With the growing prominence of digital dentistry, high-quality and cost-effective three-dimensional (3D) intra-oral scanned (IOS) tooth data has become essential in various dental applications. However, limited by sensor resolution and occlusion, the acquired tooth point clouds often suffer from sparsity and incompleteness, resulting in missing regions and insufficient 3D detail. In this paper, we propose a Hierarchical Tooth Completion Network(HTCNet), a novel image-assistance framework to integrate geometry and structure learning from 2D images and 3D point clouds for 3D tooth completion. It employs a dual-stream-based hierarchical feature extraction, utilizing a 2D stream for extracting image features and a 3D stream for processing point clouds. Additionally, we introduce 3DTeethSegX for evaluating image-assistance tooth completion to address the lack of consideration for the comprehensive situation in the current dataset. Extensive experiments on two tooth completion datasets demonstrate the superiority and robustness of HTCNet, showcasing its potential to generate high-resolution 3D tooth data from low-resolution, cost-effective sensors with the assistance of 2D images. To the best of our knowledge, it is the first to use image-assistance techniques to significantly improve the accuracy and effectiveness of 3D tooth completion. Source code will be available at: https://github.com/labiip/HTCNet.

Federated Learning (FL) emerged as a solution for collaborative medical image classification while preserving data privacy. However, label noise, which arises from inter-institutional data variability, can cause training instability and degrade model performance. Existing FL methods struggle with noise heterogeneity and the imbalance in medical data. Motivated by these challenges, we propose FedGSCA, a novel framework for enhancing robustness in noisy medical FL. FedGSCA introduces a Global Sample Selector that aggregates noise knowledge from all clients, effectively addressing noise heterogeneity and improving global model stability. Furthermore, we develop a Client Adaptive Adjustment (CAA) mechanism that combines adaptive threshold pseudo-label generation and Robust Credal Labeling Loss. CAA dynamically adjusts to class distributions, ensuring the inclusion of minority samples and carefully managing noisy labels by considering multiple plausible labels. This dual approach mitigates the impact of noisy data and prevents overfitting during local training, which improves the generalizability of the model. We evaluate FedGSCA on one real-world colon slides dataset and two synthetic medical datasets under various noise conditions, including symmetric, asymmetric, extreme, and heterogeneous types. The results show that FedGSCA outperforms the state-of-the-art methods, excelling in extreme and heterogeneous noise scenarios. Moreover, FedGSCA demonstrates significant advantages in improving model stability and handling complex noise, making it well-suited for real-world medical federated learning scenarios.

Retinal image processing is critical for the early diagnosis and management of ocular diseases. While foundation models pretrained on large-scale datasets show strong generalization, they struggle with rare diseases absent during pre-training. Fine-tuning on datasets containing these rare diseases introduces challenges like class imbalance and domain shift. To address these, we propose adaptive imbalance-to-generalization learning (AIGL), a novel pipeline that integrates imbalance-aware training and test-time feature reconstruction to enhance model performance. During training, we introduce the confidence-weighted adaptive loss, which dynamically adjusts the contribution of each sample based on the model’s confidence, addressing severe class imbalance and improving rare diseases recognition. At test-time, we design context-aware feature masked autoencoders, which operate on high-level features since structures like the macula remain consistent across raw retinal images. It improves generalization by leveraging dynamic token filtering and consistency constraints to adaptively mask and reconstruct features through self-supervised learning. Our AIGL outperforms state-of-the-art methods across multiple metrics on the custom RetinaX dataset, demonstrating superior classification ability.

ABSTRACT Convolutional neural network (CNN)‐based auxiliary diagnostic systems have been widely proposed. However, CNNs have limitations in perceiving global features and more subtle features, which makes existing methods unable to achieve ideal accuracy in tasks such as pneumoconiosis screening. To overcome these limitations, we propose MBSM‐Net, a new multi‐branch structure‐enhanced model for pneumoconiosis screening and grading based on X‐ray images. MBSM‐Net introduces an adaptive feature selection and fusion module to achieve synchronous extraction and hierarchical fusion of global and local features. In the local feature extraction module, we designed a CNN‐Mamba module. This module integrates prior information through a detailed enhancement module to compensate for the shortcomings of traditional convolutions and significantly enhances the expression of subtle lesion information. Meanwhile, the Mamba module simulates pixel‐level long‐range dependencies to extract finer‐grained texture features. In the global feature extraction module, we cleverly utilize the windowed multi‐head self‐attention (W‐MSA) mechanism, enabling the model to better understand the overall distribution and degree of fibrosis of pulmonary lesions. We validated the MBSM‐Net model on 1,760 real anonymized patient X‐ray chest films. The results showed that the accuracy of the MBSM‐Net model reached 78.6%, and the F 1 score reached 79%, both of which are superior to existing models.

Accurate medical image segmentation especially for echocardiographic images with unmissable noise requires elaborate network design. Compared with manual design, Neural Architecture Search (NAS) realizes better segmentation results due to larger search space and automatic optimization, but most of the existing methods are weak in layer-wise feature aggregation and adopt a "strong encoder, weak decoder" structure, insufficient to handle global relationships and local details. To resolve these issues, we propose a novel semi-supervised hybrid NAS network for accurate medical image segmentation termed SSHNN. In SSHNN, we creatively use convolution operation in layer-wise feature fusion instead of normalized scalars to avoid losing details, making NAS a stronger encoder. Moreover, Transformers are introduced for the compensation of global context and U-shaped decoder is designed to efficiently connect global context with local features. Specifically, we implement a semi-supervised algorithm Mean-Teacher to overcome the limited volume problem of labeled medical image dataset. Extensive experiments on CAMUS echocardiography dataset demonstrate that SSHNN outperforms state-of-the-art approaches and realizes accurate segmentation. Code will be made publicly available.

Adopting deep learning in early fundus screening images benefits ocular disease recognition and helps patients avoid blindness in recent years. The robust data representation capability of deep learning is derived from numerous data and annotations. However, the fundus images collected from hospitals or institutes have privacy issues and obvious domain gaps, which greatly influence multi-site learning performance. In this work, a learning system with differential privacy and unsupervised domain regularizer is proposed for ocular disease recognition. First, a Laplace randomized mechanism is introduced to keep the privacy of local models and a global model is constructed via a weighted sum process. Second, an unsupervised domain regularizer, which converts the last fully-connected layer into two sub-layers and then adopts an MMD loss in the element-wise layers of source and target domains, is proposed for unsupervised domain adaptation. Numerous experiments, including four different settings, verify the performance in a multi-disease ocular dataset.

As one of the effective ways of ocular disease recognition, early fundus screening can help patients avoid unrecoverable blindness. Although deep learning is powerful for image-based ocular disease recognition, the performance mainly benefits from a large number of labeled data. For ocular disease, data collection and annotation in a single site usually take a lot of time. If multi-site data are obtained, there are two main issues: 1) the data privacy is easy to be leaked; 2) the domain gap among sites will influence the recognition performance. Inspired by the above, first, a Gaussian randomized mechanism is adopted in local sites, which are then engaged in a global model to preserve the data privacy of local sites and models. Second, to bridge the domain gap among different sites, a two-step domain adaptation method is introduced, which consists of a domain confusion module and a multi-expert learning strategy. Based on the above, a privacy-preserving federated learning framework with domain adaptation is constructed. In the experimental part, a multi-disease early fundus screening dataset, including a detailed ablation study and four experimental settings, is used to show the stepwise performance, which verifies the efficiency of our proposed framework.

Fusing multi-modal radiology and pathology data with complementary information can improve the accuracy of tumor typing. However, collecting pathology data is difficult since it is high-cost and sometimes only obtainable after the surgery, which limits the application of multi-modal methods in diagnosis. To address this problem, we propose comprehensively learning multi-modal radiology-pathology data in training, and only using uni-modal radiology data in testing. Concretely, a Memory-aware Hetero-modal Distillation Network (MHD-Net) is proposed, which can distill well-learned multi-modal knowledge with the assistance of memory from the teacher to the student. In the teacher, to tackle the challenge in hetero-modal feature fusion, we propose a novel spatial-differentiated hetero-modal fusion module (SHFM) that models spatial-specific tumor information correlations across modalities. As only radiology data is accessible to the student, we store pathology features in the proposed contrast-boosted typing memory module (CTMM) that achieves type-wise memory updating and stage-wise contrastive memory boosting to ensure the effectiveness and generalization of memory items. In the student, to improve the cross-modal distillation, we propose a multi-stage memory-aware distillation (MMD) scheme that reads memory-aware pathology features from CTMM to remedy missing modal-specific information. Furthermore, we construct a Radiology-Pathology Thymic Epithelial Tumor (RPTET) dataset containing paired CT and WSI images with annotations. Experiments on the RPTET and CPTAC-LUAD datasets demonstrate that MHD-Net significantly improves tumor typing and outperforms existing multi-modal methods on missing modality situations.

Multimodal volumetric segmentation and fusion are two valuable techniques for surgical treatment planning, image-guided interventions, tumor growth detection, radiotherapy map generation, etc. In recent years, deep learning has demonstrated its excellent capability in both of the above tasks, while these methods inevitably face bottlenecks. On the one hand, recent segmentation studies, especially the U-Net-style series, have reached the performance ceiling in segmentation tasks. On the other hand, it is almost impossible to capture the ground truth of the fusion in multimodal imaging, due to differences in physical principles among imaging modalities. Hence, most of the existing studies in the field of multimodal medical image fusion, which fuse only two modalities at a time with hand-crafted proportions, are subjective and task-specific. To address the above concerns, this work proposes an integration of multimodal segmentation and fusion, namely SegCoFusion, which consists of a novel feature frequency dividing network named FDNet and a segmentation part using a dual-single path feature supplementing strategy to optimize the segmentation inputs and suture with the fusion part. Furthermore, focusing on multimodal brain tumor volumetric fusion and segmentation, the qualitative and quantitative results demonstrate that SegCoFusion can break the ceiling both of segmentation and fusion methods. Moreover, the effectiveness of the proposed framework is also revealed by comparing it with state-of-the-art fusion methods on 2D two-modality fusion tasks, our method achieves better fusion performance than others. Therefore, the proposed SegCoFusion develops a novel perspective that improves the performance in volumetric fusion by cooperating with segmentation and enhances lesion awareness.

Neural architecture search (NAS) has made incredible progress in medical image segmentation tasks, due to its automatic design of the model. However, the search spaces studied in many existing studies are based on U-Net and its variants, which limits the potential of neural architecture search in modeling better architectures. In this study, we propose a new NAS architecture named GNAS-U2Net for the joint segmentation of optic cup and optic disc. This architecture is the first application of NAS in a two-level nested U-shaped structure. The best performance achieved by the joint segmentation model designed by NAS on the REFUGE dataset has an average DICE of 92.88%. Compared to U2-Net and other related work, the model has better performance and uses only 34.79M parameters. We then verify the generalization of the model on two datasets, namely the Drishti-GS dataset and the GAMMA dataset, for which we obtain an average DICE of 92.32% and 92.11% respectively.

Recent digital pathology workflows mainly focus on mono-modality histopathology image analysis. However, they ignore the complementarity between Haematoxylin & Eosin (H&E) and Immunohistochemically (IHC) stained images, which can provide comprehensive gold standard for cancer diagnosis. To resolve this issue, we propose a cross-boosted multi-target domain adaptation pipeline for multi-modality histopathology images, which contains Cross-frequency Style-auxiliary Translation Network (CSTN) and Dual Cross-boosted Segmentation Network (DCSN). Firstly, CSTN achieves the one-to-many translation from fluorescence microscopy images to H&E and IHC images for providing source domain training data. To generate images with realistic color and texture, Cross-frequency Feature Transfer Module (CFTM) is developed to pertinently restructure and normalize high-frequency content and low-frequency style features from different domains. Then, DCSN fulfills multi-target domain adaptive segmentation, where a dual-branch encoder is introduced, and Bidirectional Cross-domain Boosting Module (BCBM) is designed to implement cross-modality information complementation through bidirectional inter-domain collaboration. Finally, we establish Multi-modality Thymus Histopathology (MThH) dataset, which is the largest publicly available H&E and IHC image benchmark. Experiments on MThH dataset and several public datasets show that the proposed pipeline outperforms state-of-the-art methods on both histopathology image translation and segmentation.

In this paper, we propose a progressive margin loss (PML) approach for unconstrained facial age classification. Conventional methods make strong assumption on that each class owns adequate instances to outline its data distribution, likely leading to bias prediction where the training samples are sparse across age classes. Instead, our PML aims to adaptively refine the age label pattern by enforcing a couple of margins, which fully takes in the in-between discrepancy of the intra-class variance, inter-class variance and class center. Our PML typically incorporates with the ordinal margin and the variational margin, simultaneously plugging in the globally-tuned deep neural network paradigm. More specifically, the ordinal margin learns to exploit the correlated relationship of the real-world age labels. Accordingly, the variational margin is leveraged to minimize the influence of head classes that misleads the prediction of tailed samples. Moreover, our optimization carefully seeks a series of indicator curricula to achieve robust and efficient model training. Extensive experimental results on three face aging datasets demonstrate that our PML achieves compelling performance compared to state of the art. Code will be made publicly.

In this paper, we propose an end-to-end open set face anti-spoofing (OSFA) approach for unseen attack recognition. Previous domain generalization approaches aim to align multiple domains beyond one common subspace, leading to performance degradation due to the discrepancy of different domains. To address this issue, our approach formulates face anti-spoofing (FAS) in an open set recognition framework, which learns compact representation for each known class in parallel to recognizing unseen attack examples. To this end, we introduce the statistical extreme value theory incorporated in our objective under the multi-task framework. Moreover, we develop an identity-aware contrastive learning method, preventing us from confusion in unseen attack examples versus hard examples. Experimental results on four datasets demonstrate the robustness of our proposed OSFA, especially under diverse categories of unseen attacks.

In this paper, we propose an invariant learning method for facial landmark mining in a self-supervised manner. The conventional methods mostly train with raw data of paired facial appearances and landmarks, assuming that they are evenly distributed. However, assumptions like this tend to lead to failures in challenging cases even undergo costly training since they usually don't hold in real-world scenarios. To address this issue, our model achieves to be invariant to facial biases by learning through the landmark-anchored distributions. Specifically, we generate faces from these distributions, then group them based on the appearance sources and the probe facial landmarks into intra-identities and intra-landmarks classes, respectively. Thus, we construct intra-class invariance losses to disentangle the spatial structures from appearances. In addition, we adopt a reconstruction loss to produce more realistic faces with probe landmarks. Extensive experimental results on four standard facial landmark datasets demonstrate that our method achieves compelling performance compared with supervised and unsupervised methods.

The guided filter has been acknowledged as an exceptional edge‐preserving filter whose output is a locally linear transform of the guidance image. However, the traditional guided filter heavily relies on the guidance image and fails to achieve the desired result when performing image denoising without a clear guidance image. In this study, to address this limitation, the authors propose a simple yet effective guided filter variant for the single image noise removing. They further show that the proposed denoising strategy can be easily realised by using the iterative framework. Moreover, the weak textured patches based image noise estimation is utilised to generate a clear intermediate image which makes the proposed method highly adaptable to the local noise level. Experimental results demonstrate that their proposed algorithm can compete with the state‐of‐the‐art local denoising methods in edge‐preserving.