Synthesis of myocardial infarction enhancement (MIE) images without contrast agents (CAs) has shown great potential to advance myocardial infarction (MI) diagnosis and treatment. It provides results comparable to late gadolinium enhancement (LGE) images, thereby reducing the risks associated with CAs and streamlining clinical workflows. The existing knowledge-and-data-driven approach has made progress in addressing the complex challenges of synthesizing MIE images (i.e., invisible myocardial scars and high inter-individual variability) but still has limitations in the interpretability of kinematic inference, morphological knowledge integration, and kinematic-morphological fusion, thereby reducing the transparency and reliability of the model and causing information loss during synthesis. In this paper, we proposed a knowledge-driven interpretative conditional diffusion model (K-ICDM), which learns kinematic and morphological information from non-enhanced cardiac MR images (CINE sequence and T1 sequence) guided by cardiac knowledge, enabling the synthesis of MIE images. Importantly, our K-ICDM introduces three key innovations that address these limitations, thereby providing interpretability and improving synthesis quality. (1) A novel cardiac causal intervention that generates counterfactual strain to intervene in the inference process from motion maps to abnormal myocardial information, thereby establishing an explicit relationship and providing the clear causal interpretability. (2) A knowledge-driven cognitive combination strategy that utilizes cardiac signal topology knowledge to analyze T1 signal variations, enabling the model to understand how to learn morphological features, thus providing interpretability for morphology capture. (3) An information-specific adaptive fusion strategy that integrates kinematic and morphological information into the conditioning input of the diffusion model based on their specific contributions and adaptively learns their interactions, thereby preserving more detailed information. Experiments on a broad MI dataset with 315 patients show that our K-ICDM achieves state-of-the-art performance in contrast-free MIE image synthesis, improving structural similarity index measure (SSIM) by at least 2.1% over recent methods. These results demonstrate that our method effectively overcomes the limitations of existing methods in capturing the complex relationship between myocardial motion and scar distribution and integrating of static and dynamic sequences, thus enabling the accurate synthesis of subtle scar boundaries.

The stable diffusion model (SDM) is a prevalent and effective model for text-to-image (T2I) and image-to-image (I2I) generation. Despite various attempts at sampler optimization, model distillation, and network quantification, these approaches typically maintain the original network architecture. The extensive parameter scale and substantial computational demands have limited research into adjusting the model architecture. This study focuses on reducing redundant computation in SDM and optimizes the model through both tuning and tuning-free methods: 1) for the tuning method, we design a model assembly strategy to reconstruct a lightweight model while preserving performance and ensuring semantic stability through distillation and 2) for the tuning-free method, we propose a feature inheritance strategy to accelerate inference by skipping local computations at the block, layer, or unit level within the network structure. We also examine multiple sampling modes for feature inheritance at the time-step level. Experiments demonstrate that both the proposed tuning and the tuning-free methods can improve the speed and performance of the SDM. The lightweight model reconstructed by the model assembly strategy increases generation speed by 22.4%, while the feature inheritance strategy enhances the SDM generation speed by 40.0%.

SUMMARY: To address the challenges in single-cell metabolomics (SCM) research, we have developed an open-source Python-based modular library, named SCMeTA, for SCM data processing. We designed standardized pipeline and inter-container communication format and have developed modular components to adapt to the diverse needs of SCM studies. The validation was carried out on multiple SCM experiment data. The results demonstrated significant improvements in batch effects, accuracy of results, metabolic extraction rate, cell matching rate, as well as processing speed. This library is of great significance in advancing the practical application of SCM analysis and makes a foundation for wide-scale adoption in biological studies. AVAILABILITY AND IMPLEMENTATION: SCMeTA is freely available on https://github.com/SCMeTA/SCMeTA and https://doi.org/10.5281/zenodo.13569643.

Training generative adversarial networks (GANs) with limited data is challenging because the discriminator is prone to overfitting. Previously proposed differentiable augmentation demonstrates improved data efficiency of training GANs. However, the augmentation implicitly introduces undesired invariance to augmentation for the discriminator since it ignores the change of semantics in the label space caused by data transformation, which may limit the representation learning ability of the discriminator and ultimately affect the generative modeling performance of the generator. To mitigate the negative impact of invariance while inheriting the benefits of data augmentation, we propose a novel augmentation-aware self-supervised discriminator that predicts the augmentation parameter of the augmented data. Particularly, the prediction targets of real data and generated data are required to be distinguished since they are different during training. We further encourage the generator to adversarially learn from the self-supervised discriminator by generating augmentation-predictable real and not fake data. This formulation connects the learning objective of the generator and the arithmetic $-$ harmonic mean divergence under certain assumptions. We compare our method with state-of-the-art (SOTA) methods using the class-conditional BigGAN and unconditional StyleGAN2 architectures on data-limited CIFAR-10, CIFAR-100, FFHQ, LSUN-Cat, and five low-shot datasets. Experimental results demonstrate significant improvements of our method over SOTA methods in training data-efficient GANs.

Conditional generative models aim to learn the underlying joint distribution of data and labels to achieve conditional data generation. Among them, the auxiliary classifier generative adversarial network (AC-GAN) has been widely used, but suffers from the problem of low intra-class diversity of the generated samples. The fundamental reason pointed out in this paper is that the classifier of AC-GAN is generator-agnostic, which therefore cannot provide informative guidance for the generator to approach the joint distribution, resulting in a minimization of the conditional entropy that decreases the intra-class diversity. Motivated by this understanding, we propose a novel conditional GAN with an auxiliary discriminative classifier (ADC-GAN) to resolve the above problem. Specifically, the proposed auxiliary discriminative classifier becomes generator-aware by recognizing the class-labels of the real data and the generated data discriminatively. Our theoretical analysis reveals that the generator can faithfully learn the joint distribution even without the original discriminator, making the proposed ADC-GAN robust to the value of the coefficient hyperparameter and the selection of the GAN loss, and stable during training. Extensive experimental results on synthetic and real-world datasets demonstrate the superiority of ADC-GAN in conditional generative modeling compared to state-of-the-art classifier-based and projection-based conditional GANs.

Although significant progress has been made in generating images from the text by using generative adversarial networks (GANs), it is still challenging to deal with long text, which contains complex semantic information like recipes. This paper focuses on generating images with high visual realism and semantic consistency from the complex text of recipes. To achieve this, we propose a GANs based method termed ChefGAN. The critical concept of ChefGAN is that a joint image-recipe embedding model is used before the generation task to provide high-quality representations of recipes, and it acts as an extra regularization during the generation to improve semantic consistency. Two modules are designed for this image text embedding module (ITEM) and a cascaded image generation module (CIGM). The generation process is carried out in 3 steps: (1) Two encoders in ITEM are trained simultaneously to generate similar representations for each image-recipe pair. (2) CIGM generates images according to the representations from ITEM's text encoder. (3) The generated image is fed into ITEM's image encoder to calculate the similarity with the given recipe. This process can provide additional regularization effect other than the impact of a discriminator. To facilitate convergence, we applied a two-stage training strategy, which generates an image with low resolution and then one with high resolution in the CIGM module. Compared with other representative state-of-the-art methods, ChefGAN demonstrates better performance both in visual realism and semantic consistency.