Search

Sea fog can seriously affect schedules and safety by reducing visibility during marine transportation. Therefore, the forecasting of sea fog is an important issue in preventing accidents. Recently, in order to forecast sea fog, several deep learning methods have been applied to time series data consisting of meteorological and oceanographic observations or image data to predict fog. However, these methods only use a single image without considering meteorological and temporal characteristics. In this study, we propose a multi-modal learning method to improve the forecasting accuracy of sea fog using convolutional neural network (CNN) and gated recurrent unit (GRU) models. CNN and GRU extract useful features from closed-circuit television (CCTV) images and multivariate time series data, respectively. CCTV images and time series data collected at Daesan Port in South Korea from 1 March 2018 to 14 February 2021 by Korea Hydrographic and Oceanographic Agency (KHOA) were used to evaluate the proposed method. We compare the proposed method with deep learning methods that only consider temporal information or spatial information. The results indicate that the proposed method using both temporal and spatial information at the same time shows superior accuracy.

Sea fog is a natural phenomenon that reduces the visibility of manned vehicles and vessels that rely on the visual interpretation of traffic. Fog clearance, also known as fog dissipation, is a relatively under-researched area when compared with fog prediction. In this work, we first analyzed meteorological observations that relate to fog dissipation in Incheon port (one of the most important ports for the South Korean economy) and Haeundae beach (the most populated and famous resort beach near Busan port). Next, we modeled fog dissipation using two separate algorithms, classification and regression, and a model with nine machine learning and three deep learning techniques. In general, the applied methods demonstrated high prediction accuracy, with extra trees and recurrent neural nets performing best in the classification task and feed-forward neural nets in the regression task.

The molecular mechanisms for epithelial differentiation have been studied by observing skin development in embryogenesis, but the early signaling modulations involved in tongue epithelial differentiation are not completely understood. Based on the gene expression patterns of the Fgf signaling molecules and previous results from Fgf10 and Fgfr2b knockout mice, it was hypothesized that there would be fundamental signaling interactions through the epithelial Fgfr2b and its mesenchymal ligand Fgf10 to regulate tongue epithelium differentiation. To elucidate these reciprocal interactions in tongue epithelial differentiation, this study employed an in vitro tongue organ culture system with antisense-oligodeoxynucleotides (AS-ODNs) and recombinant protein-soaked bead implantation for the loss-of-function and gain-of-function studies. Functional analysis of Fgf signaling revealed precise reciprocal interactions, which showed that mesenchymal Fgf10 rather than Fgf7 modulates tongue epithelial differentiation via Fgfr2b in a temporal- and spatial-specific manner.

♦ Background Human peritoneal mesothelial cells (HPMCs) undergo a transition from an epithelial phenotype to a mesenchymal phenotype (EMT) during peritoneal dialysis (PD). That transition may be directly related to failure of peritoneal membrane function. ♦ Objective In a randomized prospective controlled study, we investigated the effect of low glucose degradation product (GDP) dialysis solution on the transition of HPMCs. ♦ Methods Among new continuous ambulatory PD patients, 60 patients completed a 12-month protocol (low-GDP solution group, n = 32; high-GDP solution group, n = 28). At the 1st, 6th, and 12th months, HPMCs drained from overnight effluent were cultured on T25. When they had nearly reached confluence, cell scores were measured blindly by the same person (score 1 = cobblestone-shaped HPMCs, score 2 = mixed, score 3 = fibroblast dominant). Cell scores and clinical indices, including peritoneal markers, were compared between the low-GDP and high-GDP groups at the 1st, 6th, and 12th months. The factors associated with EMT were analyzed with generalized estimating equations using STATA 7.0 (STATA Corp., College Station, Texas, USA). In addition, vimentin and cytokeratin 8/18 stains were used to verify EMT in cultured cells and peritoneal specimens in some patients. ♦ Results ( 1 ) The low-GDP group showed higher dialysate cancer antigen 125 levels from the 1st to 12th months (55.4 ± 24.8 vs 8.8 ± 1.7, 56.7 ± 28.1 vs 22.1 ± 11.3, and 54.2 ± 28.2 vs 24.6 ± 16.5 U/mL, at the 1st, 6th, and 12th months, respectively; all p = 0.000). ( 2 ) The low-GDP group showed lower cell scores at the 1st, 6th, and 12th months (1.22, 1.22, and 1.56 vs 1.61, 1.75, and 2.14; p < 0.05, p < 0.01, and p < 0.01, respectively). ( 3 ) At the 12th month, the number of fibroblast-dominant cultures (score 3) was significantly lower in the low-GDP group [4/32 (12.5%) patients vs 14/28 (50%), p < 0.05]. ( 4 ) Both cobblestone-shaped HPMCs and fibroblastoid cells were positively stained with cytokeratin and vimentin. ( 5 ) There were many cytokeratin- and vimentin-positive cells in the submesothelial area in the peritoneal biopsy specimens. (6) The consistent factor associated with EMT was only high-GDP solution (60 patients, n = 178, beta coefficient 0.312, p = 0.000; 46 patients, n = 137, beta coef: 0.228, p = 0.000) and not numbers of peritonitis episodes, duration of angiotensin-converting enzyme inhibitor and/or angiotensin receptor blocker medication, or diabetes. ♦ Conclusion Low-GDP solution showed beneficial effects such as rapid remesothelialization and less EMT in the peritoneum with time on PD.