RESUMEN
AIM: To explore the application value of deep learning technology in automatic meibomian glands segmentation. METHODS:Infrared meibomian gland images were collected and 193 of them were picked out for establishing the database. The images were manually labeled by three clinicians. UNet++ network and automatic data expansion strategy were introduced to construct the automatic meibomian glands segmentation model. The feasibility and effectiveness of the proposed segmentation model were analyzed by precision, sensitivity, specificity, accuracy and intersection over union.RESULTS: Taking manual labeling as the gold standard, the presented method segment the glands effectively and steadily with accuracy, sensitivity and specificity of 94.31%, 82.15% and 96.13% respectively. On the average, only 0.11s was taken for glands segmentation of single image.CONCLUSIONS: In this paper, deep learning technology is introduced to realize automatic segmentation of meibomian glands, achieving high accuracy, good stability and efficiency. It would be quite useful for calculation of gland morphological parameters, the clinical diagnosis and screening of related diseases, improving the diagnostic efficiency.
RESUMEN
Effect of Fe(III) concentration on nitrogen immigration and transformation and nitrous oxide emission during the simultaneous nitrification denitrification (SND) process was investigated. Higher nitrogen removal efficiency was obtained when the Fe(III) concentration was 20 mg x L(-1), while lower nitrogen removal efficiency was observed when the Fe (III) concentration turned to 60 mg x L(-1). In addition, higher Fe(III) concentration significantly enhanced the N2O emission, as well as the N2O conversion ratio. This was mainly attributed to (1) the high concentration of nitrite accumulation during the oxic stage, which was caused by lower dehydrogenase activity at high Fe(III) concentration; (2) less PHB production during the anoxic stage, which would led to shortage of carbon source for denitrification in the following oxic stage. The results also showed that Fe(III) addition could improve the TP removal efficiency. TP removal efficiency increased with increasing Fe(III) concentration, mainly because of extra chemical reaction.
Asunto(s)
Desnitrificación , Compuestos Férricos/química , Nitrificación , Nitrógeno/química , Óxido Nitroso/química , Reactores Biológicos , Nitritos , Oxidación-Reducción , Eliminación de Residuos LíquidosRESUMEN
The nitrous oxide fluxes and ammonia-oxidizing bacterium in two typical constructed wetlands, i.e. subsurface flow (SF) and free water surface (FWS) were studied by the method of static chamber-gas chromatography. The results showed that the mean N2O fluxes were 296.5 microg x (m2 x h)(-1) and 28.2 microg x (m2 x h)(-1) respectively, and two typical wetlands were all the sources of atmosphere nitrous oxide as a whole. SF wetland exhibited a higher risk of N2O emissions, and the mean N2O flux in this system was higher than the values reported in the literature for ecosystems, e.g. farmland, forest, grassland and marsh. The nitrous oxide fluxes in test wetlands presented obvious seasonal and diurnal variation, and the highest N2O emission flux was in July. The highest flux was (762.9 +/- 239.3) microg x (m2 x h)(-1) and (91.9 +/- 20.3) microg x (m2 x h)(-1) in SF and FWS wetlands, respectively. The peak flux mostly occurred around midday, whereas the minimum flux likely occurred in the early morning. The results indicated that the growth of Phragmites australis and temperature were the key factors controlling the variation of N2O fluxes. The average N2O emission from the microsites above the inflow zones was higher than that above the outflow microsites. High influent strength promoted nitrification and denitrification, and high fluxes were obtained. The clone results showed that Nitrosomonas and Nitrosospira were the main ammonia-oxidizing microorganisms contributing to N2O production in constructed wetlands.