Search

Improving the segmentation accuracy of cells from microscopy images directly benefits medical practitioners for differential cell counting, disease diagnosis, and treatment. However, separating each cell continues to be challenging, especially when the cells appear in clumps. Recent advancements in deep learning have demonstrated success in automating cell segmentation and counting, but they necessitate large volumes of annotated data for training. Creating precise ground truth for cell segmentation is laborious and demands expertise due to the intricate and varied characteristics of microscopy images. Consequently, this complexity poses a challenge to the practicality of implementing deep neural network-based solutions for microscopy cell segmentation. This paper addresses a significant research gap by exploring the underutilized traditional image features to build a lightweight deep-learning framework for cell boundary segmentation that can be effectively trained with a limited amount of labeled data. We have developed Saliency Guided Attention Module Network (SGAM-net), which uses perceptual features inherent along the cell boundaries and poses cell segmentation as a thick boundary prediction problem with the aim to separate the cells. SGAM-net uses a specialized attention mechanism to combine the perceptual feature maps with the learned features. When tested on the three public bench-mark microscopy datasets, with respect to the base Unet model, there is an increase of 1.36%, 1.29%, and 6.35% in Dice score and 2.39%, 2.31% and 6.99% in Intersection Over Union (IOU) score. The model requires only 23.07% less parameters than Unet++ model. The potential of SGAM-net in segmenting boundaries semantically has been demonstrated, and it is further processed to compute cell counts. The model outperformed the state-of-the-art instance separation method 'Stardist' in cell counting by giving a reduction ratio of 2.34, 1.89, and 2.03 on the metric 'mean absolute error'. • Propose a novel deep learning method for boundary semantic segmentation of cells. • Utilize both perceptual and learned features of the images for segmentation. • Employs lightweight model with specialized attention gate for better segmentation. • Applicable to estimate cell counts in cancer diagnosis and drug discovery systems. [ABSTRACT FROM AUTHOR]

Cell segmentation is the most significant task in microscopic image analysis as it facilitates differential cell counting and analysis of sub-cellular structures for diagnosing cytopathological diseases. Bright-field microscopy is considered the gold standard among different types of optical microscopes used for cell analysis due to its simplicity and cost-effectiveness. However, automatic cell segmentation in bright field microscopy is challenging due to imaging artifacts, poor contrast, overlapping cells, and wide variability of cells. Also, the availability of labeled bright-field images is limited, further constraining the research in developing supervised models for automated cell segmentation. In this research, we propose a novel cell segmentation framework termed Saliency and Ballness driven U-shaped Network (SBU-net) to overcome these challenges. The proposed architecture comprises a novel data-driven feature fusion module that enhances the perceivable structure of cells using its saliency and ballness features. This, together with an encoder–decoder model having dilated convolutions and a novel combination loss function, captured the global context of cell structures and produced accurate cell segmentation results. SBU-net is evaluated using two publicly available bright-field datasets of T cells and pancreatic cancer cells. The model is subjected to 5-fold cross-validation and outperformed state-of-the-art models by producing mean Intersection over Union (IoU) scores of 0.804, 0.829, and mean Dice of 0.891, 0.906, respectively. The architecture was also tested on a fluorescent dataset to see how well it could generalize, and it came out with a mean IoU of 0.892 and a mean Dice of 0.948, outperforming other models reported in the literature. [ABSTRACT FROM AUTHOR]

Abstract: There is significant interest in the development of new materials for chemical hydrogen storage for the transportation sector. Electronic structure calculations using various methods, reveal that the borane molecule (BH3) could act as an efficient bifunctional acid–base catalyst in the H2-elimination reactions of XH n YH n systems (X,Y=B,C,N). Such a catalyst is needed as the elimination of H2 from isoelectronic ethane and borane amine compounds, proceeds with an energy barrier much higher than that of the A–B bond energy. The catalytic effect of BH3 has been probed by an analysis of the electronic densities of the transition structures using the atoms-in-molecules (AIM) and electron localization function (ELF) approaches. [Copyright &y& Elsevier]

Abstract: The boron buckyball avoids the high symmetry icosahedral cage structure. The previously reported I h symmetric structure is not an energy minimum in the potential energy surface and exhibits a spontaneous symmetry breaking to yield a puckered cage with a rare T h symmetry. The HOMO–LUMO gap is twice as large as the reported value and amounts to 1.94eV at B3LYP/6-31G(d) level. The valence orbital structure of boron buckyball is identical to the one in the carbon analogue. [Copyright &y& Elsevier]

Rapid fluctuations in the amplitude and phase of L-band GNSS signals occur, when they pass through the ionosphere, because of electron density irregularities. This so-called scintillation is a consequence of electron density irregularities mainly in the E and F regions of the ionosphere. The daytime scintillation is proven to be linked with E-region irregularities by the existence of a cloud of intense sporadic-E (E s) in the path of radio waves, which produces scintillation even during the daytime at low latitudes. In this study, we have reported the presence of L-band (lower and upper) scintillation observed using data from a newly established GNSS station (SHJ1) situated under the northern crest of the equatorial ionization anomaly (EIA) near Arabian Peninsula at 25.3°N and 55.5°E. The presence of daytime L-band scintillation is found to be consistent with the appearance of E s. A significant presence of weak, moderate and strong scintillation has been observed throughout the day, in general, and in the late afternoon hours (LT 1500–1800), in particular. However, no significant post-sunset and post-midnight scintillation has been observed. The reported pattern of scintillation occurrences over Sharjah differs in many regards to the previous studies, specifically, the local time of peak occurrences. Using RO data, it has been concluded that E s does exist around Sharjah and the occurrence of it appears to be associated with the appearance of scintillation during pre-sunset hours. [ABSTRACT FROM AUTHOR]

Abstract: Despite the remarkable success of the standard model, it is widely believed that it is just an intermediate step in understanding the properties of the elementary particles in nature and the interactions between them. Studies of parity non-conservation in atomic systems based on non-accelerator methods have advanced over the years and they have now attained the maturity to test the standard model. The most accurate data from this approach comes from the 6s2S1/2→7s2S1/2 transition in cesium. An experiment on singly ionized barium has been proposed to carry out a further test of this kind independent of cesium. We discuss here the result of our ab initio calculation of the 6s2S1/2→5d2D3/2 parity non-conserving electric dipole transition amplitude in 137Ba+ based on a novel relativistic coupled-cluster theory. We also discuss a possible improvement of the existing cesium parity non-conservation calculations. [Copyright &y& Elsevier]