Search Results (2)

Cardiovascular diseases are the most common cause of death worldwide, and they still have dramatic consequences on the patients’ lives. Murine models are often used to study the anatomical and microstructural changes caused by the diseases. Contrast-enhanced microfocus computed tomography (CECT) is a new imaging technique for 3D histology of biological tissues. In this study, we confirmed the nondestructiveness of Hf-WD 1:2 POM-based CECT and cryogenic CECT (cryo-CECT) to image the heart in 3D. The influence of the image quality (i.e., acquisition time and spatial resolution) was assessed for the characterization of the heart structural constituents: heart integrity, the coronary blood vessels and the heart valves. Coronary blood vessels were visualized and segmented in murine hearts, allowing us to distinguish veins from arteries and to visualize the 3D spatial distribution of the right coronary artery and the left main coronary artery. Finally, to demonstrate the added value of 3D imaging, the thickness distribution of the two leaflets in the mitral valve and three cusps in the aortic valve was computed in 3D. This study corroborates the added value of CECT and cryo-CECT compared to classical 2D histology to characterize ex vivo the structural properties of murine hearts and paves the way for the detailed 3D (micro)structural analyses of future cardiovascular disease models obtained in mice and rats. Full article

Salient segmentation is a critical step in biomedical image analysis, aiming to cut out regions that are most interesting to humans. Recently, supervised methods have achieved promising results in biomedical areas, but they depend on annotated training data sets, which requires labor and proficiency in related background knowledge. In contrast, unsupervised learning makes data-driven decisions by obtaining insights directly from the data themselves. In this paper, we propose a completely unsupervised self-aware network based on pre-training and attentional backpropagation for biomedical salient segmentation, named as PUB-SalNet. Firstly, we aggregate a new biomedical data set from several simulated Cellular Electron Cryo-Tomography (CECT) data sets featuring rich salient objects, different SNR settings, and various resolutions, which is called SalSeg-CECT. Based on the SalSeg-CECT data set, we then pre-train a model specially designed for biomedical tasks as a backbone module to initialize network parameters. Next, we present a U-SalNet network to learn to selectively attend to salient objects. It includes two types of attention modules to facilitate learning saliency through global contrast and local similarity. Lastly, we jointly refine the salient regions together with feature representations from U-SalNet, with the parameters updated by self-aware attentional backpropagation. We apply PUB-SalNet for analysis of 2D simulated and real images and achieve state-of-the-art performance on simulated biomedical data sets. Furthermore, our proposed PUB-SalNet can be easily extended to 3D images. The experimental results on the 2d and 3d data sets also demonstrate the generalization ability and robustness of our method. Full article