A simultaneous enhancement of accuracy and diversity of predictions remains a challenge in ambiguous medical image segmentation (AMIS) due to the inherent trade-offs. While truncated diffusion probabilistic models (TDPMs) hold strong potential with a paradigm optimization, existing TDPMs suffer from entangled accuracy and diversity of predictions with insufficient fidelity and plausibility. To address the aforementioned challenges, we propose Ambiguity-aware Truncated Flow Matching (ATFM), which introduces a novel inference paradigm and dedicated model components. Firstly, we propose Data-Hierarchical Inference, a redefinition of AMIS-specific inference paradigm, which enhances accuracy and diversity at data-distribution and data-sample level, respectively, for an effective disentanglement. Secondly, Gaussian Truncation Representation (GTR) is introduced to enhance both fidelity of predictions and reliability of truncation distribution, by explicitly modeling it as a Gaussian distribution at Ttrunc instead of using sampling-based approximations. Thirdly, Segmentation Flow Matching (SFM) is proposed to enhance the plausibility of diverse predictions by extending semantic-aware flow transformation in Flow Matching (FM). Comprehensive evaluations on LIDC and ISIC3 datasets demonstrate that ATFM outperforms SOTA methods and simultaneously achieves a more efficient inference. ATFM improves GED and HM-IoU by up to 12% and 7.3% compared to advanced methods.

Knowledge distillation can mitigate catastrophic forgetting in domain continual segmentation by transferring knowledge from the older model to the newer model. However, existing distillation-based methods primarily emphasize knowledge retention while overlooking inherent defects in the older teacher models. As a result, these teacher-originated defects, such as knowledge gaps or biases, are propagated and exacerbate forgetting. To address this challenge, we propose a Targeted Knowledge Rectification Learning framework (TKRL) to probe and correct teacher-originated defects. TKRL consists of two modules: 1) Probe-augmented Class Distillation, which generates gradient-driven "probes" to uncover underrepresented features in the older model, thereby bridging knowledge gaps by distilling hidden information into the new model; 2) Variance-guided Masked Autoencoder, which selectively masks and reconstructs critical high-uncertainty patches across multi-level semantic regions, thereby correcting biases inherited from the older model. Our experimental results show that TKRL effectively rectifies knowledge gaps and biases, thereby mitigating catastrophic forgetting and enhancing performance in domain continual segmentation.

Bias fields inevitably degrade MRI that seriously interferes the diagnosis of physicians for accurate analysis, and removing it is a crucial image analysis task. Generative models (such as GANs) are used for bias field correction, and outperform traditional methods, however are hindered by the high cost of data annotation and instability during training. Recently, the diffusion-based methods have excelled over GANs in many applications, and they are powerful in removing noise from images, while the bias field can be regarded as a smooth noise. However, it is a challenge to directly apply to 3D bias field correction due to sampling inefficiency, the heavy computational demand, and implicit correction process. We propose a Memory-Efficient Multivariate Gaussian Bias Diffusion Model (MeMGB-Diff) that is an explicit, sampling, and memory both efficient diffusion model for 3D bias field correction without using clinical labels. MeMGB-Diff extends the diffusion models to multivariate Gaussian and models the bias field as a multivariate Gaussian variable, allowing direct diffusion and removal of the 3D bias fields without Gaussian noise. For memory efficiency, MeMGB-Diff performs diffusion model in smaller readable image domain at the expense of a negligible accuracy loss, based on the strong correlation among adjacent voxels of bias field. We also propose a loss function to mainly learn the intensity trend, which mainly causes the inhomogeneity of MRI, and effectively increases the correction accuracy. For comprehensive performance comparison, we propose a synthetic method for generating more varied bias fields during testing. Both quantitative and qualitative assessments on synthetic and clinical data confirm the high fidelity and uniform intensity of our results. MeMGB-Diff reduces data size by 64 times to use less memory, improves sampling efficiency by more than 10 times compared to other diffusion-based methods, and achieves optimal metrics, including SSIM, PSNR, COCO, and CV for various tissues. Hence, our MeMGB-Diff is a state-of-the-art (SOTA) method for 3D bias field correction.

Automatic acne severity grading is crucial for the accurate diagnosis and effective treatment of skin diseases. However, the acne severity grading process is often ambiguous due to the similar appearance of acne with close severity, making it challenging to achieve reliable acne severity grading. Following the idea of fuzzy logic for handling uncertainty in decision-making, we transforms the acne severity grading task into a fuzzy label learning (FLL) problem, and propose a novel adjacency-aware fuzzy label learning (AFLL) framework to handle uncertainties in this task. The AFLL framework makes four significant contributions, each demonstrated to be highly effective in extensive experiments. First, we introduce a novel adjacency-aware decision sequence generation method that enhances sequence tree construction by reducing bias and improving discriminative power. Second, we present a consistency-guided decision sequence prediction method that mitigates error propagation in hierarchical decision-making through a novel selective masking decision strategy. Third, our proposed sequential conjoint distribution loss innovatively captures the differences for both high and low fuzzy memberships across the entire fuzzy label set while modeling the internal temporal order among different acne severity labels with a cumulative distribution, leading to substantial improvements in FLL. Fourth, to the best of our knowledge, AFLL is the first approach to explicitly address the challenge of distinguishing adjacent categories in acne severity grading tasks. Experimental results on the public ACNE04 dataset demonstrate that AFLL significantly outperforms existing methods, establishing a new state-of-the-art in acne severity grading.

Medical vision-language pretraining (VLP) that leverages naturally-paired medical image-report data is crucial for medical image analysis. However, existing methods struggle to accurately characterize associations between images and diseases, leading to inaccurate or incomplete diagnostic results. In this work, we propose MedFILIP, a fine-grained VLP model, introduces medical image-specific knowledge through contrastive learning, specifically: 1) An information extractor based on a large language model is proposed to decouple comprehensive disease details from reports, which excels in extracting disease deals through flexible prompt engineering, thereby effectively reducing text complexity while retaining rich information at a tiny cost. 2) A knowledge injector is proposed to construct relationships between categories and visual attributes, which help the model to make judgments based on image features, and fosters knowledge extrapolation to unfamiliar disease categories. 3) A semantic similarity matrix based on fine-grained annotations is proposed, providing smoother, information-richer labels, thus allowing fine-grained image-text alignment. 4) We validate MedFILIP on numerous datasets, e.g., RSNA-Pneumonia, NIH ChestX-ray14, VinBigData, and COVID-19. For single-label, multi-label, and fine-grained classification, our model achieves state-of-the-art performance, the classification accuracy has increased by a maximum of 6.69%.

Ultrasound high-resolution imaging is essential in clinic. The limited number of channels in portable ultrasound machines results in low-resolution images being acquired. Achieving high-resolution ultrasound images relies on machines with a substantial number of channels, which increases costs. Obtaining potential high-channel images from low-channel images can significantly reduce the cost of ultrasound machines. A novel physics-based deep learning method spatially-variant model (SV-Net) is proposed to deconvolve low-channel ultrasound images, yielding high-resolution outputs. SV-Net is structured with a multi-channel Wiener deconvolution layer placed before Convolutional Neural Network (CNN). The multi-channel Wiener deconvolution layer consists of various differentiable Wiener deconvolutions, which leverage knowledge of spatially-variant Point Spread Functions (PSFs) in ultrasound images. These PSFs are subsequently optimized through further training to obtain high-quality ultrasound images. Experiments show the method's efficacy in improving lateral resolution and achieving favorable performance metrics, including PSNR, SSIM, MAE, and FWHM.

The application of transformer architecture in analyzing whole slide images (WSIs) has become increasingly popular due to its remarkable ability to learn complex associations. Nevertheless, a significant drawback emerges in the multiclass analysis of WSIs. The majority of the transformer-based methods available currently rely primarily on a single, class-agnostic token. This approach might not ideally capture the subtleties of class-discriminative information. To address this challenge, we present an innovative approach tailored for multi-class WSI analysis that harnesses the power of class-specific tokens. Central to our method is a novel attention mechanism designed to foster a synergistic learning relationship between patch and class tokens, enhancing the granularity of information captured and ensuring a more comprehensive representation of the WSI. Complementing this, we introduce a dynamic class-centric training strategy designed to optimize token representation learning, ensuring each token is informatively aligned with its corresponding class. Through extensive experimentation on three challenging multi-class WSI analysis datasets, our method consistently demonstrates superior performance, underscoring its potential as a robust solution for multi-class WSI analysis tasks.

Accurate segmentation of the Intracranial Hemorrhage (ICH) in non-contrast CT images is significant for computer-aided diagnosis. Although existing methods have achieved remarkable 1 1 The code will be available from https://github.com/JohnleeHIT/SLEX-Net. results, none of them incorporated ICH's prior information in their methods. In this work, for the first time, we proposed a novel SLice EXpansion Network (SLEX-Net), which incorporated hematoma expansion in the segmentation architecture by directly modeling the hematoma variation among adjacent slices. Firstly, a new module named Slice Expansion Module (SEM) was built, which can effectively transfer contextual information between two adjacent slices by mapping predictions from one slice to another. Secondly, to perceive contextual information from both upper and lower slices, we designed two information transmission paths: forward and backward slice expansion, and aggregated results from those paths with a novel weighing strategy. By further exploiting intra-slice and inter-slice context with the information paths, the network significantly improved the accuracy and continuity of segmentation results. Moreover, the proposed SLEX-Net enables us to conduct an uncertainty estimation with one-time inference, which is much more efficient than existing methods. We evaluated the proposed SLEX-Net and compared it with some state-of-the-art methods. Experimental results demonstrate that our method makes significant improvements in all metrics on segmentation performance and outperforms other existing uncertainty estimation methods in terms of several metrics.