work, there has been no timely and comprehensive survey in the field. To fill this gap, we provide a comprehensive survey of recent advances in SFDA and organize them into a unified categorization scheme based on the framework of transfer learning. Instead of presenting each approach independently, we modularize several components of each method to more clearly illustrate their relationships and mechanisms in light of the composite properties of each method. Furthermore, we compare the results of more than 30 representative SFDA methods on three popular classification benchmarks, namely Office-31, Office-home, and VisDA, to explore the effectiveness of various technical routes and the combination effects among them. Additionally, we briefly introduce the applications of SFDA and related fields. Drawing on our analysis of the challenges confronting SFDA, we offer some insights into future research directions and potential settings.

Towards more precise automatic analysis: a systematic review of deep learning-based multi-organ segmentation

Accurate segmentation of multiple organs in the head, neck, chest, and abdomen from medical images is an essential step in computer-aided diagnosis, surgical navigation, and radiation therapy. In the past few years, with a data-driven feature extraction approach and end-to-end training, automatic deep learning-based multi-organ segmentation methods have far outperformed traditional methods and become a new research topic. This review systematically summarizes the latest research in this field. We searched Google Scholar for papers published from January 1, 2016 to December 31, 2023, using keywords "multi-organ segmentation" and "deep learning", resulting in 327 papers. We followed the PRISMA guidelines for paper selection, and 195 studies were deemed to be within the scope of this review. We summarized the two main aspects involved in multi-organ segmentation: datasets and methods. Regarding datasets, we provided an overview of existing public datasets and conducted an in-depth analysis. Concerning methods, we categorized existing approaches into three major classes: fully supervised, weakly supervised and semi-supervised, based on whether they require complete label information. We summarized the achievements of these methods in terms of segmentation accuracy. In the discussion and conclusion section, we outlined and summarized the current trends in multi-organ segmentation.

FedDUS: Lung tumor segmentation on CT images through federated semi-supervised with dynamic update strategy

BACKGROUND AND OBJECTIVE

Lung tumor annotation is a key upstream task for further diagnosis and prognosis. Although deep learning techniques have promoted automation of lung tumor segmentation, there remain challenges impeding its application in clinical practice, such as a lack of prior annotation for model training and data-sharing among centers.

METHODS

In this paper, we use data from six centers to design a novel federated semi-supervised learning (FSSL) framework with dynamic model aggregation and improve segmentation performance for lung tumors. To be specific, we propose a dynamically updated algorithm to deal with model parameter aggregation in FSSL, which takes advantage of both the quality and quantity of client data. Moreover, to increase the accessibility of data in the federated learning (FL) network, we explore the FAIR data principle while the previous federated methods never involve.

RESULT

The experimental results show that the segmentation performance of our model in six centers is 0.9348, 0.8436, 0.8328, 0.7776, 0.8870 and 0.8460 respectively, which is superior to traditional deep learning methods and recent federated semi-supervised learning methods.

CONCLUSION

The experimental results demonstrate that our method is superior to the existing FSSL methods. In addition, our proposed dynamic update strategy effectively utilizes the quality and quantity information of client data and shows efficiency in lung tumor segmentation. The source code is released on (https://github.com/GDPHMediaLab/FedDUS).

An efficient Fusion-Purification Network for Cervical pap-smear image classification

BACKGROUND AND OBJECTIVES

In cervical cell diagnostics, autonomous screening technology constitutes the foundation of automated diagnostic systems. Currently, numerous deep learning-based classification techniques have been successfully implemented in the analysis of cervical cell images, yielding favorable outcomes. Nevertheless, efficient discrimination of cervical cells continues to be challenging due to large intra-class and small inter-class variations. The key to dealing with this problem is to capture localized informative differences from cervical cell images and to represent discriminative features efficiently. Existing methods neglect the importance of global morphological information, resulting in inadequate feature representation capability.

METHODS

To address this limitation, we propose a novel cervical cell classification model that focuses on purified fusion information. Specifically, we first integrate the detailed texture information and morphological structure features, named cervical pathology information fusion. Second, in order to enhance the discrimination of cervical cell features and address the data redundancy and bias inherent after fusion, we design a cervical purification bottleneck module. This model strikes a balance between leveraging purified features and facilitating high-efficiency discrimination. Furthermore, we intend to unveil a more intricate cervical cell dataset: Cervical Cytopathology Image Dataset (CCID).

RESULTS

Extensive experiments on two real-world datasets show that our proposed model outperforms state-of-the-art cervical cell classification models.

CONCLUSIONS

The results show that our method can well help pathologists to accurately evaluate cervical smears.

Source-free unsupervised domain adaptation: A survey

Unsupervised domain adaptation (UDA) via deep learning has attracted appealing attention for tackling domain-shift problems caused by distribution discrepancy across different domains. Existing UDA approaches highly depend on the accessibility of source domain data, which is usually limited in practical scenarios due to privacy protection, data storage and transmission cost, and computation burden. To tackle this issue, many source-free unsupervised domain adaptation (SFUDA) methods have been proposed recently, which perform knowledge transfer from a pre-trained source model to the unlabeled target domain with source data inaccessible. A comprehensive review of these works on SFUDA is of great significance. In this paper, we provide a timely and systematic literature review of existing SFUDA approaches from a technical perspective. Specifically, we categorize current SFUDA studies into two groups, i.e., white-box SFUDA and black-box SFUDA, and further divide them into finer subcategories based on different learning strategies they use. We also investigate the challenges of methods in each subcategory, discuss the advantages/disadvantages of white-box and black-box SFUDA methods, conclude the commonly used benchmark datasets, and summarize the popular techniques for improved generalizability of models learned without using source data. We finally discuss several promising future directions in this field.

Dual-space high-frequency learning for transformer-based MRI super-resolution

BACKGROUND AND OBJECTIVE

Magnetic resonance imaging (MRI) can provide rich and detailed high-contrast information of soft tissues, while the scanning of MRI is time-consuming. To accelerate MR imaging, a variety of Transformer-based single image super-resolution methods are proposed in recent years, achieving promising results thanks to their superior capability of capturing long-range dependencies. Nevertheless, most existing works prioritize the design of transformer attention blocks to capture global information. The local high-frequency details, which are pivotal to faithful MRI restoration, are unfortunately neglected.

METHODS

In this work, we propose a high-frequency enhanced learning scheme to effectively improve the awareness of high frequency information in current Transformer-based MRI single image super-resolution methods. Specifically, we present two entirely plug-and-play modules designed to equip Transformer-based networks with the ability to recover high-frequency details from dual spaces: 1) in the feature space, we design a high-frequency block (Hi-Fe block) paralleled with Transformer-based attention layers to extract rich high-frequency features; while 2) in the image intensity space, we tailor a high-frequency amplification module (HFA) to further refine the high-frequency details. By fully exploiting the merits of the two modules, our framework can recover abundant and diverse high-frequency information, rendering faithful MRI super-resolved results with fine details.

RESULTS

We integrated our modules with six Transformer-based models and conducted experiments across three datasets. The results indicate that our plug-and-play modules can enhance the super-resolution performance of all foundational models to varying degrees, surpassing the capabilities of existing state-of-the-art single image super-resolution networks.

CONCLUSION

Comprehensive comparison of super-resolution images and high-frequency maps from various methods, clearly demonstrating that our module possesses the capability to restore high-frequency information, showing huge potential in clinical practice for accelerated MRI reconstruction.

Multi-organ segmentation: a progressive exploration of learning paradigms under scarce annotation

Precise delineation of multiple organs or abnormal regions in the human body from medical images plays an essential role in computer-aided diagnosis, surgical simulation, image-guided interventions, and especially in radiotherapy treatment planning. Thus, it is of great significance to explore automatic segmentation approaches, among which deep learning-based approaches have evolved rapidly and witnessed remarkable progress in multi-organ segmentation. However, obtaining an appropriately sized and fine-grained annotated dataset of multiple organs is extremely hard and expensive. Such scarce annotation limits the development of high-performance multi-organ segmentation models but promotes many annotation-efficient learning paradigms. Among these, studies on transfer learning leveraging external datasets, semi-supervised learning including unannotated datasets and partially-supervised learning integrating partially-labeled datasets have led the dominant way to break such dilemmas in multi-organ segmentation. We first review the fully supervised method, then present a comprehensive and systematic elaboration of the 3 abovementioned learning paradigms in the context of multi-organ segmentation from both technical and methodological perspectives, and finally summarize their challenges and future trends.

U-shaped convolutional transformer GAN with multi-resolution consistency loss for restoring brain functional time-series and dementia diagnosis

Introduction

The blood oxygen level-dependent (BOLD) signal derived from functional neuroimaging is commonly used in brain network analysis and dementia diagnosis. Missing the BOLD signal may lead to bad performance and misinterpretation of findings when analyzing neurological disease. Few studies have focused on the restoration of brain functional time-series data.

Methods

In this paper, a novel U-shaped convolutional transformer GAN (UCT-GAN) model is proposed to restore the missing brain functional time-series data. The proposed model leverages the power of generative adversarial networks (GANs) while incorporating a U-shaped architecture to effectively capture hierarchical features in the restoration process. Besides, the multi-level temporal-correlated attention and the convolutional sampling in the transformer-based generator are devised to capture the global and local temporal features for the missing time series and associate their long-range relationship with the other brain regions. Furthermore, by introducing multi-resolution consistency loss, the proposed model can promote the learning of diverse temporal patterns and maintain consistency across different temporal resolutions, thus effectively restoring complex brain functional dynamics.

Results

We theoretically tested our model on the public Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, and our experiments demonstrate that the proposed model outperforms existing methods in terms of both quantitative metrics and qualitative assessments. The model's ability to preserve the underlying topological structure of the brain functional networks during restoration is a particularly notable achievement.

Conclusion

Overall, the proposed model offers a promising solution for restoring brain functional time-series and contributes to the advancement of neuroscience research by providing enhanced tools for disease analysis and interpretation.

Deep learning for unsupervised domain adaptation in medical imaging: Recent advancements and future perspectives

Deep learning has demonstrated remarkable performance across various tasks in medical imaging. However, these approaches primarily focus on supervised learning, assuming that the training and testing data are drawn from the same distribution. Unfortunately, this assumption may not always hold true in practice. To address these issues, unsupervised domain adaptation (UDA) techniques have been developed to transfer knowledge from a labeled domain to a related but unlabeled domain. In recent years, significant advancements have been made in UDA, resulting in a wide range of methodologies, including feature alignment, image translation, self-supervision, and disentangled representation methods, among others. In this paper, we provide a comprehensive literature review of recent deep UDA approaches in medical imaging from a technical perspective. Specifically, we categorize current UDA research in medical imaging into six groups and further divide them into finer subcategories based on the different tasks they perform. We also discuss the respective datasets used in the studies to assess the divergence between the different domains. Finally, we discuss emerging areas and provide insights and discussions on future research directions to conclude this survey.

Deep Learning Framework for Liver Segmentation from T1-Weighted MRI Images

The human liver exhibits variable characteristics and anatomical information, which is often ambiguous in radiological images. Machine learning can be of great assistance in automatically segmenting the liver in radiological images, which can be further processed for computer-aided diagnosis. Magnetic resonance imaging (MRI) is preferred by clinicians for liver pathology diagnosis over volumetric abdominal computerized tomography (CT) scans, due to their superior representation of soft tissues. The convenience of Hounsfield unit (HoU) based preprocessing in CT scans is not available in MRI, making automatic segmentation challenging for MR images. This study investigates multiple state-of-the-art segmentation networks for liver segmentation from volumetric MRI images. Here, T1-weighted (in-phase) scans are investigated using expert-labeled liver masks from a public dataset of 20 patients (647 MR slices) from the Combined Healthy Abdominal Organ Segmentation grant challenge (CHAOS). The reason for using T1-weighted images is that it demonstrates brighter fat content, thus providing enhanced images for the segmentation task. Twenty-four different state-of-the-art segmentation networks with varying depths of dense, residual, and inception encoder and decoder backbones were investigated for the task. A novel cascaded network is proposed to segment axial liver slices. The proposed framework outperforms existing approaches reported in the literature for the liver segmentation task (on the same test set) with a dice similarity coefficient (DSC) score and intersect over union (IoU) of 95.15% and 92.10%, respectively.

Tailored multi-organ segmentation with model adaptation and ensemble

Multi-organ segmentation, which identifies and separates different organs in medical images, is a fundamental task in medical image analysis. Recently, the immense success of deep learning motivated its wide adoption in multi-organ segmentation tasks. However, due to expensive labor costs and expertise, the availability of multi-organ annotations is usually limited and hence poses a challenge in obtaining sufficient training data for deep learning-based methods. In this paper, we aim to address this issue by combining off-the-shelf single-organ segmentation models to develop a multi-organ segmentation model on the target dataset, which helps get rid of the dependence on annotated data for multi-organ segmentation. To this end, we propose a novel dual-stage method that consists of a Model Adaptation stage and a Model Ensemble stage. The first stage enhances the generalization of each off-the-shelf segmentation model on the target domain, while the second stage distills and integrates knowledge from multiple adapted single-organ segmentation models. Extensive experiments on four abdomen datasets demonstrate that our proposed method can effectively leverage off-the-shelf single-organ segmentation models to obtain a tailored model for multi-organ segmentation with high accuracy.

Detection and Classification of Melanoma Skin Cancer Using Image Processing Technique

Human skin cancer is the most common and potentially life-threatening form of cancer. Melanoma skin cancer, in particular, exhibits a high mortality rate. Early detection is crucial for effective treatment. Traditionally, melanoma is detected through painful and time-consuming biopsies. This research introduces a computer-aided detection technique for early melanoma diagnosis-sis. In this study, we propose two methods for detecting skin cancer and focus specifically on melanoma cancerous cells using image data. The first method employs convolutional neural networks, including AlexNet, LeNet, and VGG-16 models, and we integrate the model with the highest accuracy into web and mobile applications. We also investigate the relationship between model depth and performance with varying dataset sizes. The second method uses support vector machines with a default RBF kernel, using feature parameters to categorize images as benign, malignant, or normal after image processing. The SVM classifier achieved an 86.6% classification accuracy, while the CNN maintained a 91% accuracy rate after 100 compute epochs. The CNN model is deployed as a web and mobile application with the assistance of Django and Android Studio.

EDRL: Entropy-guided disentangled representation learning for unsupervised domain adaptation in semantic segmentation

BACKGROUND AND OBJECTIVE

Deep learning-based approaches are excellent at learning from large amounts of data, but can be poor at generalizing the learned knowledge to testing datasets with domain shift, i.e., when there exists distribution discrepancy between the training dataset (source domain) and the testing dataset (target domain). In this paper, we investigate unsupervised domain adaptation (UDA) techniques to train a cross-domain segmentation method which is robust to domain shift, eliminating the requirement of any annotations on the target domain.

METHODS

To this end, we propose an Entropy-guided Disentangled Representation Learning, referred as EDRL, for UDA in semantic segmentation. Concretely, we synergistically integrate image alignment via disentangled representation learning with feature alignment via entropy-based adversarial learning into one network, which is trained end-to-end. We additionally introduce a dynamic feature selection mechanism via soft gating, which helps to further enhance the task-specific feature alignment. We validate the proposed method on two publicly available datasets: the CT-MR dataset and the multi-sequence cardiac MR (MS-CMR) dataset.

RESULTS

On both datasets, our method achieved better results than the state-of-the-art (SOTA) methods. Specifically, on the CT-MR dataset, our method achieved an average DSC of 84.8% when taking CT as the source domain and MR as the target domain, and an average DSC of 84.0% when taking MR as the source domain and CT as the target domain.

CONCLUSIONS

Results from comprehensive experiments demonstrate the efficacy of the proposed EDRL model for cross-domain medical image segmentation.

Constructing brain functional network by Adversarial Temporal-Spatial Aligned Transformer for early AD analysis

Introduction

The brain functional network can describe the spontaneous activity of nerve cells and reveal the subtle abnormal changes associated with brain disease. It has been widely used for analyzing early Alzheimer's disease (AD) and exploring pathological mechanisms. However, the current methods of constructing functional connectivity networks from functional magnetic resonance imaging (fMRI) heavily depend on the software toolboxes, which may lead to errors in connection strength estimation and bad performance in disease analysis because of many subjective settings.

Methods

To solve this problem, in this paper, a novel Adversarial Temporal-Spatial Aligned Transformer (ATAT) model is proposed to automatically map 4D fMRI into functional connectivity network for early AD analysis. By incorporating the volume and location of anatomical brain regions, the region-guided feature learning network can roughly focus on local features for each brain region. Also, the spatial-temporal aligned transformer network is developed to adaptively adjust boundary features of adjacent regions and capture global functional connectivity patterns of distant regions. Furthermore, a multi-channel temporal discriminator is devised to distinguish the joint distributions of the multi-region time series from the generator and the real sample.

Results

Experimental results on the Alzheimer's Disease Neuroimaging Initiative (ADNI) proved the effectiveness and superior performance of the proposed model in early AD prediction and progression analysis.

Discussion

To verify the reliability of the proposed model, the detected important ROIs are compared with clinical studies and show partial consistency. Furthermore, the most significant altered connectivity reflects the main characteristics associated with AD.

Conclusion

Generally, the proposed ATAT provides a new perspective in constructing functional connectivity networks and is able to evaluate the disease-related changing characteristics at different stages for neuroscience exploration and clinical disease analysis.

Cervical Cell Image Classification-Based Knowledge Distillation

Current deep-learning-based cervical cell classification methods suffer from parameter redundancy and poor model generalization performance, which creates challenges for the intelligent classification of cervical cytology smear images. In this paper, we establish a method for such classification that combines transfer learning and knowledge distillation. This new method not only transfers common features between different source domain data, but also realizes model-to-model knowledge transfer using the unnormalized probability output between models as knowledge. A multi-exit classification network is then introduced as the student network, where a global context module is embedded in each exit branch. A self-distillation method is then proposed to fuse contextual information; deep classifiers in the student network guide shallow classifiers to learn, and multiple classifier outputs are fused using an average integration strategy to form a classifier with strong generalization performance. The experimental results show that the developed method achieves good results using the SIPaKMeD dataset. The accuracy, sensitivity, specificity, and F-measure of the five classifications are 98.52%, 98.53%, 98.68%, 98.59%, respectively. The effectiveness of the method is further verified on a natural image dataset.

Application of Deep Learning to IVC Filter Detection from CT Scans

IVC filters (IVCF) perform an important function in select patients that have venous blood clots. However, they are usually intended to be temporary, and significant delay in removal can have negative health consequences for the patient. Currently, all Interventional Radiology (IR) practices are tasked with tracking patients in whom IVCF are placed. Due to their small size and location deep within the abdomen it is common for patients to forget that they have an IVCF. Therefore, there is a significant delay for a new healthcare provider to become aware of the presence of a filter. Patients may have an abdominopelvic CT scan for many reasons and, fortunately, IVCF are clearly visible on these scans. In this research a deep learning model capable of segmenting IVCF from CT scan slices along the axial plane is developed. The model achieved a Dice score of 0.82 for training over 372 CT scan slices. The segmentation model is then integrated with a prediction algorithm capable of flagging an entire CT scan as having IVCF. The prediction algorithm utilizing the segmentation model achieved a 92.22% accuracy at detecting IVCF in the scans.