More Like this

1. Single image multi-scale enhancement for rock Micro-CT super-resolution using residual U-Net

   https://doi.org/10.1016/j.acags.2024.100165  

   Shan, Liqun; Liu, Chengqian; Liu, Yanchang; Tu, Yazhou; Chilukoti, Sai Venkatesh; Hei, Xiali (June 2024, Applied Computing and Geosciences)

   Micro-CT, also known as X-ray micro-computed tomography, has emerged as the primary instrument for pore-scale properties study in geological materials. Several studies have used deep learning to achieve super-resolution reconstruction in order to balance the trade-off between resolution of CT images and field of view. Nevertheless, most existing methods only work with single-scale CT scans, ignoring the possibility of using multi-scale image features for image reconstruction. In this study, we proposed a super-resolution approach via multi-scale fusion using residual U-Net for rock micro-CT image reconstruction (MS-ResUnet). The residual U-Net provides an encoder-decoder structure. In each encoder layer, several residual sequential blocks and improved residual blocks are used. The decoder is composed of convolutional ReLU residual blocks and residual chained pooling blocks. During the encoding-decoding method, information transfers between neighboring multi-resolution images are fused, resulting in richer rock characteristic information. Qualitative and quantitative comparisons of sandstone, carbonate, and coal CT images demonstrate that our proposed algorithm surpasses existing approaches. Our model accurately reconstructed the intricate details of pores in carbonate and sandstone, as well as clearly visible coal cracks. 

   more » « less
2. A comparison of diceCT and histology for determination of nasal epithelial type

   https://doi.org/10.7717/peerj.12261  

   Smith, Timothy D.; Corbin, Hayley M.; King, Scot E.; Bhatnagar, Kunwar P.; DeLeon, Valerie B. (January 2021, PeerJ)

   Diffusible iodine-based contrast-enhanced computed tomography (diceCT) has emerged as a viable tool for discriminating soft tissues in serial CT slices, which can then be used for three-dimensional analysis. This technique has some potential to supplant histology as a tool for identification of body tissues. Here, we studied the head of an adult fruit bat ( Cynopterus sphinx ) and a late fetal vampire bat ( Desmodus rotundus ) using diceCT and µCT. Subsequently, we decalcified, serially sectioned and stained the same heads. The two CT volumes were rotated so that the sectional plane of the slice series closely matched that of histological sections, yielding the ideal opportunity to relate CT observations to corresponding histology. Olfactory epithelium is typically thicker, on average, than respiratory epithelium in both bats. Thus, one investigator (SK), blind to the histological sections, examined the diceCT slice series for both bats and annotated changes in thickness of epithelium on the first ethmoturbinal (ET I), the roof of the nasal fossa, and the nasal septum. A second trial was conducted with an added criterion: radioopacity of the lamina propria as an indicator of Bowman’s glands. Then, a second investigator (TS) annotated images of matching histological sections based on microscopic observation of epithelial type, and transferred these annotations to matching CT slices. Measurements of slices annotated according to changes in epithelial thickness alone closely track measurements of slices based on histologically-informed annotations; matching histological sections confirm blind annotations were effective based on epithelial thickness alone, except for a patch of unusually thick non-OE, mistaken for OE in one of the specimens. When characteristics of the lamina propria were added in the second trial, the blind annotations excluded the thick non-OE. Moreover, in the fetal bat the use of evidence for Bowman’s glands improved detection of olfactory mucosa, perhaps because the epithelium itself was thin enough at its margins to escape detection. We conclude that diceCT can by itself be highly effective in identifying distribution of OE, especially where observations are confirmed by histology from at least one specimen of the species. Our findings also establish that iodine staining, followed by stain removal, does not interfere with subsequent histological staining of the same specimen. 

   more » « less
3. Generalizability test of a deep learning-based CT image denoising method

   Zeng, Rongping; Lin, Claire Yilin; Li, Qin; Lu, Jiang; Fessler, Jeffrey A.; Myers, Kyle J (January 2020, The 6th International Conference on Image Formation in X-Ray Computed Tomography)

   Deep learning (DL) has been increasingly explored in low-dose CT image denoising. DL products have also been submitted to the FDA for premarket clearance. While having the potential to improve image quality over the filtered back projection method (FBP) and produce images quickly, generalizability of DL approaches is a major concern because the performance of a DL network can depend highly on the training data. In this work we take a residual encoder-decoder convolutional neural network (REDCNN)-based CT denoising method as an example. We investigate the effect of the scan parameters associated with the training data on the performance of this DL-based CT denoising method and identifies the scan parameters that may significantly impact its performance generalizability. This abstract particularly examines these three parameters: reconstruction kernel, dose level and slice thickness. Our preliminary results indicate that the DL network may not generalize well between FBP reconstruction kernels, but is insensitive to slice thickness for slice-wise denoising. The results also suggest that training with mixed dose levels improves denoising performance. 

   more » « less
4. Deep-3D microscope: 3D volumetric microscopy of thick scattering samples using a wide-field microscope and machine learning

   https://doi.org/10.1364/BOE.444488  

   Li, Bowen; Tan, Shiyu; Dong, Jiuyang; Lian, Xiaocong; Zhang, Yongbing; Ji, Xiangyang; Veeraraghavan, Ashok (December 2021, Biomedical Optics Express)

   Confocal microscopy is a standard approach for obtaining volumetric images of a sample with high axial and lateral resolution, especially when dealing with scattering samples. Unfortunately, a confocal microscope is quite expensive compared to traditional microscopes. In addition, the point scanning in confocal microscopy leads to slow imaging speed and photobleaching due to the high dose of laser energy. In this paper, we demonstrate how the advances in machine learning can be exploited to teach a traditional wide-field microscope, one that’s available in every lab, into producing 3D volumetric images like a confocal microscope. The key idea is to obtain multiple images with different focus settings using a wide-field microscope and use a 3D generative adversarial network (GAN) based neural network to learn the mapping between the blurry low-contrast image stacks obtained using a wide-field microscope and the sharp, high-contrast image stacks obtained using a confocal microscope. After training the network with widefield-confocal stack pairs, the network can reliably and accurately reconstruct 3D volumetric images that rival confocal images in terms of its lateral resolution, z-sectioning and image contrast. Our experimental results demonstrate generalization ability to handle unseen data, stability in the reconstruction results, high spatial resolution even when imaging thick (∼40 microns) highly-scattering samples. We believe that such learning-based microscopes have the potential to bring confocal imaging quality to every lab that has a wide-field microscope. 

   more » « less
5. A thinner jacket for frosty and windy climates? Global patterns in leaf cuticle thickness and its environmental associations

   https://doi.org/10.1111/nph.70397  

   Li, Xin'e; Burrows, Geoff E; Dong, Ning; Jordan, Gregory J; Kattge, Jens; Lenz, Tanja I; Niinemets, Ülo; Onoda, Yusuke; Poorter, Hendrik; Sack, Lawren; et al (October 2025, New Phytologist)

   Summary Plant cuticles protect the interior tissues from ambient hazards, including desiccation, UV light, physical wear, herbivores and pathogens. Consequently, cuticle properties are shaped by evolutionary selection.We compiled a global dataset of leaf cuticle thickness (CT) and accompanying leaf traits for 1212 species, mostly angiosperms, from 293 sites representing all vegetated continents. We developed and tested 11 hypotheses concerning ecological drivers of interspecific variation in CT.CT showed clear patterning according to latitude, biome, taxonomic family, site climate and other leaf traits. Species with thick leaves and/or high leaf mass per area tended to have thicker cuticles, as did evergreen relative to deciduous woody species, and species from sites that during the growing season were warmer, had fewer frost days and lower wind speeds, and occurred at lower latitudes. CT–environment relationships were notably stronger among nonwoody than woody species.Heavy investment in cuticle may be disadvantaged at sites with high winds and frequent frosts for ‘economic’ or biomechanical reasons, or because of reduced herbivore pressure. Alternatively, cuticles may become more heavily abraded under such conditions. Robust quantification of CT–trait–environment relationships provides new insights into the multiple roles of cuticles, with additional potential use in paleo‐ecological reconstruction. 

   more » « less