Unsupervised Domain Adaptation for EM Image Denoising with Invertible Networks.

Electron microscopy (EM) image denoising is critical for visualization and subsequent analysis. Despite the remarkable achievements of deep learning-based non-blind denoising methods, their performance drops significantly when domain shifts exist between the training and testing data. To address this issue, unpaired blind denoising methods have been proposed. However, these methods heavily rely on image-to-image translation and neglect the inherent characteristics of EM images, limiting their overall denoising performance. In this paper, we propose the first unsupervised domain adaptive EM image denoising method, which is grounded in the observation that EM images from similar samples share common content characteristics. Specifically, we first disentangle the content representations and the noise components from noisy images and establish a shared domain-agnostic content space via domain alignment to bridge the synthetic images (source domain) and the real images (target domain). To ensure precise domain alignment, we further incorporate domain regularization by enforcing that: the pseudo-noisy images, reconstructed using both content representations and noise components, accurately capture the characteristics of the noisy images from which the noise components originate, all while maintaining semantic consistency with the noisy images from which the content representations originate. To guarantee lossless representation decomposition and image reconstruction, we introduce disentanglement-reconstruction invertible networks. Finally, the reconstructed pseudo-noisy images, paired with their corresponding clean counterparts, serve as valuable training data for the denoising network. Extensive experiments on synthetic and real EM datasets demonstrate the superiority of our method in terms of image restoration quality and downstream neuron segmentation accuracy. Our code is publicly available at https://github.com/sydeng99/DADn.

MRI advances in the imaging diagnosis of tuberculous meningitis: opportunities and innovations.

Tuberculous meningitis (TBM) is not only one of the most fatal forms of tuberculosis, but also a major public health concern worldwide, presenting grave clinical challenges due to its nonspecific symptoms and the urgent need for timely intervention. The severity and the rapid progression of TBM underscore the necessity of early and accurate diagnosis to prevent irreversible neurological deficits and reduce mortality rates. Traditional diagnostic methods, reliant primarily on clinical findings and cerebrospinal fluid analysis, often falter in delivering timely and conclusive results. Moreover, such methods struggle to distinguish TBM from other forms of neuroinfections, making it critical to seek advanced diagnostic solutions. Against this backdrop, magnetic resonance imaging (MRI) has emerged as an indispensable modality in diagnostics, owing to its unique advantages. This review provides an overview of the advancements in MRI technology, specifically emphasizing its crucial applications in the early detection and identification of complex pathological changes in TBM. The integration of artificial intelligence (AI) has further enhanced the transformative impact of MRI on TBM diagnostic imaging. When these cutting-edge technologies synergize with deep learning algorithms, they substantially improve diagnostic precision and efficiency. Currently, the field of TBM imaging diagnosis is undergoing a phase of technological amalgamation. The melding of MRI and AI technologies unquestionably signals new opportunities in this specialized area.